Octree map construction method based on closed-loop trajectory

Abstract

The invention discloses a closed-loop trajectory-based octree map construction method, which comprises the following steps of: 1) performing pose estimation and optimization according to two adjacent frames of images, and determining whether a key frame is generated or not; 2) generating map points according to the newly added key frame; 3) carrying out closed-loop detection and closed-loop correction; 4) storing the image file and the track file; 5) reading the track file to obtain a timestamp and a pose corresponding to the key frame; 6) reading the image file to generate a three-dimensional point cloud; 7) filtering the three-dimensional point cloud; 8) converting the three-dimensional point cloud into an octree map; and 9) dyeing the newly generated octree map according to the height. According to the method, the pose estimation accumulative error is reduced through closed-loop detection and closed-loop correction, so that the established map is more consistent with the reality, and the mapping accuracy is improved; and direct filtering and voxel filtering are carried out on the three-dimensional point cloud, so that the data volume of the three-dimensional point cloud is reduced, the conversion time of the octree map is shortened, and the real-time performance of the algorithm is ensured.

Description

technical field [0001] The invention relates to the technical field of three-dimensional map construction, in particular to an octree map construction method. Background technique [0002] Three-dimensional maps can describe the three-dimensional space environment in more detail, and have advantages over two-dimensional maps in providing positioning, navigation, obstacle avoidance and other service functions for UAVs and robots. The methods of constructing a three-dimensional map in the prior art include constructing a three-dimensional point cloud map and constructing an octree map. However, because the 3D point cloud map is sparse, it is impossible to accurately and clearly see the structure of the original 3D scene, distinguish feasible areas and obstacles, so it cannot be directly used for navigation, while the octree map can overcome the shortcomings of the 3D point cloud map. [0003] However, when constructing an octree map in the prior art, due to the accumulated er...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the purpose of the present invention is to provide a method for constructing an octree map based on a closed-loop trajectory, so as to solve the problems caused by accumulative errors in pose estimation based on adjacent frame images when constructing an octree map in the prior art. The problems of uneven reference plane and inconsistency between the angle and the actual situation appear in the established map, which lead to the technical problem of reducing the accuracy of the built map; and solve the problem that the existing technology takes a long time to convert the 3D point cloud into an octree map and it is difficult to guarantee Algorithm real-time technical problems

Images