Research On AGV Path Selection And Navigation Based On Multi-dimensioned Features

AGV is a power-driven intelligent equipment with a certain load capacity.With the help of the control system on the vehicle,it achieve functions of positioning,obstacle avoidance and path planning.In order to finish automatic guided transmission,AGV can be automatically guided to move and stop by the preset trajectory,in which case it meets the needs of flexible manufacturing and automated logistics transshipment.In recent years,with the development of machine vision technology,the visualnavigation is a research hotspot in AGV,at present through a specific single feature works well in AGV self-navigation,such as inspection line navigation.However,this navigation mode performs well only in the linear path navigation and its function is single.The content of the image contains abundant information and has wide detection range.Moreover,machine vision has incomparable advantages in object detection and object tracking.But the amount of computation is too large,so how to use a single vision sensor to achieve fast and precise path selection and target location is the current research in visual navigation in AGV.In this thesis,the image processing method is used to analyze the road path of AGV,and the deflection angle of the vehicle body and the route line is calculated according to the straight line feature.According to the camera imaging principle and machine vision knowledge,we build the correspondence between the image plane and the road based on the homography matrix to achieve the image of any two points of distance measurement.In order to solve the problem that the traditional AGV travel path is single and the workflow is fixed,this paper proposes a path selection scheme based on the recognition of the road sign,and sets the text road sign on the branch path.In the course of the AGV movement,the text region in the image is positioned and identified,and the path selection is made with the result of the recognition of the text sign.