An application of parameter extraction for AGV navigation based on computer vision

In this paper we realize an application of a computer vision method to acquire some basic information to control the Automated Guided Vehicles (AGV). Firstly, we design markers refer to some special meanings (such as “Go Straight”, “Left Tum”, “Right Tum” and “Stop”). Secondly, we use camera to capture the markers, using HOG feature and classify the markers with SVM predictor. After that we calculate the distance and deviation angel from the markers to the AGV. Finally, we send the information to the AGV and control the movement of it.