So recently I have been trying to detect objects by training my own Cascade Classifier. Specifically, I wish to be able to detect drones in the sky. However, my dataset only includes drones of specific colors and I want it to detect drones of all colors.
I thought about maybe running Edge Detection on my training images and using those images to train the Classifier. Is this a good approach? Any suggestions on how to train a Cascade Classifier to be color invariant?
