Tensorflow

[SphaeroX]

I just read through the readme, I think what you are looking for can be found here:

https://www.tensorflow.org/tutorials/keras/classification

Here, image recognition is trained on the basis of a dataset, i.e. it is possible to classify the 30x30 squares in a very "simple" way. What I don't understand yet is: what do you want to achieve with this?

Do you want to replace the squares that have few features with a white square so that the photogrammetry software does not have to calculate too much?