Mobile Robot Visual Servo Control System

With the development of image processing technology and computer processing ability in recent years, vision-based mobile robot technology is rapid development. Use image segmentation and recognition, complete target identification and tracking function for mobile robot system.In this paper, we designed an automatic material transport mobile robot, which is used in unmanned factory. It used machine vision sensors for sensing the external environment, completed vehicle body control based on fuzzy intelligent control algorithm, have the ability to running in accordance with pre-set orbit in the factory.Through research of this paper, get the following results and conclusions:(1) Based on ATOM processor Industrial control computer and ARM7 processor motion controller, we designed a mobile robot visual control platform with the advantage of rich computing power, easy development and low power consumption.(2) By analyzing the characteristics of road signs, we designed a road sign extraction algorithm based on parallel feature, and used geometric invariant moment algorithm to verify the target selection of this method.(3) Try adaptive threshold segmentation, region growing algorithm, color segmentation and watershed segmentation to deal with the environmental features of this paper, using the ultimate measurement accuracy segmentation assessment criteria (UMA) to evaluate this segmented result.(4) Based on template matching method to achieve the goal of continuous image sequences tracking, and use image pyramid to optimize this method. It reduces the computational complexity and shorter the processing time.(5) Based the image extraction of interest points, we using the template matching algorithm to calculate their three-dimensional coordinates in stereo vision system. By clustering processing of the interest points, we obtain the obstacle information in the image.(6) Analysis the dynamics model of differential wheel system. According to the difficult to establish the specific characteristics of visual processing system, based on fuzzy control algorithm, using point tracking control strategy for mobile robot to complete this control system. Using MATLAB for simulation to verify this control system's feasibility for mobile robot visual control.