Currently lost.

[R.P..R]

Hello gang.

Last week, I was mapping a small area. I usually do a double grid if there is a structure and set the gimbal at -45deg at 100ft alt. and do two separate missions.

1) The first data, came out really BAD!


2) Second data a week after.



Any idea of what might have caused this? Camera? Exposure? (My settings are the same all through my workflow)

 

[Outta Control]

What were the environmental and flight conditions?

Temp, Wind, Shutter speed setting, Exposure, Humidity, etc can all affect it.

 

[R.P..R]

What were the environmental and flight conditions?

Temp, Wind, Shutter speed setting, Exposure, Humidity, etc can all affect it.

First mission: July 5th



Second mission: July 13th

 

[Outta Control]

Did you use the same SD card? Same brand?

Phantom 4 Pro correct?

ND Filter?

Manual settings?

 

[R.P..R]

i1 Oly 12mm

I hardly use ND filters to map. Manual, expose to ground, with no crushed shadows and blown highlights.

And....... No, I did not change the card. I usually do, but I do not have an excuse of why I didn't on this particual week. I am just getting around to all of my data that I dumpped in my HD.

 

[R.P..R]

I think, I know what it is now. but not a 100% sure. I could have another app running in the background. As I recall this happened when I had Go App open while running Litchi and ended up with a burly footage at the center. On this mission, I could have forgotten to exit Go after I did a compass calibration.

 

[Outta Control]

So here are my thoughts about figuring out the issue you brought up. So take this as constructive criticism.

I don't have an answer for you. Not because I don't know the problem but there is not enough true information to go by.

The images are surely different from one another, from within one week.

Surely, there should not be such as significant difference from each other.

The biggest problem you have is you do not have accurate on-site information.

Information such as "on-site" weather, wind, and environmental conditions.

You presented a historical data of a general weather condition from a weather station within a specified region, which always gives a vague and averaged results.

Did you have an portable weather station to record the actual conditions on site? It would have been beneficial to get actual on-site records not only to help diagnose an aircraft or image problem but it is recorded documentation of the conditions you were working with.

This is nothing new. A proficient sUAS Thermographer commercial remote pilot, should be recording these types of information to eliminate and validate the captured data.

 

[Outta Control]

I think, I know what it is now. but not a 100% sure. I could have another app running in the background. As I recall this happened when I had Go App open while running Litchi and ended up with a burly footage at the center. On this mission, I could have forgotten to exit Go after I did a compass calibration.

Possibly, don't know.

Which is why we have check list and recorded documentations.

I do hope it is as simple as that, and moving forward to make sure the data is accurate.

 

[R.P..R]

The biggest problem you have is you do not have accurate on-site information.

You are absolutely right, my life is governed by apps, and forgetting to use simple, yet vital basic apparatus, due to getting things done faster, and the end results are unusable data.

Thanks, a bunch.

 

[Meta4]

Temp, humidity, SD card etc are unlikely to have had any effect
I'd think the first thing to check would be the quality of the individual images for the first shoot.
Are they clear and sharply focused?
How do the rooftop a/c units near the top of the composite look in the individual shots?

think, I know what it is now. but not a 100% sure. I could have another app running in the background. As I recall this happened when I had Go App open while running Litchi and ended up with a burly footage at the center. On this mission, I could have forgotten to exit Go after I did a compass calibration.

Was there any particular reason to be calibrating the compass?
It should be completely unnecessary.
You used Litchi for mapping? How do you ensure overlaps? What overlap % was used?

 

[dougcjohn]

I’d look more toward images sharpness; The 12 Oly is a great sharp lens... you’re meaning the Zuiko 12mm correct? If assuming focus was Infinity on manual, 2-3 f minimal above f2; then sharpness should be max obtainable in the Zuiko’s sweet spot.

May be an oddity, and depends how quickly you launched for this next possible. The 12mm Zuiko Oly has several internal glass lenses. Here in IA, we can be very humid & hot... run AC in route to site, if camera cools and popped out into heat, you get a little light fog on or in lens.

Same in Winter with extreme cold... removing lens from warm (hot) car to cold... I normally give camera & lens time to adjust to environment.

Even more important to prevent fog or moisture going from cold to warm environment... I normally put camera asmb in case and leave sit to adjust temp more slowly if extreme difference.

 

[dougcjohn]

One other thought... I’ve noticed the X5 start up differently from 1 time to the next... an oddity I think has been discussed in several threads. The X5Raw is know to not only produce a better EV color range, it’s known to be much more stable for photos on SD card.

I can mount the X5, then X5R and easily see the difference.
I don’t recall which you’re using... just a thought.

 

[R.Perry]

Interesting thread. I had one of my mappings distort about thirty acres of a 160 acre mapping. All setting were the same as previous mapping. Flying the I2 with X5 and 15mm lens set at 90 degrees. Using Drone Deploy and no one could figure out why. I do this same mapping every two weeks, haven't had a problem since.

 

[dougcjohn]

Interesting thread. I had one of my mappings distort about thirty acres of a 160 acre mapping. All setting were the same as previous mapping. Flying the I2 with X5 and 15mm lens set at 90 degrees. Using Drone Deploy and no one could figure out why. I do this same mapping every two weeks, haven't had a problem since.

That is interesting... distort as in out of focus or image actually distorted, garbled?
Was problem in the initial shots, somewhere middle or last series of shots?

BTW, just to separate the two... the I1 uses the X5 & X5R, the I2 uses the X5S.
The X5S isn't known for any problems similar to the X5.

 

[R.Perry]

That is interesting... distort as in out of focus or image actually distorted, garbled?
Was problem in the initial shots, somewhere middle or last series of shots?

BTW, just to separate the two... the I1 uses the X5 & X5R, the I2 uses the X5S.
The X5S isn't known for any problems similar to the X5.

The first five or six rows were totally distorted, and obviously didn't stitch properly. DD couldn't give me any explanation as to what happened. I didn't have any other programs running in the background. Actually it was everything up until the first battery change as I remember.

I also had another mapping that just failed to capture two rows, but that was an easy fix.

 

[dougcjohn]

The first five or six rows were totally distorted, and obviously didn't stitch properly. DD couldn't give me any explanation as to what happened. I didn't have any other programs running in the background. Actually it was everything up until the first battery change as I remember.

I also had another mapping that just failed to capture two rows, but that was an easy fix.

That is odd, and sorry to hear that... hope it doesn't repeat.
Taking a "guess" based on information...
You mention "until first battery change", that might indicate electronic circuits that didn't initialize correctly or program (FW / Flight Code) didn't load correctly. New battery, new restart & refresh of circuit initialization.
We're still dealing with micro circuits and chips that do have a failure rate... very minimal but present, and that can result with an improper initialization... if 1-16 bits initialized incorrectly, that could cause a ripple effect in the executed code downstream as corrupted logic.

If that was the situation, then I'd watch the initialization over time. If it repeated or repeated with any sporadic frequency, it'd be a possible indicator that a circuit board was probably failing or flawed.

 

[R.P..R]

Temp, humidity, SD card etc are unlikely to have had any effect
I'd think the first thing to check would be the quality of the individual images for the first shoot.
Are they clear and sharply focused?
How do the rooftop a/c units near the top of the composite look in the individual shots?

Was there any particular reason to be calibrating the compass?
It should be completely unnecessary.
You used Litchi for mapping? How do you ensure overlaps? What overlap % was used?

Focus is set to infinity and verified by my Peak Focus Threshold (I set mine to high)

I calibrate if I transfer from on field to another specially during the early stages of the build.

I’m not using Litchi for mapping, my overlap is pretty consistent not below 60% and not higher than 70% depending on the output.

 

[R.P..R]

Sorry for the late reply, gang. But this certainly is mind boggling, and glad to know that Perry also ran across this problem.

I also posted a thread in DD and it seems like I am not the only one, and this phenomenon has no real explanations.

The rest of the data that I collected for 2wks are great. I will be back to work by the end of next week, and I’m hoping this problem, whatever it is will not return to haunt me.

Aloha, y’all
-Akoni

 

[R.Perry]

I was going to post the corrupted mapping, but Multivista pulled it. I guess we don't live in a perfect world.

 

[Meta4]

Focus is set to infinity and verified by my Peak Focus Threshold (I set mine to high)

Were the individual images well focused or blurry? - was the problem with the image capture or the stitching?

I calibrate if I transfer from on field to another specially during the early stages of the build.

Completely unnecessary and a waste of time.
Compass calibration is unrelated to geographic location, moving to another area makes no difference.