Test agrob bags from 20-5-2020

miguelriemoliveira commented 4 years ago

Hi @aaguiar96 ,

I was really anxious to try out the new bag file.

I ran but could not see the images.

Did you change the topic names? Perhaps you need to fix the playback republish topics?

Thanks for the help, Miguel

miguelriemoliveira commented 4 years ago

BTW, I ran like this:

roslaunch interactive_calibration agrob_calibration.launch bag:=/home/mike/bagfiles/agrob/agrob_calibration_2020-05-20-12-2/calibration_2020-05-20-12-25-58_0.bag

aaguiar96 commented 4 years ago

Hi @miguelriemoliveira

Yes, we had that problem with the topics last time. I fixed the issue this time. I committed the changes to agrob and atlascar right now.

aaguiar96 commented 4 years ago

Hi @miguelriemoliveira and @eupedrosa

Apart from the dimensions, that I can measure, are these parameters correct for the charuco that I used yesterday?

    "calibration_pattern" : {
        "link": "chessboard_link",
        "parent_link": "base_link",
        "origin": [-0.37, 0.0, 0.0, 0, 0.0, 3.8],
        "fixed": false,
        "pattern_type": "charuco",
        "dictionary": "DICT_5X5_100",
        "border_size": 0.03,
        "dimension": {"x": 11, "y": 8},
        "size": 0.06,
        "inner_size": 0.045
    }

I'm trying to create a dataset, but the pattern detections are unstable...

eupedrosa commented 4 years ago

You have the "x" and "y" swapped... EDIT: sorry.. no you do not have them swap....

eupedrosa commented 4 years ago

@aaguiar96, did you used a ruler to measure the size and inner_size?

aaguiar96 commented 4 years ago

@aaguiar96, did you used a ruler to measure the size and inner_size?

Yes, but I also saw from calib.io that the square size is 60mm.

miguelriemoliveira commented 4 years ago

Thanks.

On Thu, May 21, 2020, 09:46 André Aguiar notifications@github.com wrote:

Hi @miguelriemoliveira https://github.com/miguelriemoliveira

Yes, we had that problem with the topics last time. I fixed the issue this time. I committed the changes to agrob and atlascar right now.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/lardemua/AtlasCarCalibration/issues/153#issuecomment-631966088, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACWTHVVYGZTU5NX7YRD4W4TRSTS5DANCNFSM4NGLRE2A .

aaguiar96 commented 4 years ago

@miguelriemoliveira,

Can you execute the system to see the detections? Do you think that is good enough?

miguelriemoliveira commented 4 years ago

The pattern is far away from the cameras. Next bag file if you can bring it closed. Also the illumination sometimes creates reflections which hamper the detection.

I tweked a bit with the aruco detector's parameters. Now we can get detect about 50% of the time.

aaguiar96 commented 4 years ago

Thanks @miguelriemoliveira

I'll try to save a dataset.

miguelriemoliveira commented 4 years ago

Yes,

4ea46400a2209b5a2436377ddabbfbe76e3b6b76

aaguiar96 commented 4 years ago

Hi @miguelriemoliveira and @eupedrosa

I was able to record a dataset with 19 collections. Check it out! :)

agrob_second_dataset.zip

miguelriemoliveira commented 4 years ago

Hi,

Great!

Can you tell me how many collections have all 3 sensors detected?

On Thu, May 21, 2020, 15:21 André Aguiar notifications@github.com wrote:

Hi @miguelriemoliveira https://github.com/miguelriemoliveira and @eupedrosa https://github.com/eupedrosa

I was able to record a dataset with 19 collections. Check it out! :)

agrob_second_dataset.zip https://github.com/lardemua/AtlasCarCalibration/files/4662894/agrob_second_dataset.zip

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/lardemua/AtlasCarCalibration/issues/153#issuecomment-632113568, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACWTHVUKJ44D5YA2FPJAXB3RSU2GDANCNFSM4NGLRE2A .

aaguiar96 commented 4 years ago

All of them, I think! :)

aaguiar96 commented 4 years ago

Yes, all you them!

miguelriemoliveira commented 4 years ago

It looks great. Now you can really test the optimization.

If you need help with something let me know.

eupedrosa commented 4 years ago

@aaguiar96, why is the data_collected.json so large (22MB)? Do you save the (complete) velodine data?

miguelriemoliveira commented 4 years ago

Hi @eupedrosa ,

yes, by design we save the complete sensor data, besides de labels.

eupedrosa commented 4 years ago

Should we follow the same strategy as the images? save on a separate file?

miguelriemoliveira commented 4 years ago

We could. I thought of saving into pcl or ply. It would have to be a format we could read by itself.

The bottomline was, nobody had time to work on it ...

Regards, Miguel

Miguel Riem de Oliveira email: mriem@ua.pt | m.riem.oliveira@gmail.com office: 22.2.18 | ext: 23887 Dep. de Engenharia Mecânica, Universidade de Aveiro Campus Universitário de Santiago, 3810-193 Portugal*

On Thu, 21 May 2020 at 16:07, Eurico F. Pedrosa notifications@github.com wrote:

Should we follow the same strategy as the images? save on a separate file?

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/lardemua/AtlasCarCalibration/issues/153#issuecomment-632138744, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACWTHVT7URYNNVAL46H5JSDRSU7U7ANCNFSM4NGLRE2A .

eupedrosa commented 4 years ago

PLY is a good format. I like that idea.

The bottomline was, nobody had time to work on it ...

I volunteer @aaguiar96 ...

aaguiar96 commented 4 years ago

I volunteer @aaguiar96 ...

Ahahah, sure, I'll work on it (https://github.com/lardemua/AtlasCarCalibration/issues/155)

:D

lardemua / atom

Test agrob bags from 20-5-2020 #153