Pedestrian detection and tracking for mobile service robots
Shengsheng Han, Ping Ye, Baohua Zhu, Hanxu Sun · 2017
Pedestrian detection and tracking is an important research direction in the field of computer vision. With the smart home applications developing, mobile service robots have also drawn more and more attention. But the complex dynamic background and mobile pedestrians have brought great challenges to such problems. In addition, we found that it is always not enough only to know the information of pedestrian's location in the picture for mobile service robots in practical applications, the robots often need to know the spatial relationship between their locations and the corresponding tracked pedestrian's location, in order to move there to provide the appropriate services. However, the general detection and tracking algorithm only relied on the picture RGB or gray information can't do this. This paper presents a pedestrian detection and tracking method based on depth information and visual SLAM, which aims to solve the problem of pedestrian detection and tracking in the field of mobile service robots. We are clever to use tracking to the corresponding point cloud to achieve the purpose of tracking pedestrians. And with the help of camera motion estimation by visual SLAM, it can create a global position relationship for us.