SAM classifier for remote sensing images. Python libraries required -Ratserio,OpenCV,Numpy,Tkinter,matplot. Allows user to click on a part of image to act as training data for a class. Threshold for classification angle < 0.1 radians else unclassified. Steps of execution the python file:
- Run the program.
- Browse and select the 4 band input image.( If inpput image is not 4 band it will give a error message saying " NOT A 4 BAND Image"
- Click on "Show TCC&FCC". Two windows will open, titled as Figure1 corresponding to TCC & Figure 2 corresponding to FCC.
- Double click on pixels which you want to consider as reference pixels( In FCC image i.e. Figure2)
- After selecting reference pixels close all windows including main window and wait for the classified image to pop up in some time.
- If you want to save classified image you can save it.
- After closing the classified image a new window will appear showing the TCC, FCC and Classified image at once.
- There is a default palatte loaded to classify, which is also incuded in folder named as "palette.jpg"