Improved Gamma Correction for Visual SLAM in Low-Light Scenes

Abstract

Visual SLAM in low-light scenes is important for mobile robots to use in complex industrial scenarios. In this paper, based on the RGB-D SLAM model, an improved Gamma adaptive correction algorithm is added. According to the gray average of the acquired image, the Gamma value is adjusted to achieve adaptive correction. The processed image is substituted into the pose estimation based on the ORB feature, then after key frame extraction and pose graph optimization, and finally the positioning and mapping work based on the RGB-D camera is completed. The experiments show that in different low-light environments, this method can effectively improve the accuracy and stability of image matching compared with other methods. It can be applied to the positioning and mapping of mobile robots in low-light scenes in the future.