Vision positioning method in dynamic environment

Abstract

The invention discloses a stereoscopic vision positioning method capable of realizing accurate positioning in a dynamic environment. An environment image is acquired in real time by a binocular camera. Characteristic points in the image are extracted by using a characteristic extraction algorithm. Four images acquired by the binocular camera at two adjacent moments are stereoscopically matched. The three-dimensional information of the characteristic points is recovered by using a double-vision geometric principle. Matching accuracy is improved in a loop-back matching mode. The characteristic points on a dynamic object are removed by using a field flow method. A field flow method effect is improved by considering error causing factors and calculating a covariance matrix. Motion parameters of a robot are obtained by using a Gaussian Newton iteration method according to the position information of the characteristic points. Vision positioning accuracy is further improved by using an RANSIC algorithm. The whole process is continuously iterative, and the gesture and position of the robot are calculated in real time.

Description

Technical field: [0001] The invention relates to a mobile robot self-positioning technology related to image processing and machine vision. The technology can use visual information as an input to realize the mobile robot self-positioning in complex dynamic scenes. Background technique: [0002] In the field of intelligent robots, mobile robot positioning and navigation technology has always been a key research issue. The accuracy and speed of positioning directly affect the efficiency and accuracy of the robot to complete the given task. The previous positioning methods, such as encoders or wheel odometers, cannot overcome the positioning errors caused by nonlinear factors such as friction and sliding. The visual positioning methods that have emerged in recent years can solve nonlinear errors well, but the existing There are certain defects in the visual positioning method: [0003] 1: When there are areas with similar textures and a large number of similar repetitive str...

AI Technical Summary

Problems solved by technology

[0003] 1: When there are areas with similar textures and a large number of similar repetitive structures in the environment, it is easy to cause mis-matching of feature points, resulting in positioning errors

[0004] 2: When there are a large number of independently moving objects in the environment, the feature points located on the dynamic objects will interfere with the odometer results, resulting in a decrease in positioning accuracy

[0005] In order to make visual localization technology applicable in complex dynamic scenes, certain methods are needed to eliminate feature points located on dynamic objects. However, current algorithms, such as outlier measurement algorithms and image understanding algorithms, cannot simultaneously satisfy the Real-time and effectiveness requirements for positioning

Images