Contrast Limited Adaptive Histogram Equalization for an Advanced Stereo Visual SLAM System
Teams: Beijing Jiaotong University
Writers: Wenkao Yang; Xiangwei Zhai
Visual SLAM (Simultaneous Location and Mapping) has gained popularity in the field of autonomous driving, virtual reality and augmented reality recently. However, the popular feature-based visual SLAM systems are usually sensitive to brightness and translation. To make our system more robust, we present a real-time robust feature-based visual SLAM system with low drift and high precision. We contribute an image pre-processing module together with the CLAHE (Contrast Limited Adaptive Histogram Equalization) algorithm to locally enhance image contrast and get more feature details. Experimental evaluation shows that our system is significantly better than ORB SLAM2.