Feature-point extraction based on an improved SIFT algorithm

SIFT (Scale Invariant Feature Transform) is an algorithm that extracts the feature data from an input image. It comprises robust characteristics that prevent image transformations such as the image size and rotation in the matching of feature points. However, it is disadvantageous because it is difficult to extract the feature points if the brightness distribution of the image or the image itself is concentrated in a specific range. In this paper, we propose an improved SIFT algorithm for disturbances such as sunflare by applying the CLAHE (Contrast Limited Adaptive Histogram Equalization), a histogram-equalization method, as a preprocessing SIFT method. For this paper, we implemented the algorithm using Visual Studio 2013, which can use the C ++ language, and implemented the histogram analyses in the MATLAB environment.