OGIL : An Improved ORB Algorithm Based on GMS, Information Entropy and LC

The ORB feature extraction and matching algorithm is in commonly used. It is based on single-channel grayscale images for feature point extraction and matching, which results in a huge loss of information to cause inaccurate feature extraction. To solve this problem, we propose OGIL (Improved ORB algorithm based on GMS, information entropy and LC), the OGIL algorithm based on the ORB algorithm, which clearly outperforms the original algorithm. The OGIL understands RGB images as a sequential superposition of three independent layers and uses the information entropy metric of RGB images to determine the gray shift weights based on the information entropy of the three layers, which fuses into a grayscale image that most effectively retains the information contained in the image. Secondly, the LC (Luminance Contrast) saliency detection algorithm based on gray value is introduced to filter out redundant information, extract feature points quickly, and speed up the operation speed. To figure out the problem of mismatching in ORB feature extraction, the OGIL brings in a GMS (Grid-based Motion Statistics) algorithm. The comparison results on the dataset images show that the OGIL algorithm is significantly better.