Research On Binocular Stereo Camera For AGV Navigation

With the continuous development of society and the continuous progress of science and technology,the modernization and upgrading of enterprise logistics system is imminent.As an efficient automated transportation means,AGV gradually enters people’s vision and plays a pivotal role in the process of replacing the traditional human transportation mode.The rapid development of machine vision and artificial intelligence technology makes the AGV integrated with vision more intelligent.Compared with monocular vision,binocular vision can use parallax to calculate the three-dimensional information of the target,so it has become a key research topic in the field of vision.In this paper,we design a binocular stereo camera for AGV navigation,which is limited by hardware devices such as cameras,lenses and some related algorithms,resulting in poor environmental adaptation,difficult feature matching and low efficiency of visual navigation.Based on the review of domestic and foreign literature,understanding of navigation technology and the current status of binocular camera research,the binocular stereo camera for AGV visual navigation is built for the visual navigation needs of industrial sites;Considering the need for real-time monitoring of visual navigation,the upper computer software is developed using MFC to realize the relevant functions of visual navigation,and the AGV navigation system is designed.For the needs of unstructured path recognition navigation,visual labels including position information and direction information are designed with reference to the image features of X corners;The segment test is used to extract X corners,X corners matching algorithm is designed to realize the correct pairing of feature points,and the binocular vision principle is used to convert two-dimensional coordinates into three-dimensional information,which in turn realizes functions such as label recognition and position detection.To analyze the various factors affecting the navigation accuracy in the process of visual navigation,mainly including illumination,threshold and equipment accuracy;To design an automatic exposure algorithm for the influence of illumination on camera image acquisition;To select the best threshold in the process of corner point detection,an iterative method is used to find the best threshold in different environments;To find the influence of equipment accuracy on X corners localization,a Kalman filter is used to estimate the 3D coordinates of X corners;Design relevant experiments to verify the effectiveness of the above algorithms,the results show that the above algorithm is effective in improving the navigation accuracy.In view of the complex working environment of the site,the visual navigation logic is designed,and various AGV control methods and motion modes are developed to meet the working requirements of the site;The binocular camera is applied to the AGV for verification,and the results show that the binocular camera built in this paper can provide reliable information for the visual navigation of the AGV,and the designed AGV navigation system gives full play to the advantages of visual navigation,and the AGV can complete the transportation of items within a certain range,which improves the productivity of industrial sites.