Closed elboyran closed 7 years ago
The predict
method of the ImageCategoryClassifier
can be used to predict the labels of an image datastore:
[labelIdx,score] = predict(categoryClassifier,imds)
but geenrating tiles for ALL pixels makes more than 30 million images- too much storage!
Classifying one pixel using s tile around it (100x100px = 80x80m.) takes about 350 ms. But segmenting the whole image of more than 30 million (6223 x 4872 = 30 318 456) pixels would take in current implementation 12 days. Therefore, I've processed every 5th pixel in both dimensions , that took about 11 hours. Interpolation to the missing pixels should be a separate issue.
[x] Decide on how to do it- via datastore or not. Check of the method
predict
can work on single image tile or on datastore.[x] Create function and script for predicting the label of each pixel based on tile (100x100px = 80x80m) around it using a learned classifier.
[x] Segment the whole image using codes: 1- Slum, 2- NonBuiltUp and 3 BuiltUp