Line feature visual odometer method combining depth map inference

Abstract

The invention provides a line feature visual odometer method combining depth map inference. The method comprises the following steps: acquiring a color image and a depth image by using a depth camera;extracting 2D line features in the color image by using a line segment detection algorithm, then determining each rectangular region containing the line features, mapping each rectangular region intoa depth map, and performing depth inference on each rectangular region by using a pixel filtering method for a depth missing problem; sampling points in the 2D line features, combining with the depthmap back projection after depth deduction to obtain 3D points, finding out the optimal two points passing through the straight line by using a random sampling consistency method to fit a 3D line segment, and estimating the camera pose according to the matching relationship of the 3D line segment; and finally, constructing a Jacobian matrix containing a distance error and an angle error, and optimizing the pose by using a Gauss-Newton method. According to the method, the fitting reliability of the 3D line segment is improved, and angle error information is added, so that the accuracy of the optimized pose is further improved.

Description

Technical field [0001] The invention relates to a visual odometer inference method, in particular to a line feature visual odometer method combined with depth map inference, belonging to the field of robot vision SLAM (Simultaneous Localization and Mapping, synchronous positioning and map construction). Background technique [0002] The purpose of SLAM technology is to allow mobile robots that are in an unknown environment and are uncertain about their position to perform positioning and map construction at the same time. It is a key technology to realize robot autonomy. At present, with the development of modern computer technology and artificial intelligence, a new upsurge has emerged in the research of intelligent robots. For example, the use of autonomous mobile robots for dangerous space exploration, the use of drones for military reconnaissance, and underwater robots for marine environment detection, etc., and the implementation of the above applications is inseparable from...

AI Technical Summary

Problems solved by technology

[0004] The use of depth cameras in visual odometry will provide great convenience for the 3D estimation of features, but in the process of using depth cameras, some places in the depth map collected by the depth camera will have depth loss, which is likely to cause 2D features Unable to backproject to 3D features

Images