A Loop Closure Detection Method Based on Point Cloud Fragment Matching Constraints and Trajectory Drift Optimization

Abstract

A loop detection method based on point cloud segment matching constraints and trajectory drift optimization, using real-time lidar odometer and mapping LOAM system, but the LOAM system will have a large drift after a long time of mapping, so a point cloud is proposed A Loop Closure Detection Method Optimized for Segment Matching Constraints and Trajectory Drift. By segmenting, describing, and matching the point cloud acquired by lidar, the loop-closing relationship is found. After the loop is found, the explicit loop-closing ELCH algorithm is used to adjust the loop, and an optimization algorithm is proposed to eliminate the local inconsistency in the point cloud map adjustment process. After optimization, the present invention proposes a pose prediction and compensation algorithm to predict and compensate the pose after loop closure. In this way, the effect of reducing the drift error can be maximized.

Description

technical field [0001] The invention relates to technical fields such as computer vision and machine learning, and in particular to a loop closure detection method based on point cloud segment matching constraints and trajectory drift optimization. Background technique [0002] In many robotic applications, such as aerial mapping, disaster relief, space exploration and other tasks, there is usually no pre-built map information. The emergence of SLAM system provides a good solution to this problem, it can give the robot the ability to explore the unknown environment. The robot perceives the environment and realizes simultaneous positioning and mapping in an unknown environment through SLAM technology. It is a variety of sensors that allow robots to perceive the outside world. Commonly used sensors include cameras, lidar, inertial sensors, GPS, etc. Most current SLAM systems use cameras as sensors. With the development of unmanned vehicles and unmanned aerial vehicles, peo...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to design a loop detection method based on point cloud segment matching constraints and trajectory drift optimization, to solve the problem of trajectory drift in the process of lidar mapping, and to enhance the robustness of lidar mapping

Images