Abstract:Autonomous vehicles (AVs) play an important role in the next-generation intelligent transportation system (ITS), which requires AVs to have the ability of rapid decision-making and control. As a part of ITS, simultaneous localization and mapping (SLAM) technology is the basis for AVs to operate in a complex and unknown environment. In this article, a stereo vision odometry method in a dynamic environment is proposed, which can not only effectively overcome the influence of dynamic objects but also detect the position of dynamic objects. An optical flow filtering algorithm based on the quantitative histogram (QH) and the optical flow angle histogram (OFAH) of the feature points is proposed to obtain dynamic points. Furthermore, a multifeature fusion mechanism is used to perform binary segmentation and get the bounding box of dynamic object. Experiments show that the proposed method can improve the accuracy of pose estimation and detect moving objects in a dynamic environment.

Optical Flow-Based Stereo Visual Odometry With Dynamic Object Detection