V-SLAM pose estimation method based on Kinect camera

Abstract

The invention discloses a V-SLAM pose estimation method based on a Kinect camera, and the method comprises the following steps: 1) collecting image information through Kinect, obtaining a color imageand a depth image, carrying out the registration of the color image and the depth image, and obtaining a three-dimensional point cloud; 2) performing feature point detection on the acquired color image by adopting an improved FAST, and calculating a descriptor; 3) performing feature point matching on the to-be-matched image and the reference image by using an FLANN matching method to obtain a feature point matching pair; 4) eliminating mismatching of the matching pairs in the step 3); and 5) solving the pose of the accurate matching pair in the step 4). According to the method, an adaptive threshold value is designed according to the gray scale of the image to change the robustness of the algorithm, and meanwhile, dense feature points are screened out by using a non-maximum suppression method after feature point extraction, so that the problem of non-uniform distribution of the feature points is improved. The method has a good pose estimation effect and can be widely applied to the aspects of robots, VR, unmanned aerial vehicles and the like.

Description

technical field [0001] The invention relates to computer vision technology, in particular to a Kinect camera-based V-SLAM pose estimation method. Background technique [0002] SLAM (Simultaneous Localization and Mapping) technology is widely used in the fields of mobile robots, unmanned driving and drones. The intelligence level of the robot. With the continuous development of visual sensors, visual SLAM has gradually emerged. Among them, the Kinect visual sensor is widely used in mobile robots because of its rich environmental information, high precision and direct measurement of depth. [0003] V-SLAM is mainly divided into two parts: the front-end visual odometry part and the back-end optimization part. The function of visual odometry is to estimate the motion of the camera between two adjacent frames and determine the pose of the camera. Estimating the pose of the camera with feature points has always been the mainstream method of visual odometry, and the distance betw...

AI Technical Summary

Problems solved by technology

The function of visual odometry is to estimate the motion of the camera between two adjacent frames and determine the pose of the camera. Estimating the pose of the camera with feature points has always been the mainstream method of visual odometry, and the distance between two adjacent frames of the camera is Low matching efficiency and low precision will affect the camera pose estimation, which will directly affect the accuracy and performance of robot positioning and map construction.

Images