Research Of Fast Object Tracking Algorithm In 3D Point Cloud Environment

Moving object detection is a hot topic in computer vision and digital image processing.It is widely used in robot navigation,intelligent video surveillance,industrial inspection,aerospace and many other fields.With the rapid development of three-dimensional sensors such as Kinect and TOF cameras,the cost of acquiring point clouds has become cheaper,and the tracking of moving objects in three-dimensional environments has gradually become a research hotspot.In recent years,although researchers have proposed a large number of methods for target tracking in 3D scenes,there has not been a unified tracking framework so far.In addition,there are no specific reports that provide in-depth analysis of the influencing factors of these tracking methods for point clouds.This topic proposes a moving object tracking framework based on PCL(Point Cloud Library)in 3D environment.Through the Pass-Through filtering of the scene,the interested area where the target is located is obtained.The number of point clouds is reduced by sampling the area,the interest area including the moving target is segmented,the original moving target model is extracted,the model is filtered to reduce the error,and finally it is taken as the reference model that is applied to the adaptive KLD particle filter framework to achieve moving target tracking.The proposed target tracking framework is divided into four parts.In each part,the algorithm with better overall performance is selected.Finally,a fast PCL-based target tracking framework in the 3D environment is constructed.The tests were performed on a modular robotic arm sorting system and the results showed the effectiveness of the framework.