In my opencv project I try to compute sift features, I detect keypoints and compute descriptors. What I know it should contains 4 parameter: X, Y, Scale and Orientation. The opencv keypoint structure has pt (X, Y coordination) and angle (Orientation), but I could not understand where is the Scale parameter! can you explain me about this parameter?
the scale use for building pyramid which mean you can choose how much change of the size(scale) of the objects can be change. for example if the object(s) moving in distance you should have more levels of pyramid in order to have better recognition.