Hotspot Effects in Micro Copter Data

A shots from last year vegetation point cloud mapping in Brandenburg. Strong Hotspot-Effect visible with easily visible shadow projection of the quadrocopter shape on the tree crown texture. Nice example of this direct back-reflection effect on vegetation (crown surface in this example), although you can also clearly see the same effect on the grassland vegetation in the neighborhood of the tree crown.



Punktwolkenmodelle für die Archäologie

Die unterschiedlichen Datensätze der Befliegungen für die Grabungen auf dem Alten Gleisberg und bei Kuckenburg in 2017 sind nun vollständig an das Seminar für Ur- und Frühgeschichtliche Archäologie übergeben worden. Die abgeleiteten Datenlayer umfassen jeweils Punktwolkenmodelle im LAS Format, DSM (Digitales Oberflächenmodell ohne Normalisierung) und jeweils das Orthomosaik mit Georeferenzierung von der Befliegung. Alle Daten wurden nach UTM32 WGS84 georeferenziert. Insgesamt wurden Daten im Umfang von über 100 GB verarbeitet. Die Befliegungen fanden jeweils in drei verschiedenen Flughöhen statt und liefern nun nach der Berechnung der Punktwolke eine cm-genaue Darstellung der Grabungsflächen bzw. deren 3D-Volumenwerte.

Die präzise Verarbeitung und Georeferenzierung der Copter Daten von diesen Befliegungen ist durchaus aufwändig. Jedoch methodisch eher unproblematisch schritt. Schwieriger wird die präzise Punktwolkenfusion aus verschiedenen Grabungen werden. Hier ist die absolute Georeferenzierungsgenauigkeit entscheidend.



Abb. 1: Grabungsareal der Kuckenburg-Befliegung vom September 2017 (Tiefbefliegung mit der Phantom 4 Pro, Punktdatenmodell mit 70 Mio Punkten abgeleitet aus einem Bilddatenpool von 250 Fotos).



Abb. 2: Grabungsareal der Kuckenburg-Befliegung – Hochbefliegung.



Abb. 3. Befliegung am alten Gleisberg – in dieser Abbildung:  das Punktmodell der Hochbefliegung vom August 2017.


Abb. 4. Befliegung am Alten Gleisberg – Detailbefliegung des Grabungsareals.

Absolut Accuracy of the Phantom 4 Pro Ortho Mosaic

Some results from DGPS point measurements with a Stonex S9III GNSS DGPS receiver indicate an offset of 1.5-3 m to the Phantom 4 Pro ortho mosaic images. The relative accuracy is however much better. The Phantom 4 Pro derived ortho mosaic (without reference points) has a standard deviation of 0.31 to 0.36 m.

When comparing the overall positional accuracy of the reference points its surprising that the UTM coordinates that come directly from the Phantom 4 Pro are not far off: I measured 12 DGPS points and found that the offset is between 150 cm and 290cm (compare with Fig. 6).

RMSE between the Stonex S9III DGPS (SAPOS) measured points and the DJI Phantom 4 pro POI mode ortho image mosaic measured points is 1,98 m (stdev 0,31 m) and the RMSE between the Stonex Points and P4P NADIR mode ortho image mosaic measured points is 2,95 m (stdev 0,36 m). The difference is clearly also depending on overall atmosphere conditions and Stonex DGPS accuracy. 




Fig. Comparison of measured DGPS Stonex GNSS points with in-Copter derived GPS geometry in POI ortho mosaic and with Nadir ortho mosaic processed in Agisoft in September 2017.

Phantom 4 Pro – lens flare issues

Some more tests with the Phantom 4 Pro: one major “issue” with this system for stills / photography is imo clearly the flare sensitivity of the lens/sensor combination. The into the sun shots just capture a lot of flare and the sun star is very much Samyang 14mm like. This is an issue and it reduces the image acuity for 3D point cloud modeling for object measurements as well as for the landscape panorama approach.

The against the light shots look very much like the Samyang/Rokinon  14mm lens character. The sunstar has light rays that increase in size from the center and these rays cover the full frame of the 1inch sensor when the illumination comes from one of the outer edges. It clearly degrades the full image when the lens is stopped down to f.e. f8 but gets better when the lens is driven wide open. There seem to be some diffraction effects at work that are clearly not so great. You can drive the camera to generate sharp sun stars, but as usual here the characteristics of this star make a difference to some of us.  The Samyang like star is not so well received. The linked flare issue is much more a problem because it is hard to avoid when you have the sun in your frame. For some scenarios a lens hood may be useful but its only functional for those shooting scenarios where the sun is not within your FoV.

To be honest I am bit puzzled … this is the first affordable 1inch sensor platform with nice resolution and better DR, but – the flare issue is likely killing some of the potential ideas that you might develop with this machine.

Some examples:

smh-lim15-20170214-d4c5abaf23a147c434e47d2a0105ac83-0281 smh-lim15-20170214-d4c5abaf23a147c434e47d2a0105ac83-0300 smh-lim15-20170214-d4c5abaf23a147c434e47d2a0105ac83-0304

@f4 things get a bit better, but with this image below also the light levels have changed:

smh-lim15-20170214-d4c5abaf23a147c434e47d2a0105ac83-0366 smh-lim15-20170214-d4c5abaf23a147c434e47d2a0105ac83-0424 smh-lim15-20170214-d4c5abaf23a147c434e47d2a0105ac83-0447

Comparison shot done with the Phantom 3A, clearly shows that the lens/sensor combi wont give the same amount of flare here (again different light level and different sun illumination angle):


mtk, Sören

Phantom 4 Pro arrived

The Phantom 4 Pro arrived – first integration tests started and I am testing the image quality of the 1-Inch sensor – at least from my perspective the most important part of the new P4 series. Especially comparing dynamic range and sharpness to the tiny P3A sensor/lens combo and systems will be interesting (which is similar to the original P4 sensor).