Automatic Landing Flyaway

[slup]

@sar104 ... one thing puzzles me looking at the TXT log on Airdata.com & one of the pics in the log that have GPS position. That GPS position doesn't match the likely photo position in reality.

What the log says match reality, both position & yaw angle wise to get that photo ... so from where comes that GPS position under the pic?

The log:



Photo from Airdata + positions in GE:

 

[sar104]

@sar104 ... one thing puzzles me looking at the TXT log on Airdata.com & one of the pics in the log that have GPS position. That GPS position doesn't match the likely photo position in reality.

What the log says match reality, both position & yaw angle wise to get that photo ... so from where comes that GPS position under the pic?

The log:

View attachment 96658

Photo from Airdata + positions in GE:

View attachment 96660

No idea why AirData got that wrong - the position data are correct in the txt log. There are no position EXIF data in the file from the log, so it's determined from the txt data.

 

[sar104]

AirData gets the photo location correct on the media map:

 

[slup]

No idea why AirData got that wrong - the position data are correct in the txt log. There are no position EXIF data in the file from the log, so it's determined from the txt data.

When I saw that my thought was if it's something really off position wise ... but can't find anything, it's also as you say correct in the media map.

 

[BudWalker]

Did you notice that the offset time and tick data in the DAT file contain errors?

View attachment 96654

The tick# you see produced by DatCon is a computed value. The "real" tick# looks like a saw tooth wave form. It's represented by 64 bit integer and therefore has to roll over when it reaches 4294967295. That's why the computed tick#s seen in the .csv can be greater than 4294967295. It's a complicated business because, in addition, records can occur out of sequence and DatCon has to reorder the records when this happens.

In this .DAT there was a real tick# roll over during the 20 minute gap and DatCon just couldn't manage to figure it out. It came close though.

BTW, a few of the records have their own tick# but with the label cnt. And they have a much lower frequency and don't have to roll over in a typical flight. Might be easier to use than the GPS:time Here is the VIO:cnt which is at 20 Hz.