Thanks for the reply.
Just so we’re on the same page I’ve included a representative snippet from the sample RGB image that Micasense have on their website as sample data from the Sequoia and another example that Parrot shared. They’re both rendered as 1:1 in terms of pixel size.
The jpeg compression setting that Sequoia are using of 80 means that so much detail is lost and you end up with an oil painting effect.
Are customers really happy enough with this sort of RGB image quality?
As I mentioned we’ve been using a USB camera from e-con systems which has the same image sensor size as the Sequoia RGB sensor and the same sort of 5mm lens and the image quality is much better, it doesn’t have this oil painting effect anymore.
It used to have a similar effect before, or though not quite as bad/harsh. The difference was that their camera used to try and use a jpeg image quality of 95 but if it couldn’t keep up with the expected frame rate then it would dynamically lower the jpeg image quality. Now they use a fixed image quality of 95 and let the frame rate drop lower if need be from their default frame rate of 14fps for the 13 mega pixel size we’re using.
Given the limits of the CPU chosen by Parrot for the Sequoia would it be possible to configure the RGB camera to output a TIFF image instead of a JPEG image. The trade-off if we selected the TIFF option would be much larger file sizes on the SD card and longer write times to the SD which would potentially limit the capture frequency, but it would allow us to then subsequently choose our own JPEG compression level to use for transmitting the images from our customer to ourselves for processing.
I did take a look at the Sequoia API and didn’t see any such configuration option currently to have the RGB camera output a TIFF instead of a JPEG.