Research on Environment Intelligent Perception with Depth Camera

Environmental intelligent perception technology is among the key technologies for intelligent mobile robots. This paper proposes an environment intelligent perception scheme based on depth cameras to address mobile robot environments intelligent perception challenges. Firstly, the system initiates with direct filtering and point cloud gridding filtering to eliminate noise interference. Secondly, we employ Euclidean clustering extraction to identify non-ground objects and convex hull method to get their location and size. Finally, for 3D scene reconstruction, utilizing Scale-Invariant Feature Transform (SIFT) operator to extract features and match adjacent intensity images to find the corresponding 3D point set. The Iterative Closest Point (ICP) Algorithm is used to compute camera motion matrices to alignment adjacent point clouds.