Triangles of AIS targets are layered over the radar image

keesverruijt / BR24radar_pi

OpenCPN radar plugin for Navico Broadband Radars (BR24, 3G, 4G models) [HISTORIC]

17 stars 25 forks source link

Triangles of AIS targets are layered over the radar image #16

Closed hduerr closed 10 years ago

hduerr commented 10 years ago

Triangles of AIS targets are layered over the radar image. This makes it hard or impossible to verify the AIS target is also a radar contact - which is important, I think. Given the fact that the radar image can be made partly transparent and the AIS triangles cannot, I propose to put the radar image as the topmost layer with chart lowest, then AIS, then radar. Making AIS triangles partly transparent would start a competition in "translucentness" between radar blips and AIS triangles which likely results in both being invisible.

keesverruijt commented 10 years ago

I don't know how O settles on the Z layering. Maybe somebody would like to check out how the AIS plugin does this?

Also, I am not sure I agree. What do the plotters do?

hduerr commented 10 years ago

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

guymadison commented 10 years ago

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

— Reply to this email directly or view it on GitHub.

keesverruijt commented 10 years ago

Guy,

That would be -extremely- helpful. I am a bit busy with work and several open source projects and organising a 600+ entry sailing event. Learning OpenGL is just too big a project right now.

All drawing is done in br24radar_pi::DrawRadarImage in br4radar_pi.cpp.

cowelld commented 10 years ago

Hi Guy, Our problem has been the level of the PC video drivers support for OpenGL and visa versa. We historically had to keep the constructs at the most basic 1.0 level. Going into the textures and higher video features offered in OpenGL 4.0 would be nice but .... At one time I created the option in the Plug-in preferences Display dialog called "GL3+" with the idea of exploring that. If you want to experiment with 4.0 coding you can avail yourself of that section and no-one will be the wiser. I have both AIS and BR24 images on my NSS7 and to be honest they don't work well together. Zooming will often create a huge AIS triangle target which will stay for a couple of cycles before re-sizing. The hollow triangle is apparently a standard symbol for chart imagery. I think there's even an international standard. The AIS imagery isn't done with a plug-in and so we can't influence it's display, it's not even purely on OpenGL canvas. I think Pavel did the AIS sections, at one time we were talking about the AIS and DSC NMEA protocol and getting a recorded file for testing. Unfortunately hd-sf.com seems to have gone away so does anyone know of a free AIS server on line?

Hasta, Dave

----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Sent: Tuesday, April 22, 2014 7:31 AM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

— Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41046543

guymadison commented 10 years ago

Hi Dave,

I will take a look at the code to understand OpenCPN / plug-ins architecture. The framebuffer object features have been in the standard since 2.1, I did the work almost 8 years ago at Apple so it should be available in almost any PC. I am contracting at nVidia for the next few months, I can say PC's lead Apple in OpenGL features by miles.

The glChartCanvas.cpp uses FBO's currently and I have OpenGL enabled on my Mac. I have used OpenCPN on my boat in the past but I would like to use OpenCPN integrated with my NSS8 radar / all the other stuff too so this looks like a good place to start.

Guy

On Apr 22, 2014, at 11:27 AM, Dave Cowell notifications@github.com wrote:

Hi Guy, Our problem has been the level of the PC video drivers support for OpenGL and visa versa. We historically had to keep the constructs at the most basic 1.0 level. Going into the textures and higher video features offered in OpenGL 4.0 would be nice but .... At one time I created the option in the Plug-in preferences Display dialog called "GL3+" with the idea of exploring that. If you want to experiment with 4.0 coding you can avail yourself of that section and no-one will be the wiser. I have both AIS and BR24 images on my NSS7 and to be honest they don't work well together. Zooming will often create a huge AIS triangle target which will stay for a couple of cycles before re-sizing. The hollow triangle is apparently a standard symbol for chart imagery. I think there's even an international standard. The AIS imagery isn't done with a plug-in and so we can't influence it's display, it's not even purely on OpenGL canvas. I think Pavel did the AIS sections, at one time we were talking about the AIS and DSC NMEA protocol and getting a recorded file for testing. Unfortunately hd-sf.com seems to have gone away so does anyone know of a free AIS server on line?

Hasta, Dave

----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Sent: Tuesday, April 22, 2014 7:31 AM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

— Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41046543 — Reply to this email directly or view it on GitHub.

marcus-aa commented 10 years ago

Hi all,

I will provide AIS feeds to everyone involved in this project, as long as we keep them to ourselves. My server handles >20.000 AIS reports per minute currently but I can provide partial feeds, geofilters or variable downsampling.

A problem with the AIS presentation in OpenCPN is that it's not IHO S52 preslib compliant. All AIS targets should be a hollow triangle, no fill colour is allowed. Normal targets are always blue. Dangerous targets (if implemented) are red. Sleeping targets are blue and smaller. When scale allows, targets can be drawn to scale around the GPS position, with or without the not-to scale symbol inside as well. Using these S52 standard symbols there is rarely any problem with hiding information, neither chart-based nor radar-derived.

Targets based on radar information are green, for example ARPA targets are green circles (since their orientation is not known). Most professional systems do target correlation between AIS and radar, and a target that has identity and heading from AIS, but position updates from radar, are green triangles in every ECDIS system I have seen.

I have attached a sample of a S52 compliant presentations. Also the own-ship presentation is S52-compliant, something I dearly miss in OpenCPN. I'm running parallel to Jutlandica on her starboard side. You can see both to-scale and standard AIS symbols. img_0038_2

img_0040_2

Red (Dangerous) AIS target:

Transas Navi-Sailor ECDIS, AIS & ARPA target fusion: img_2117_2

/Marcus

rgleason commented 10 years ago

Marcus, A very useful offer. Do Bdcat and Nohal get a copy of your email? If not, I am sure they would find it useful and I hope you send them then offer. Thanks Rick Gleason

On 4/23/2014 6:11 PM, marcus-aa wrote:

Hi all,

I will provide AIS feeds to everyone involved in this project, as long as we keep them to ourselves. My server handles >20.000 AIS reports per minute currently but I can provide partial feeds, geofilters or variable downsampling.

img_0161 https://cloud.githubusercontent.com/assets/3657845/2783810/0786f15c-cb32-11e3-9e43-13d681e94698.PNG

A problem with the AIS presentation in OpenCPN is that it's not IMO S52 preslib compliant. All AIS targets should be a hollow triangle, no fill colour is allowed. Normal targets are always blue. Dangerous targets (if implemented) are red. Sleeping targets are blue and smaller. When scale allows, targets can be drawn to scale around the GPS position, with or without the not-to scale symbol inside as well. Using these S52 standard symbols there is rarely any problem with hiding information, neither chart-based nor radar-derived.

Targets based on radar information are green, for example ARPA targets are green circles (since their orientation is not known). Most professional systems do target correlation between AIS and radar, and a target that has identity and heading from AIS, but position updates from radar, are green triangles in every ECDIS system I have seen.

I have attached a sample of a S52 compliant presentations. Also the own-ship presentation is S52-compliant, something I dearly miss in OpenCPN. I'm running parallel to Jutlandica on her starboard side. You can see both to-scale and standard AIS symbols. img_0038_2 https://cloud.githubusercontent.com/assets/3657845/2783758/3d7e28d0-cb31-11e3-8cfd-93d1310eb777.PNG

img_0040_2 https://cloud.githubusercontent.com/assets/3657845/2783786/94f4545e-cb31-11e3-8e5d-99b9fe057395.PNG

Red (Dangerous) AIS target: img_0112 https://cloud.githubusercontent.com/assets/3657845/2783842/a1111e60-cb32-11e3-8a06-a9fd6bc4fcd1.PNG

Transas Navi-Sailor ECDIS, AIS & ARPA target fusion: img_2117_2 https://cloud.githubusercontent.com/assets/3657845/2783915/f22e8156-cb33-11e3-915e-90bfa9f7397f.JPG

/Marcus

— Reply to this email directly or view it on GitHub https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41221283.

rgleason commented 10 years ago

Also could you attach the files and detail to your Tracker request (I think it was you...) if not already there?

Thanks

On 4/23/2014 6:11 PM, marcus-aa wrote:

Hi all,

I will provide AIS feeds to everyone involved in this project, as long as we keep them to ourselves. My server handles >20.000 AIS reports per minute currently but I can provide partial feeds, geofilters or variable downsampling.

img_0161 https://cloud.githubusercontent.com/assets/3657845/2783810/0786f15c-cb32-11e3-9e43-13d681e94698.PNG

A problem with the AIS presentation in OpenCPN is that it's not IMO S52 preslib compliant. All AIS targets should be a hollow triangle, no fill colour is allowed. Normal targets are always blue. Dangerous targets (if implemented) are red. Sleeping targets are blue and smaller. When scale allows, targets can be drawn to scale around the GPS position, with or without the not-to scale symbol inside as well. Using these S52 standard symbols there is rarely any problem with hiding information, neither chart-based nor radar-derived.

Targets based on radar information are green, for example ARPA targets are green circles (since their orientation is not known). Most professional systems do target correlation between AIS and radar, and a target that has identity and heading from AIS, but position updates from radar, are green triangles in every ECDIS system I have seen.

I have attached a sample of a S52 compliant presentations. Also the own-ship presentation is S52-compliant, something I dearly miss in OpenCPN. I'm running parallel to Jutlandica on her starboard side. You can see both to-scale and standard AIS symbols. img_0038_2 https://cloud.githubusercontent.com/assets/3657845/2783758/3d7e28d0-cb31-11e3-8cfd-93d1310eb777.PNG

img_0040_2 https://cloud.githubusercontent.com/assets/3657845/2783786/94f4545e-cb31-11e3-8e5d-99b9fe057395.PNG

Red (Dangerous) AIS target: img_0112 https://cloud.githubusercontent.com/assets/3657845/2783842/a1111e60-cb32-11e3-8a06-a9fd6bc4fcd1.PNG

Transas Navi-Sailor ECDIS, AIS & ARPA target fusion: img_2117_2 https://cloud.githubusercontent.com/assets/3657845/2783915/f22e8156-cb33-11e3-915e-90bfa9f7397f.JPG

/Marcus

— Reply to this email directly or view it on GitHub https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41221283.

marcus-aa commented 10 years ago

On 2014-04-24, at 00:19, Rick Gleason wrote:

A very useful offer. Do Bdcat and Nohal get a copy of your email? If not, I am sure they would find it useful and I hope you send them then offer.

No, I don't know if they read this here. But I will extend my offer to them as well if they are interested.

Also could you attach the files and detail to your Tracker request (I think it was you...) if not already there?

I have never used the tracker but I can add it there as well. Do you have a # for it I can find ?

rgleason commented 10 years ago

http://opencpn.org/ocpn/flyspray/ http://opencpn.org/ocpn/

You can register from the Tracker Tab. Projects where you can enter your requests (use drop down tab on right):

Opencpn Feature Requests -Start title with "AIS" http://opencpn.org/ocpn/flyspray/index.php?do=index&project=6&do=index&project=6&order=summary&sort=asc These should be sorted by clicking on "Summary" to sort, and bring up all AIS requests. One of the interests is to get less clutter...seems to be a possible driver.

Gradar_pi http://opencpn.org/ocpn/flyspray/index.php?project=26&do=index&switch=1

BR24 Radar (beta) http://opencpn.org/ocpn/flyspray/index.php?project=48&do=index&switch=1

Dave [Bdcat] and Nohal [Pavel] are the main programmers. They can be found on the forum. You could send a private message to them by clicking on their name...

See this forum http://www.cruisersforum.com/forums/f134/tracker-ais-feature-requests-111372-3.html#post1525543 Bdcat recently entered a post.... above.. http://www.cruisersforum.com/forums/f134/tracker-ais-feature-requests-111372-3.html#post1519937

Thanks very much.

On 4/23/2014 6:47 PM, marcus-aa wrote:

On 2014-04-24, at 00:19, Rick Gleason wrote:

A very useful offer. Do Bdcat and Nohal get a copy of your email? If not, I am sure they would find it useful and I hope you send them then offer.

No, I don't know if they read this here. But I will extend my offer to them as well if they are interested.

Also could you attach the files and detail to your Tracker request (I think it was you...) if not already there?

I have never used the tracker but I can add it there as well. Do you have a # for it I can find ?

/M

— Reply to this email directly or view it on GitHub https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41224073.

hduerr commented 10 years ago

Status of this issue:

I understand the consensus is that AIS triangles should be "hollow" (only consist of outline) following a well established standard. Please comment if you disagree.
If this is our desired solution, then this issue is no longer valid, and will be closed by me, because we don't want the radar image to be layered over the AIS triangles.
Compared to the standard, there are a number of other issues with the current implementation of AIS in OCPN. marcus-aa has volunteered to create issues for the main application.
Marcus, please let me know via comment or email when this is done. I will then close this issue.

cowelld commented 10 years ago

Hey Guy, That's great! In line with the previously mentioned, how about sending me a Linux driver for ... oh yeah, that's a SIS671 chip, not nVidia. Unfortunately with Windows DirectX in competition with OpenGL and the driving force for OGL being the chip makers it's a bit like tail-wagging-dog. We have to keep from going outside the OpenCPN guidelines too. As far as integration with the NSS all I do is use the TCP/IP feature for the NMEA data stream. It saves a serial connection since everything I have is SIMNET. It's too bad SIM/LOW didn't use a UDP structure to do the NMEA since it'd allow multiple PC's to monitor the data simultaneously (and make PlayCap work for recordings). I've tested the GOFREE apps with a WIFI unit and that seems to be a better methodology for and alternate remote control of the NSS. I thought of using a HR20 system but haven't had a chance to sniff that protocol due to BT. (I need to find the serial out of the micro before it goes into the BT controller.) I'd bet it's only CANBUS over BT. Where will you be working? Silicon Valley? See if they'll spot you a Tegra

I haven't checked that out but guess it's probably another ARM machine. I have an ARM7v dual in a Wandboard that I used to load Arch Linux, the CPP compilers and then finally natively compile OpenGL, wxWidgets and then OpenCPN. (Some of that was downloadable from the ARCH USR group though.) Next time use my direct email address. Going through the Notification portal seems to be inviting surveillance.

Hasta, Dave ----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Cc: "Dave Cowell" david.cowell@ymail.com Sent: Tuesday, April 22, 2014 8:55 PM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

Hi Dave,

I will take a look at the code to understand OpenCPN / plug-ins architecture. The framebuffer object features have been in the standard since 2.1, I did the work almost 8 years ago at Apple so it should be available in almost any PC. I am contracting at nVidia for the next few months, I can say PC's lead Apple in OpenGL features by miles.

The glChartCanvas.cpp uses FBO's currently and I have OpenGL enabled on my Mac. I have used OpenCPN on my boat in the past but I would like to use OpenCPN integrated with my NSS8 radar / all the other stuff too so this looks like a good place to start.

Guy

On Apr 22, 2014, at 11:27 AM, Dave Cowell notifications@github.com wrote:

Hi Guy, Our problem has been the level of the PC video drivers support for OpenGL and visa versa. We historically had to keep the constructs at the most basic 1.0 level. Going into the textures and higher video features offered in OpenGL 4.0 would be nice but .... At one time I created the option in the Plug-in preferences Display dialog called "GL3+" with the idea of exploring that. If you want to experiment with 4.0 coding you can avail yourself of that section and no-one will be the wiser. I have both AIS and BR24 images on my NSS7 and to be honest they don't work well together. Zooming will often create a huge AIS triangle target which will stay for a couple of cycles before re-sizing. The hollow triangle is apparently a standard symbol for chart imagery. I think there's even an international standard. The AIS imagery isn't done with a plug-in and so we can't influence it's display, it's not even purely on OpenGL canvas. I think Pavel did the AIS sections, at one time we were talking about the AIS and DSC NMEA protocol and getting a recorded file for testing. Unfortunately hd-sf.com seems to have gone away so does anyone know of a free AIS server on line?

Hasta, Dave

----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Sent: Tuesday, April 22, 2014 7:31 AM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

— Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41046543 — Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41122195

cowelld commented 10 years ago

Hi Marcus, Which language did you use for your work, Python or C++? When I was working with Nohan on the DSC issues I took a major dive into the AIS software. If you haven't seen these websites: http://gpsd.berlios.de/AIVDM.html has all the coding data. and this one by Kurt Schwehr was developed for the US Coast Guard https://pypi.python.org/pypi/noaadata-py/ while he was a post-grad at UNH. Reading his blog is very interesting. He now works for Google here in Cal.

Hasta, Dave

----- Original Message ----- From: "marcus-aa" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Cc: "Dave Cowell" david.cowell@ymail.com Sent: Wednesday, April 23, 2014 3:11 PM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

Hi all,

I will provide AIS feeds to everyone involved in this project, as long as we keep them to ourselves. My server handles >20.000 AIS reports per minute currently but I can provide partial feeds, geofilters or variable downsampling.

A problem with the AIS presentation in OpenCPN is that it's not IMO S52 preslib compliant. All AIS targets should be a hollow triangle, no fill colour is allowed. Normal targets are always blue. Dangerous targets (if implemented) are red. Sleeping targets are blue and smaller. When scale allows, targets can be drawn to scale around the GPS position, with or without the not-to scale symbol inside as well. Using these S52 standard symbols there is rarely any problem with hiding information, neither chart-based nor radar-derived.

Targets based on radar information are green, for example ARPA targets are green circles (since their orientation is not known). Most professional systems do target correlation between AIS and radar, and a target that has identity and heading from AIS, but position updates from radar, are green triangles in every ECDIS system I have seen.

I have attached a sample of a S52 compliant presentations. Also the own-ship presentation is S52-compliant, something I dearly miss in OpenCPN. I'm running parallel to Jutlandica on her starboard side. You can see both to-scale and standard AIS symbols.

Red (Dangerous) AIS target:

Transas Navi-Sailor ECDIS, AIS & ARPA target fusion:

/Marcus

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41221283

guymadison commented 10 years ago

Thanks Dave,

I have to figure out the the TCP/IP for my setup, I bought the Simrad wireless unit also is there software to connect to that directly?

Mike

On Apr 25, 2014, at 12:23 PM, Dave Cowell notifications@github.com wrote:

Hey Guy, That's great! In line with the previously mentioned, how about sending me a Linux driver for ... oh yeah, that's a SIS671 chip, not nVidia. Unfortunately with Windows DirectX in competition with OpenGL and the driving force for OGL being the chip makers it's a bit like tail-wagging-dog. We have to keep from going outside the OpenCPN guidelines too. As far as integration with the NSS all I do is use the TCP/IP feature for the NMEA data stream. It saves a serial connection since everything I have is SIMNET. It's too bad SIM/LOW didn't use a UDP structure to do the NMEA since it'd allow multiple PC's to monitor the data simultaneously (and make PlayCap work for recordings). I've tested the GOFREE apps with a WIFI unit and that seems to be a better methodology for and alternate remote control of the NSS. I thought of using a HR20 system but haven't had a chance to sniff that protocol due to BT. (I need to find the serial out of the micro before it goes into the BT controller.) I'd bet it's only CANBUS over BT. Where will you be working? Silicon Valley? See if they'll spot you a Tegra

I haven't checked that out but guess it's probably another ARM machine. I have an ARM7v dual in a Wandboard that I used to load Arch Linux, the CPP compilers and then finally natively compile OpenGL, wxWidgets and then OpenCPN. (Some of that was downloadable from the ARCH USR group though.) Next time use my direct email address. Going through the Notification portal seems to be inviting surveillance.

Hasta, Dave ----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Cc: "Dave Cowell" david.cowell@ymail.com Sent: Tuesday, April 22, 2014 8:55 PM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

Hi Dave,

I will take a look at the code to understand OpenCPN / plug-ins architecture. The framebuffer object features have been in the standard since 2.1, I did the work almost 8 years ago at Apple so it should be available in almost any PC. I am contracting at nVidia for the next few months, I can say PC's lead Apple in OpenGL features by miles.

The glChartCanvas.cpp uses FBO's currently and I have OpenGL enabled on my Mac. I have used OpenCPN on my boat in the past but I would like to use OpenCPN integrated with my NSS8 radar / all the other stuff too so this looks like a good place to start.

Guy

On Apr 22, 2014, at 11:27 AM, Dave Cowell notifications@github.com wrote:

Hi Guy, Our problem has been the level of the PC video drivers support for OpenGL and visa versa. We historically had to keep the constructs at the most basic 1.0 level. Going into the textures and higher video features offered in OpenGL 4.0 would be nice but .... At one time I created the option in the Plug-in preferences Display dialog called "GL3+" with the idea of exploring that. If you want to experiment with 4.0 coding you can avail yourself of that section and no-one will be the wiser. I have both AIS and BR24 images on my NSS7 and to be honest they don't work well together. Zooming will often create a huge AIS triangle target which will stay for a couple of cycles before re-sizing. The hollow triangle is apparently a standard symbol for chart imagery. I think there's even an international standard. The AIS imagery isn't done with a plug-in and so we can't influence it's display, it's not even purely on OpenGL canvas. I think Pavel did the AIS sections, at one time we were talking about the AIS and DSC NMEA protocol and getting a recorded file for testing. Unfortunately hd-sf.com seems to have gone away so does anyone know of a free AIS server on line?

Hasta, Dave

----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Sent: Tuesday, April 22, 2014 7:31 AM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

— Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41046543 — Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41122195 — Reply to this email directly or view it on GitHub.

guymadison commented 10 years ago

Dave,

I work out of Seattle, I ordered the Jetson TK1 last week (http://blogs.nvidia.com/blog/2014/04/25/win-jetson-tk1) its another ARM board (quad core and a serious GPU) BUT it has a SATA connector on it which makes it ideal for a real machine onboard and something that can stay onboard. It has a HDMI connector, RS232, network, all the std stuff.

There is a GOFREE dev page (http://www.simrad-yachting.com/en-US/Products/GoFree-Wireless-Technology/GoFree-Toolkit/), download these pages before Simrad changes their mind.

Looking at OpenCPN again, I need to figure out how to import Canadian charts to make it really useful for me. I think I can buy electronic versions but they may be online.

Lots of stuff to figure out in the next few days.

I guess you know my internet alias :) Guy was my best dog..

Thanks,

Guy

On Apr 25, 2014, at 12:23 PM, Dave Cowell notifications@github.com wrote:

Hey Guy, That's great! In line with the previously mentioned, how about sending me a Linux driver for ... oh yeah, that's a SIS671 chip, not nVidia. Unfortunately with Windows DirectX in competition with OpenGL and the driving force for OGL being the chip makers it's a bit like tail-wagging-dog. We have to keep from going outside the OpenCPN guidelines too. As far as integration with the NSS all I do is use the TCP/IP feature for the NMEA data stream. It saves a serial connection since everything I have is SIMNET. It's too bad SIM/LOW didn't use a UDP structure to do the NMEA since it'd allow multiple PC's to monitor the data simultaneously (and make PlayCap work for recordings). I've tested the GOFREE apps with a WIFI unit and that seems to be a better methodology for and alternate remote control of the NSS. I thought of using a HR20 system but haven't had a chance to sniff that protocol due to BT. (I need to find the serial out of the micro before it goes into the BT controller.) I'd bet it's only CANBUS over BT. Where will you be working? Silicon Valley? See if they'll spot you a Tegra

I haven't checked that out but guess it's probably another ARM machine. I have an ARM7v dual in a Wandboard that I used to load Arch Linux, the CPP compilers and then finally natively compile OpenGL, wxWidgets and then OpenCPN. (Some of that was downloadable from the ARCH USR group though.) Next time use my direct email address. Going through the Notification portal seems to be inviting surveillance.

Hasta, Dave ----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Cc: "Dave Cowell" david.cowell@ymail.com Sent: Tuesday, April 22, 2014 8:55 PM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

Hi Dave,

I will take a look at the code to understand OpenCPN / plug-ins architecture. The framebuffer object features have been in the standard since 2.1, I did the work almost 8 years ago at Apple so it should be available in almost any PC. I am contracting at nVidia for the next few months, I can say PC's lead Apple in OpenGL features by miles.

The glChartCanvas.cpp uses FBO's currently and I have OpenGL enabled on my Mac. I have used OpenCPN on my boat in the past but I would like to use OpenCPN integrated with my NSS8 radar / all the other stuff too so this looks like a good place to start.

Guy

On Apr 22, 2014, at 11:27 AM, Dave Cowell notifications@github.com wrote:

Hi Guy, Our problem has been the level of the PC video drivers support for OpenGL and visa versa. We historically had to keep the constructs at the most basic 1.0 level. Going into the textures and higher video features offered in OpenGL 4.0 would be nice but .... At one time I created the option in the Plug-in preferences Display dialog called "GL3+" with the idea of exploring that. If you want to experiment with 4.0 coding you can avail yourself of that section and no-one will be the wiser. I have both AIS and BR24 images on my NSS7 and to be honest they don't work well together. Zooming will often create a huge AIS triangle target which will stay for a couple of cycles before re-sizing. The hollow triangle is apparently a standard symbol for chart imagery. I think there's even an international standard. The AIS imagery isn't done with a plug-in and so we can't influence it's display, it's not even purely on OpenGL canvas. I think Pavel did the AIS sections, at one time we were talking about the AIS and DSC NMEA protocol and getting a recorded file for testing. Unfortunately hd-sf.com seems to have gone away so does anyone know of a free AIS server on line?

Hasta, Dave

----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Sent: Tuesday, April 22, 2014 7:31 AM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

— Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41046543 — Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41122195 — Reply to this email directly or view it on GitHub.

cowelld commented 10 years ago

Hi Mike, Sorry about calling you Guy. The GOFREE apps are free downloads from the Simrad site. They're set up for Androids and iPads on WIFI only though. I'd bet there's probably a PC program that'll emulate those systems or something or another. It'd make a lot of this OpenCPN programming unnecessary

Hasta, Dave.

On 4/26/2014 9:03 AM, guymadison wrote:

Thanks Dave,

I have to figure out the the TCP/IP for my setup, I bought the Simrad wireless unit also is there software to connect to that directly?

Mike

keesverruijt commented 10 years ago

Well, actually the GoFree remote apps are on the Play store (Android) and the App store (iOS). Given how the image appears it looks to me like some form of VNC with a lossy compression method. AFAIK this has not yet been reverse engineered.

The Simrad GoFree WiFi-1 is a WiFi access point that is special in that it can be administered from the chart plotter. Maybe they filter out some of the multicast packets, but a standard AP wouldn’t send those out without a listener anyway. In my experience the multicast over WiFi is pretty much useless, as most APs send them at the lowest possible rate (= 1 Mbps) so that all listeners can keep up, and then the link is saturated. Which is why we are saying: don’t.

The third part of GoFree is NMEA 2000 and NMEA 0183 streams over TCP/IP to/from the plotters. NMEA 0183 reading is publicly accessible, anything more requires a license with Navico. AFAIK this has not yet been reverse engineered.

Kees

cowelld commented 10 years ago

Hi Again, Which GPU? I'm stuck with a Vivante GPU that doesn't have the full support for Linux and OpenGL yet. They've a proprietary API but it's not generally available yet. There's been a ton of op system hacks for it but they all lack that graphics capability to support the OpenGL we use in the radar program. My version is the dual core Freescale i.MX6 which can connect to SATA but has dual memory chip ports and with the 16 GB chips should be enough without the added power load. I wish I'd bought a model with built in CANBUS though since that'd allow future NMEA 2000. Did you try making GOOGLE charts? You can convert the satellite images to KAP versions but of course they don't have the standard chart symbols. Of a curious side one of the group, Matt, was using the radar and charts in the Marshal Islands and found out how off the charts were. None of the echos matched the charts but when he downloaded a Google image it matched perfectly. If you haven't seen the SeaClear web references then maybe there's a solution there. SeaClear is a precursor of OpenCPN and has had the same issues with a lot of people resolving them.

Hasta, Dave.

On 4/26/2014 9:25 AM, guymadison wrote:

Dave,

I work out of Seattle, I ordered the Jetson TK1 last week (http://blogs.nvidia.com/blog/2014/04/25/win-jetson-tk1) its another ARM board (quad core and a serious GPU) BUT it has a SATA connector on it which makes it ideal for a real machine onboard and something that can stay onboard. It has a HDMI connector, RS232, network, all the std stuff.

There is a GOFREE dev page (http://www.simrad-yachting.com/en-US/Products/GoFree-Wireless-Technology/GoFree-Toolkit/), download these pages before Simrad changes their mind.

Looking at OpenCPN again, I need to figure out how to import Canadian charts to make it really useful for me. I think I can buy electronic versions but they may be online.

Lots of stuff to figure out in the next few days.

I guess you know my internet alias :) Guy was my best dog..

Thanks,

Guy

On Apr 25, 2014, at 12:23 PM, Dave Cowell notifications@github.com wrote:

Hey Guy, That's great! In line with the previously mentioned, how about sending me a Linux driver for ... oh yeah, that's a SIS671 chip, not nVidia. Unfortunately with Windows DirectX in competition with OpenGL and the driving force for OGL being the chip makers it's a bit like tail-wagging-dog. We have to keep from going outside the OpenCPN guidelines too. As far as integration with the NSS all I do is use the TCP/IP feature for the NMEA data stream. It saves a serial connection since everything I have is SIMNET. It's too bad SIM/LOW didn't use a UDP structure to do the NMEA since it'd allow multiple PC's to monitor the data simultaneously (and make PlayCap work for recordings). I've tested the GOFREE apps with a WIFI unit and that seems to be a better methodology for and alternate remote control of the NSS. I thought of using a HR20 system but haven't had a chance to sniff that protocol due to BT. (I need to find the serial out of the micro before it goes into the BT controller.) I'd bet it's only CANBUS over BT. Where will you be working? Silicon Valley? See if they'll spot you a Tegra

I haven't checked that out but guess it's probably another ARM machine. I have an ARM7v dual in a Wandboard that I used to load Arch Linux, the CPP compilers and then finally natively compile OpenGL, wxWidgets and then OpenCPN. (Some of that was downloadable from the ARCH USR group though.) Next time use my direct email address. Going through the Notification portal seems to be inviting surveillance.

Hasta, Dave ----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Cc: "Dave Cowell" david.cowell@ymail.com Sent: Tuesday, April 22, 2014 8:55 PM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

Hi Dave,

I will take a look at the code to understand OpenCPN / plug-ins architecture. The framebuffer object features have been in the standard since 2.1, I did the work almost 8 years ago at Apple so it should be available in almost any PC. I am contracting at nVidia for the next few months, I can say PC's lead Apple in OpenGL features by miles.

The glChartCanvas.cpp uses FBO's currently and I have OpenGL enabled on my Mac. I have used OpenCPN on my boat in the past but I would like to use OpenCPN integrated with my NSS8 radar / all the other stuff too so this looks like a good place to start.

Guy

On Apr 22, 2014, at 11:27 AM, Dave Cowell notifications@github.com wrote:

Hi Guy, Our problem has been the level of the PC video drivers support for OpenGL and visa versa. We historically had to keep the constructs at the most basic 1.0 level. Going into the textures and higher video features offered in OpenGL 4.0 would be nice but .... At one time I created the option in the Plug-in preferences Display dialog called "GL3+" with the idea of exploring that. If you want to experiment with 4.0 coding you can avail yourself of that section and no-one will be the wiser. I have both AIS and BR24 images on my NSS7 and to be honest they don't work well together. Zooming will often create a huge AIS triangle target which will stay for a couple of cycles before re-sizing. The hollow triangle is apparently a standard symbol for chart imagery. I think there's even an international standard. The AIS imagery isn't done with a plug-in and so we can't influence it's display, it's not even purely on OpenGL canvas. I think Pavel did the AIS sections, at one time we were talking about the AIS and DSC NMEA protocol and getting a recorded file for testing. Unfortunately hd-sf.com seems to have gone away so does anyone know of a free AIS server on line?

Hasta, Dave

----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Sent: Tuesday, April 22, 2014 7:31 AM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

— Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub:

https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41046543 — Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41122195 — Reply to this email directly or view it on GitHub.

— Reply to this email directly or view it on GitHub https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41473114.

hduerr commented 10 years ago

Re GoFree: AFAIK GoFree is an API that can be used (a) to pull NMEA0183 data from a Navico MFD and (b) to view and remote control a touch-based Navico MFD (for Simrad this is NSS NSSevo2 and NSO, not NSE). It was my understanding that it is free to get and use by application developers and users. There is at least one implementation for a PC, a commercial application, a logbook software for pleasure boats, from TripCon www.tricon.com, made in Berlin, Germany. I have a test version installed and consider getting a license at EUR 199 for remote-controlling the MFD. Here are 2 screenshots: 20140428_gofree_1 20140428_gofree_2

guymadison commented 10 years ago

Its a Tegra K1 GPU, its pretty stout with 192 processing units.

I will take a look at SeaClear, Jeppsen makes a mobile app with all the Canadian charts but doesn't sell a PC / Mac version. I inquired about this since iOS uses the same framework and actually was put in touch with the software manager of the group... seems they can sell the mobile app for just about nothing @ $29 but have no interest in a PC / Mac version or would prefer to screw their PC / Mac customers with digital maps for $185.

Mike

On Apr 28, 2014, at 8:28 AM, Dave Cowell notifications@github.com wrote:

Hi Again, Which GPU? I'm stuck with a Vivante GPU that doesn't have the full support for Linux and OpenGL yet. They've a proprietary API but it's not generally available yet. There's been a ton of op system hacks for it but they all lack that graphics capability to support the OpenGL we use in the radar program. My version is the dual core Freescale i.MX6 which can connect to SATA but has dual memory chip ports and with the 16 GB chips should be enough without the added power load. I wish I'd bought a model with built in CANBUS though since that'd allow future NMEA 2000. Did you try making GOOGLE charts? You can convert the satellite images to KAP versions but of course they don't have the standard chart symbols. Of a curious side one of the group, Matt, was using the radar and charts in the Marshal Islands and found out how off the charts were. None of the echos matched the charts but when he downloaded a Google image it matched perfectly. If you haven't seen the SeaClear web references then maybe there's a solution there. SeaClear is a precursor of OpenCPN and has had the same issues with a lot of people resolving them.

Hasta, Dave.

On 4/26/2014 9:25 AM, guymadison wrote:

Dave,

I work out of Seattle, I ordered the Jetson TK1 last week (http://blogs.nvidia.com/blog/2014/04/25/win-jetson-tk1) its another ARM board (quad core and a serious GPU) BUT it has a SATA connector on it which makes it ideal for a real machine onboard and something that can stay onboard. It has a HDMI connector, RS232, network, all the std stuff.

There is a GOFREE dev page (http://www.simrad-yachting.com/en-US/Products/GoFree-Wireless-Technology/GoFree-Toolkit/), download these pages before Simrad changes their mind.

Looking at OpenCPN again, I need to figure out how to import Canadian charts to make it really useful for me. I think I can buy electronic versions but they may be online.

Lots of stuff to figure out in the next few days.

I guess you know my internet alias :) Guy was my best dog..

Thanks,

Guy

On Apr 25, 2014, at 12:23 PM, Dave Cowell notifications@github.com wrote:

Hey Guy, That's great! In line with the previously mentioned, how about sending me a Linux driver for ... oh yeah, that's a SIS671 chip, not nVidia. Unfortunately with Windows DirectX in competition with OpenGL and the driving force for OGL being the chip makers it's a bit like tail-wagging-dog. We have to keep from going outside the OpenCPN guidelines too. As far as integration with the NSS all I do is use the TCP/IP feature for the NMEA data stream. It saves a serial connection since everything I have is SIMNET. It's too bad SIM/LOW didn't use a UDP structure to do the NMEA since it'd allow multiple PC's to monitor the data simultaneously (and make PlayCap work for recordings). I've tested the GOFREE apps with a WIFI unit and that seems to be a better methodology for and alternate remote control of the NSS. I thought of using a HR20 system but haven't had a chance to sniff that protocol due to BT. (I need to find the serial out of the micro before it goes into the BT controller.) I'd bet it's only CANBUS over BT. Where will you be working? Silicon Valley? See if they'll spot you a Tegra

I haven't checked that out but guess it's probably another ARM machine. I have an ARM7v dual in a Wandboard that I used to load Arch Linux, the CPP compilers and then finally natively compile OpenGL, wxWidgets and then OpenCPN. (Some of that was downloadable from the ARCH USR group though.) Next time use my direct email address. Going through the Notification portal seems to be inviting surveillance.

Hasta, Dave ----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Cc: "Dave Cowell" david.cowell@ymail.com Sent: Tuesday, April 22, 2014 8:55 PM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

Hi Dave,

I will take a look at the code to understand OpenCPN / plug-ins architecture. The framebuffer object features have been in the standard since 2.1, I did the work almost 8 years ago at Apple so it should be available in almost any PC. I am contracting at nVidia for the next few months, I can say PC's lead Apple in OpenGL features by miles.

The glChartCanvas.cpp uses FBO's currently and I have OpenGL enabled on my Mac. I have used OpenCPN on my boat in the past but I would like to use OpenCPN integrated with my NSS8 radar / all the other stuff too so this looks like a good place to start.

Guy

On Apr 22, 2014, at 11:27 AM, Dave Cowell notifications@github.com wrote:

Hi Guy, Our problem has been the level of the PC video drivers support for OpenGL and visa versa. We historically had to keep the constructs at the most basic 1.0 level. Going into the textures and higher video features offered in OpenGL 4.0 would be nice but .... At one time I created the option in the Plug-in preferences Display dialog called "GL3+" with the idea of exploring that. If you want to experiment with 4.0 coding you can avail yourself of that section and no-one will be the wiser. I have both AIS and BR24 images on my NSS7 and to be honest they don't work well together. Zooming will often create a huge AIS triangle target which will stay for a couple of cycles before re-sizing. The hollow triangle is apparently a standard symbol for chart imagery. I think there's even an international standard. The AIS imagery isn't done with a plug-in and so we can't influence it's display, it's not even purely on OpenGL canvas. I think Pavel did the AIS sections, at one time we were talking about the AIS and DSC NMEA protocol and getting a recorded file for testing. Unfortunately hd-sf.com seems to have gone away so does anyone know of a free AIS server on line?

Hasta, Dave

----- Original Message ----- From: "guymadison" notifications@github.com To: "canboat/BR24radar_pi" BR24radar_pi@noreply.github.com Sent: Tuesday, April 22, 2014 7:31 AM Subject: Re: [BR24radar_pi] Triangles of AIS targets are layered over the radar image (#16)

If you want to do proper layering do it with textures and frame buffer objects then composite textures drawn as quads.

I worked in the OpenGL team at Apple... kind of know OpenGL :0

I will download what you have to see where I can help.

Guy

On Apr 21, 2014, at 4:21 PM, hduerr notifications@github.com wrote:

I don't know from my own experience because I don't have AIS interfaced to the NSS (yet - my AIS is NMEA0183 only; since I also have the cockpit Zeus Touch, I want to do this with NMEA2000 to avoid running 10 meters of wiring and another deck gland; an NAIS-400 is on my wishlist). However, I find a screenshot in the NSS user manual (page 62) showing radar only with AIS targets. It is a poor resolution but indicates that the triangles are actually on top of the radar image but the triangles are transparent inside (only consist of outline), so the radar image is visible inside the triangle. This would also be an acceptable way for OCPN IMO. But one way or the other, I consider it important to be able to tell if an AIS contact is also a radar contact - without having to constantly use the "Hide AIS targets" toolbar button. Can we (the radar people) just decide to put our image over everything else? Or do we have to ask the AIS people to back down one layer? If it's our decision, then lets go for it - ideally fixing the transparency at the same time. Otherwise there is a danger of blanking out AIS targets. If it's not our decision, we could suggest to the AIS people to use outline-only (transparent inside) triangles or to make this a configurable option.

Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub:

https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41046543 Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub: https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41122195 Reply to this email directly or view it on GitHub.

Reply to this email directly or view it on GitHub https://github.com/canboat/BR24radar_pi/issues/16#issuecomment-41473114.

Reply to this email directly or view it on GitHub.

cowelld commented 10 years ago

Hi again, Do you have a recent copy of the BR24radar_pi.dll for Windows handy? I'm down at the boat and just realized my current version is several iterations old. You're the only other in my time zone.

Thanks, Dave.

keesverruijt commented 10 years ago

Dave, it's on its way.

keesverruijt commented 10 years ago

Note that I've created a new Github organisation so that we now have a site to publish binary releases at https://opencpn-navico-radar-plugin.github.io.

Henning,

As the page referenced by @guymadison states there are 5 tiers to GoFree, only the first 2 are free.

The GoFree open protocol data strategy groups data into 5 different Tiers. Access to data varies, but essentially all data is free to use if it enhances the desirability of Simrad MFDs. Specifications for Tier 1 and Tier 2 (view only) are available to download for free - follow the links below. Any developer wishing to use Tier 2 control or Tiers 3 to 5 should contact gofree@navico.com with a proposal.Support for Tiers 2-5 will be included in the Simrad NSS RTM3 Software release, available from late 2012. Prior to the RTM3 software production release, a special “developer’s build” will be available for testing against. It is important to note that Simrad retains the right to change the Tier 2 specification prior to the features being enabled in the first production release.

guymadison commented 10 years ago

I think any of the tiers above tier 2 view are for control, should we put together a proposal ;)

I will start here, GoFree supports a Bonjour style internet connection which I can bring all my stuff (and future stuff) together with.

Mike

On Apr 28, 2014, at 3:09 PM, canboat notifications@github.com wrote:

Note that I've created a new Github organisation so that we now have a site to publish binary releases at https://opencpn-navico-radar-plugin.github.io.

Henning,

As the page referenced by @guymadison states there are 5 tiers to GoFree, only the first 2 are free.

The GoFree open protocol data strategy groups data into 5 different Tiers. Access to data varies, but essentially all data is free to use if it enhances the desirability of Simrad MFDs. Specifications for Tier 1 and Tier 2 (view only) are available to download for free - follow the links below. Any developer wishing to use Tier 2 control or Tiers 3 to 5 should contact gofree@navico.com with a proposal.Support for Tiers 2-5 will be included in the Simrad NSS RTM3 Software release, available from late 2012. Prior to the RTM3 software production release, a special “developer’s build” will be available for testing against. It is important to note that Simrad retains the right to change the Tier 2 specification prior to the features being enabled in the first production release.

— Reply to this email directly or view it on GitHub.

keesverruijt commented 10 years ago

On 29 Apr 2014, at 04:22, guymadison notifications@github.com wrote:

I think any of the tiers above tier 2 view are for control, should we put together a proposal ;)

I will start here, GoFree supports a Bonjour style internet connection which I can bring all my stuff (and future stuff) together with.

Not sure what you meant here. If you are thinking of working with the Navico JSON web sockets protocol may I suggest the Signal K project — we are trying to come up with a standardised way for open source marine apps to communicate: https://signalk.github.io

— Kees

guymadison commented 10 years ago

That looks like a great place to start

On Apr 30, 2014, at 3:05 AM, canboat notifications@github.com wrote:

On 29 Apr 2014, at 04:22, guymadison notifications@github.com wrote:

I think any of the tiers above tier 2 view are for control, should we put together a proposal ;)

I will start here, GoFree supports a Bonjour style internet connection which I can bring all my stuff (and future stuff) together with.

Not sure what you meant here. If you are thinking of working with the Navico JSON web sockets protocol may I suggest the Signal K project — we are trying to come up with a standardised way for open source marine apps to communicate: https://signalk.github.io

— Kees — Reply to this email directly or view it on GitHub.

keesverruijt commented 10 years ago

This was a nice discussion, but since we can't change the way OpenCPN layers our display let's close this issue.