DroidPlanner / Tower

Ground Control Station for Android Devices
https://play.google.com/store/apps/details?id=org.droidplanner.android
Other
621 stars 554 forks source link

Idea: Effective Survey #637

Closed ghost closed 10 years ago

ghost commented 10 years ago

Following my suggestions on survey planing (https://github.com/DroidPlanner/droidplanner/issues/325) I have some idea to (hopefully) make a survey more effective in terms of flight time. What reduces endurence in a grid survey design is that the copter slows down when reaching a waypoint and then has to speed up again. This takes time and also drains the batteries. To minimize this effect I was thinking of some "phase shift" approach to grid surveys as shown in the figure. A is the common approach, B the phase shift concept. The disadvantage of B is that the flight path needs to be a little longer to cover the same area. However, this should be compensated, because the first and last waypoint are neighbouring and thus the flight path overhead (dashed line in A) is smaller (at least in this design). Maybe this is all too complicated in terms of optimizing line spacing, radii, navigation speed, etc. but I wanted to share the idea anyway to start a discussion on making surveys more effective.

effectivesurvey

arthurbenemann commented 10 years ago

Thats a good idea i have even implemented it before while developing the survey code. But I called it interleaved, like TV line scans.

About the circling on the edges of the survey, that will take some changes on AC code, but from what I understand the developers are already working on that. On Feb 17, 2014 5:29 AM, "ThorstenDP" notifications@github.com wrote:

Following my suggestions on survey planing (#325https://github.com/DroidPlanner/droidplanner/issues/325) I have some idea to (hopefully) make a survey more effective in terms of flight time. What reduces endurence in a grid survey design is that the copter slows down when reaching a waypoint and then has to speed up again. This takes time and also drains the batteries. To minimize this effect I was thinking of some "phase shift" approach to grid surveys as shown in the figure. A is the common approach, B the phase shift concept. The disadvantage of B is that the flight path needs to be a little longer to cover the same area. However, this should be compensated, because the first and last waypoint are neighbouring and thus the flight path overhead (dashed line in A) is smaller (at least in this design). Maybe this is all too complicated in terms of optimizing line spacing, radii, navigation speed, etc. but I wanted to share the idea anyway to start a discussion on making surveys more effective.

[image: effectivesurvey]https://f.cloud.github.com/assets/5606072/2183384/a31cc818-97ab-11e3-8452-fa83febc2c7e.png

Reply to this email directly or view it on GitHubhttps://github.com/DroidPlanner/droidplanner/issues/637 .

ghost commented 10 years ago

Great! Looking forward to it! Not sure if some special circle or spline is needed in AC code. We have WPNAV_RADIUS (Defines the distance from a waypoint, that when crossed indicates the wp has been hit. http://copter.ardupilot.com/wiki/5819-2/). If this is set to a value of maybe 5 m and the distance between the lines is 20 m it should already have some effect comapred to 2 and 10m, respectively. At least I hope so :-) I would be happy to make some systematic tests for fruther refinement once it is implemented.

ghost commented 10 years ago

Another related thing (I first wanted to open a new "Issue" for this, but it is ok here as well): When summarizing the results of the latest DIYDrones T3 "contest" discussion I had another idea which might be worth to implement in the survey section: concatenated missions. This might be interesting for 3D reconstruction where both nadir and oblique images are needed (http://www.diydrones.com/profiles/blogs/t3-season-2-the-model?id=705844%3ABlogPost%3A1324807&page=16#comments). So the first auto mission might be a nadir grid mission followed by a ROI mission where the angle of the camera (using a motorized gimbal for example) is changed for example to 45°. [OT: Some have pending patents on related things (like Parrot: https://www.sensefly.com/news-single/article/senseflys-patent-pending-oblique-imagery.html, https://www.sensefly.com/about/company-profile.html). No problem with closed source and paying for some software but I do not like patents like this. So better to post it here before...] Another example is overlapping grid surveys at different hatch angles. If this is already implemented as well I would be happy to help testing it.

concatenatedsurveys

bbasso commented 10 years ago

The next version of Survey should have a Mission Creation Wizard that ask you some questions....

-What kind of Survey mission? (lawmower, grid, maybe a few others) -Plane or Copter? (changes how the ends of survey lines are treated) -Add auto takeoff and landing points? (again, different for plane and copter) -Constraint: Time, Area, or custom? #325

ghost commented 10 years ago

Perfect! I am happy to help testing!

ghost commented 10 years ago

@bbasso I would be good if the Mission Creation Wizard could be switched of, because I think that at least some of these questions only need to be answered by a user once. If it is part of paged view layout as just proposed by @arthurbenemann here https://github.com/DroidPlanner/droidplanner/issues/325#issuecomment-35319861 - perfect!

arthurbenemann commented 10 years ago

Moved to #884