Hi all,

looks that with the spread of processing methods and technology,

we can make things that were available only to national security agencies, once upon a time.

Take a look at the maps of the flight made May 18th:

one day to fly, one day to upload, one day for processing, (one day for grill party), one day to download.



Few of you might remember the 3km long flight with EasyStar (a single line), stitched with Microsoft ICE.

The map above is 2.2x2.2km. It took 746 photos to make this map.

Funny fact: if the aircraft would fly straight, the surface covered would be >16km2 instead of just 4.5km2 (null side overlap).

The overlap used was 70% side, 80% along.

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones


  • T3
    This is very impressive!
    Thanks for sharing.
  • T3

    "24 hours.... I think that it is or will be in less than a year."

    No, because in practice uploading 1GB of photos and downloading 1GB of results (from remote country, through 20 gateways and across 2-3 frontiers) will not take 2h. The processing on cloud of computers may or may not bring the price to the sky (you need order of magnitude of acceleration here). Maintaing your personal cloud service might shave 24h from that equation, but I highly doubt it is worth it.

    Downloading and time-synchronising the photos takes time. It may be one hour (get all things from the car, get the SD card, exclude blurred photos, verify all data from the logs), but making it all in one day is unrealistic unless this is national security agency (if it is commercial and not round-clock virtual entertainment squad, 24h requirement means you HAVE TO do it in 8h between 9am and 5pm).

    Making it in less than a week is a world class solution given the price levels involved and is a blitz compared to a few months of delay from satellite or a few weeks of delay (typically) from full scale airplane mapping company. The 4-5 days delay is real and truly involving humans, one of them 'worked' on sunday downloading the data.

    I don't think many countries have a processing chain and operators for large scale mapping from the UAV. Japan certainly hasn't had a few months ago.

  • Thanks for the response, I presume this was done with the "cloud" version of their service. Have you tested the "lite" version for UAV?
  • I would like to produce such orthos in less then 24 hours. Is it possible? I think that it is or will be in less than a year. The only problem is high quality seamline creation. Especially in urban area (buildings). True ortho is a good option but edge 3D geometry extraction is still far away from adequate for automated approach.
  • T3

    It is funny fact, but once you are in G class airspace, maintaining higher altitude is safer. At 300m AGL, Pteryx UAV is identifiable and can be maneuvered in assisted mode (full stabilisation) from 1500-1800m. Not to mention it is well into RC range then and it is almost never obstructed by obstacles or buildings (some 20deg over horison).

    In that flight we had verified air traffic plans and had the radio freq allowing contact with local paraglider enthusiasts. Plus 2 spotters and a parachute, flown 60% of the time over fields and got 100% endurance margin.

    Things are easy enough to run a mapping business only because PIX4D turned out to be so reliable.

  • Moderator
    @ Jack some people are cleared to 1000' in the UK now, ah the joy of proper regs!

    Great work Chris, you are making a tricky thing look easy.
  • T3

    Generally speaking, all stitching programs end somewhere around bitmap sizes around 10x10kpix. This alone prevents maps larger than 1x1km at 10cm/pixel from MS ICE. Unfortunately, those maps are completely not orthorectified and the results are not georeferenced. They are good for showcase by the amateurs.

    PIX4D results have commercial value as they are true maps that are standalone and better positioned than any existing free maps, because of being made from low flying UAV there is comparable resolution and more contrast than from the best full scale airplanes using linear scanners.

    Because of the processing time, stitching that map with ICE would require a week of hard work improving, bending the results and would require measuring additional reference points on the ground. Some maps, however, by luck, stitch well and accurately with any stitcher: only flat areas, typically less than 100 photos and a lot of luck - only 25% of the cases work - but when they work a little faster than true mapping services - but you cannot make business out of that since usually flying several times is not an option. Usually stitching software tends to struggle over build-up areas or when there is too much contrast, but they sometimes can save the day when there is minimal contrast (snow).

  • The thrill of flying 700ft over people.  There's so much we could do if our country wasn't lawsuit crazy & parts were more reliable.
  • Hello and nice job.


    So what are your thoughts on pix4d so far? How have your results with MS ICE been generally speaking?



This reply was deleted.