I'm having some issues getting the correct altitude value to write onto the picture EXIF file. I seem to be getting relative altitude as my output onto the photo. Further, it is not outputting any info to the first 4 photos, and I'm getting massive errors in Photoscan (several tens of thousands of meters).
I'm using the tlog with a 307.8 second offset. CAM message offsets are 1, 4, 3, 5, 2, 8, 7, 6; GPS are 2, 3, 9, 6, 7, 8 from left to right top to bottom in Mission Planner 1.3.7 Build 1.1.5307.24027.
I am waiting on an IR trigger and plan on triggering based on distance, but would like to get this working in the meantime. Let me know if you need some of my pictures (100 photo set @ 7 MB each).
Thanks for any help!
Replies
I thought Yannik was talking about delay and not frequency?
Delay that will not allow to map APM releasing as seen in
the logged data to the actual position when the cam is shooting.
edit: see Reply by Yannick on August 22, 2014 at 7:42am
To give an update on this.
I've measured the delay on my NEX-5R having soldered an opto coupler to the shutter button.
Cam uses RAW and everything manual. Comparing to the gentLED-SHUTTER (IR) cable.
The APM is not involved but the cam directly released by the opto and cable.
avg. 66 msec (range 41..73 msec) shutter button
avg. 288 msec (range 241..323 msec) IR cable
Beside this forward path I found an interesting discussion related to this
https://github.com/diydrones/ardupilot/pull/2383
What the impact on post-processing is I'm not experienced. Just to show these figures.
Wolfgang R.
Hey all!
Yes, i was talking about the delay. The cable modification gave me a constant delay of maybe 0,1s. With the IR-Shutter i had everything: 0,1 - 1s, or even no shutter when the sunlight came in. So i got a big improvement in reliability AND precision.
The triggering speed of the NEX is fast enough in my opinion. My plane has a cruise speed of ~15m/s, so the groundspeed ist normally under 20m/s. My typical shutter distance is 23m, so everything about 1s is okay.
For the PostProcessing i also use Agisoft PS. I get very good results without manual ajustments. The accuracy is around 1m - thats far enough for our needs.
This is a screen from a 2800 pictures project. At this time i don't use GCP. The improvement in accuracy is not needed.
Wolfgang, how precise are your results?
Kannst mir auch gerne eine PN schicken!
Regards,
Yannick
edit:
sorry iskess for not answering your questions one year ago. I hope you were successful anyway?
Thanks Yannick,
I am going to try this out to check whether APM is logging CAM events or not.
The other Do_Digicam_control uses position based image capturing right? Do i need to change any settings to use that option or just select that option while preparing a survey mission?
Thnx once again for the assistance.
Hi all!
@iksess,
i had the same problem with the IR-Trigger and NEX. Missing pictures and the delay was not constant.
So i modified my NEX 5N in this way:
http://photoshipone.com/10/sony-nex5n-trigger-modification/
(attached a cable to the shutter button). This gives a low delay (i think about 100-200 ms with MF) and misses no pictures. But there is still a arduplane bug. Sometimes there is one cam massage missing although the picture was taken. (1 per 1000 pictures)
To connect the NEX with wired shutter, i used a simple optocoupler and the Pixhawk relay output.
@ karthik,
To use "do_digicam_control" you need a "dummy" waypoint before. When the waypoint is reached, the APM/Pixhawk will switch to the "do_digicam_control" command and takes a pictures.
This option has to big disadvantages:
- You need two "waypoints" per picture, so imagine a mission with 1000 pix... impossible, even with Pixhawk (700 waypoints storage)
- The accuracy is not very good. Depending on the waypoint radius, the cam might be triggered meters before the waypoint.
I am not sure if the "do_digicam_control" command creates a CAM-message in the log.
I would suggest to everyone who wants to take pictures for surveying options, to use the CAM Distance trigger.
You have two options:
- Set CAM_TRIGG_DIST = "distance in meters" in the MissionPlanner settings -> it will be used regardless of your flightmode, even when you walk with your model on the groud -> good for testing!
- Use the MissionCommand "DO_SET_CAM_TRIGG_DIST" (needs a dummy waypoint before) to your desired trigger distance in meters. You can even set this value to zero after your last waypoint to disable the trigger for RTH and landing.
With default log settings, ArduPlane logs the CAM-Message. I am not sure if the Copter does, but i think so...
I got georef maps with an ref. error below one meter.
But there is still room for improvement. A customable CAM-Message delay would be very nice, to compensate the trigger delay from the camera.
Nice would be also, when the MissionPlanner could disable the Distance_trigger between turning points for ArduPlane.
Please excuse my bad english, i am from Germany ;-)
Do you need a 2 stage switch for the two stage trigger button? Or does a simple switch provide a full press of the trigger button?
Ps. In the newer firmware versions I believe the need for a dummy WP has been removed, at least in Plane.
Yannick, Thnx a lot mate!!! You have enlightened me on this issue.
I have tested whether the CAM messages are recording or not by setting CAM_TRIGG_DIST to 5 meters. Voila, it is recorded only in the dataflash log file but not in the tlog file.
So i have to remove the APM from my plane every time i land?
I am attaching the files for your reference.
Also could u help me out with camera trigger, currently i use a servo pressing the shutter button on top of my camera. I use a RICOH GRIII and i want to ease up the trigger mechanism.
I am developing a circuit based on Arduino to trigger the camera via a micro USB, Any inputs on wat needs to be changed in the mission planner settings.
Thnx for the kind assistance mate.
2014-08-23 15-38-40.log
2014-08-23 15-39-31.tlog
What are you using to trigger your NEX?
However the last mission, my first with a Pixhawk, is missing several images. I don't know the cause. It could be the IR, but it wasn't a speed issue since it required a much slower trigger rste.
@Yannick,
What other way could you use to trigger a NEX if not IR?
Hey David,
please upload your fotoset and the log to a one-click-host (http://uploaded.net/ would be nice (free for upload)). I will give them a try.
Besides from some missing pictures, i am very happy with the cam message ref.