Research Of Robot Localization Method Based On Corner Feature Matching And Artificial Landmark

Autonomous mobile robot has always been an important field of robot research.Robot localization technology is the research focus of mobile robot.Based on the road sign matching location,some special structural objects are known in the environment as the road sign,and the mobile robot realizes the autonomous localization by the perception and recognition of the road sign.The artificial road signs can be designed according to actual needs,such as designed form,color and even location.Therefore,it is flexible,stable and accurate in positioning.A reasonable road sign design combined with corresponding recognition algorithm can play a key role in mobile robot localization.In traditional mobile robot localization methods,it is usually difficult to balance the accuracy and real-time of the positioning system.Although some of the positioning navigation methods have a good effect,they usually use high sensors or complex recognition algorithms,so it is difficult to use the control cost to apply to the home robot.In this paper,an artificial landmark based on corner points is designed and a landmark recognition algorithm based on corner feature matching is designed.In view of the problems mentioned above,this paper designs a new design of artificial Road and combines the Microsoft Kinect depth camera sensor to form a set of mobile robot positioning system.Experiments show that the recognition rate of landmark recognition algorithm is high and the recognition time is short.It can satisfy the positioning requirement of mobile robot.The main contents of this paper include the following:(1)In this paper,a mobile robot location system based on artificial landmark is designed.The traditional road sign design scheme is complex in design and low in recognition rate,so its practicality is difficult to meet the needs of mobile robot location and navigation.In this paper,an artificial landmark based on corner feature is designed.The road sign is simple and easy to identify.By adjusting the distribution of corners,it can also encode and increase the connotation information of road signs.The system collects the image and depth information of the surrounding environment through the Kinect depth sensor,and identifies the artificial road sign,and uses the depth information to calculate the distance data of the way out.Finally,according to the triangle positioning principle,the current position coordinates are calculated and the autonomous positioning is completed.(2)A landmark recognition algorithm based on corner feature matching is proposed in this paper.After studying the traditional corner detection method,an improved BW corner detection algorithm is proposed.It improves the speed of algorithm recognition while saving hardware memory consumption.Based on the analysis of corner jamming,a corner density de-noising algorithm based on distribution density is proposed.The recognition rate and accuracy of the road sign are improved.(3)A landmark location method based on Kinect depth camera is studied.Aiming at the problem of inconsistency between color image and depth image of depth camera,a depth image registration method based on binocular correction is studied.In order to measure the distance between landmark and camera,the depth distance correction of depth camera based on pinhole imaging principle is proposed.(4)Through comparison experiments,the speediness,accuracy and robustness of landmark recognition algorithm based on corner feature matching in different environments are verified.Finally,a snow removal robot mobile platform is built,and the performance of the snow robot positioning system is verified by static positioning experiments and dynamic positioning experiments.