关键词: 弱光环境, 同步定位与地图构建, 图像增强, 多尺度高斯函数, 自适应校正

Abstract: Aiming at the localization failure and tracking loss of visual simultaneous localization and map building (SLAM) in complex environments such as weak lighted or even totally dark, a vision SLAM algorithm suitable for weak light environment is proposed based on ORB-SLAM2, to which, a new adaptive image enhancement algorithm is applied. Image brightness is adapted by means of correction factor , which can be dynamically adjusted according to illuminance component of input image extracted by multi-scale Gaussian function. Performance of the algorithm is tested on public dataset. Simulation results show that the algorithm can efficiently help feature matching in complex environments such as weak lighted or even totally dark, consequently, the robustness of ORB-SLAM is improved effectually. 

Key words: low light environment, simultaneous localization and mapping, image enhancement, multi-scale Gaussian function, adaptive correction