Isasues with Micasense Rededge-M data

[rider on the storm]

Hi There,

I just processed a set of Micasense RedEdge-M data, following the instructions on the manual.
When I now check the values in QGis, it appears that the values of band 4 (which is supposed to be the NIR-band), are much higher on bare soil than on vegetation. But it should be the opposite. When I check the single raw image of band 4, the values are much higher on vegetation than on bare soil, as it should be.

Appart from that, the processed mosaic appears very dark.

Did I do a mistake in the processing?

Thanks and kind regards
Andreas

[Alexey Pasumansky]

Hello Andreas,

Can you please check in PhotoScan that fourth band is NIR (you can see it in the Raster Transform dialog or in the Camera Calibration window)?

Latest version of PhotoScan is using central wavelength value according to MicaSense recommendations to sort the input bands on load. According to the central wavelength, the order would be the following: Blue, Green, Red, Red Edge, NIR.

[rider on the storm]

Hi Alexey,

thanks for your reply. Concernign the camera calibration panel, the fourth band is rededge. concerning MicaSense, the fourth band should be NIR -> https://support.micasense.com/hc/en-us/articles/115003704854-RedEdge-M-Manual-Specifications

Cheerio
Andreas

[Alexey Pasumansky]

Hello Andreas,

PhotoScan Pro creates output in accordance to the MicaSense Atlas output:
https://support.micasense.com/hc/en-us/articles/220233768-RedEdge-Sequoia-and-Atlas-GeoTIFF-Band-Order
The band order is defined by the central wavelength value, and not by the RigCameraIndex tag.

[rider on the storm]

Hi Alexey,

thanks for the info!
I just calculated an NDVI with band 5 as NIR and it looks ok. The darkness of the mosaic seems to stem from the reflectance calibration. When I process without it, it looks much brighter. But anyway - for further analysis the brightness doesnt matter much.

Thanks
Andreas

[Alexey Pasumansky]

Hello Andreas,

The brightness may be lower, due to the fact that 32768 value from 16-bit output corresponds to the 100% reflectance value. So in the most cases only the first half of the 16-bit interval contains the reflectance values.
Yet again, the link to MicaSense knowledge base regarding the output (since PhotoScan is following the same convention):
https://support.micasense.com/hc/en-us/articles/215460518-What-are-the-units-of-the-Atlas-GeoTIFF-output-