DJI drone flight log viewer
flightreader.com
flightreader.com
We are currently developing a open source flight log database & plotting tool for ardupilot logs (graphs, maps, stats, images using PostgreSQL, FastAPI, Next.JS) covered by the german prototype fund (sorry german only link) : https://prototypefund.de/project/searchwing/
Source is here: https://gitlab.com/searchwing/development/groundstation/flig...
Maybe this is also helpful for others.
I wonder if there are more such opportunities…
Last time I recorded video with the drone I am suprised there is a SRT file bundled. Turns out that is the telemetry data in plain text recorded at 10 Hz. It should be trivial to write a script that transform it into KML or some GIS-friendly format.
And I am sure I do not need any shareware for this...
I built an online flight log analyzer about 10 years ago ( https://www.youtube.com/watch?v=8XBCBC2oJ70 , https://github.com/foobarbecue/afterflight ) . When DJI hit the scene I used some a DJI log decoder someone else wrote to get my Phantom logs into the system, but never integrated the converter with the website.
Good ideas are a dime a dozen. It's the execution and the follow-through that matters.
Somehow, in my mind, I'd like to find a niche where I can start building something (software-related). Something that is of use, and which solves a problem people are willing to pay for. Not a "get rich in two months"-scheme, but something sustainable. I have no problem with commitments, but it’s the problem to solve that I'm currently lacking.
Apparently, this tool does it. I think there are not many tools like that on the market (I might be wrong).
And https://airdata.com which has a free tier. It can also sync automatically with DJI and third party apps (I use Litchi).
> View over 200 data points collected every 1/10 of a second during the flight
Is there some significance to these numbers (20 seconds and 30 seconds of data) or is that just an attempt at "big numbers sound good"?
The generic data channels number easily less than 50 - time, X,Y,Z, axis velocity, battery life, yaw, tilt, pitch, etc.
I dare say some of the special domain applications generate more channels of data from mounted instrumentation - all logged to a common DJI family log format - for extraction on landing or broadcast during flight etc.
The 200 channel Vs 300 channel thing is likely just poor copywriting from looking at max channels in two different application areas.
On top of that come RTK and ADS-B receiver data, and the really large ones (Matrice, Agras) add more engines, payload status (weight, flow rate), multi-battery health, RF quality and other data channels.
To be honest, it's amazing just how many features modern drones carry while still allowing sometimes well over half an hour of flight duration.
PX4 also has lots of "topics" logged at around 1/10, e.g. hover thrust estimate every 100ms:
https://github.com/PX4/PX4-Autopilot/blob/main/src/modules/l...
Running on a single-board computer with an applications processor like a Raspberry Pi or BeagleBone, with multiple cores at 1 GHz+, gigabytes of DDR3, with on-chip SDXC drivers? Logging loads are irrelevant. The only thing to watch out for is that if you want to use the logs to figure out why you had a catastrophic crash, the OS may have write-buffered a lot of data in RAM.
If you're running on a crossover MCU, like a Cortex M7, with a core or maybe two at a couple hundred MHz? Think about it what you log.
Running on a Cortex M3 with a few dozen MHz and a few dozen kB of RAM? Yeah, appending to and rolling over a whole 512-byte page in RAM and finding the time to push it over the SPI peripheral is going to be tough, you may want to use a binary protocol so you're not burning unnecessary cycles in itoa() and you get more data per byte of buffer.
The DJI drones discussed here have big Cortex-A53 SOCs that are already logging 4k video at 30 FPS to that SD card; the log file is only noticeable as a distraction from that task. I think the crossovers are the sweet spot, though, because a 'fast' clock or interrupt on that Linux board is slow in comparison to the more tightly integrated RTOSes.
When I was working on an aerial georeferencing project[1] I found that the on-device flight logs didn’t contain all the info in the serverside logs (eg heading, gimbal orientation, GPS coords, altitude)
- There are no mandatory monthly/yearly fees
- Flight logs are not uploaded to the Internet
- Many more data points are exposed
- The generated logs/reports are customizable
Also I think this software is Windows only.
Yeah, most people would call that Windows only
But most of the world does of course run Windows. So I’m not in the main group of people this software was for. That’s fine.
Thanks for the recommendations, that's why we have HN comments, to help each other :-)
[1] https://lba-openuav.de/onlinekurs/lehrmaterial/wichtige-defi...
In any case, flying with a spotter is always a good idea, particularly with smaller drones or in areas where other people are present that may interrupt you. A spotter can take care of dealing with that.
[1] https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A...