The Research On Binoeula Vision System Of The Autonomous Mobile Robot

The research of mobile robot is an important embranchment of Robotics.At present,along with the continuous advancement of robot technology,the apply field of mobile robot has been more and more abroad,and its main development aspect is all independence,that is navigation,avoiding barriers and orientation self-determinatively.And the premise is the development of robot vision techniques,so vision technology of mobile robot is one of the hot topics in robot field recently.Images are taken from binocular vision system of mobile robot.After that the grayscale original image through the method of computing the gray-scale transformation was acquired in order to improve the computation speed.Taking into account the edge pixels to transform the flat,and perpendicular to the edge of the pixel to change dramatically,This paper proposes a framework for image acquisition and preprocessing:the expansion of the operator Robert edge detection method to detect the edge of the image point,and then use the Hough transform to improve straight-line fitting method by linking into the outline of a certain strategy,which constitutes the edge of the image.It is important that the camera calibration of binocular stereo vision system and robot working environment in the three-dimensional global coordinates relative to the position and posture in this article.Recover objects from the images of three-dimensional information space must be known to the object coordinate system with it in the image plane as the corresponding relation between,and this relationship is determined by the location of the camera,and imaging properties of the model parameters.Therefore,it is essential that the camera calibration.However,Traditional calibration methods need to create complex mathematical models to calculate a large amount of bad real-time.In view of the above shortcomings,this paper proposes an approach based on radial basis function(RBF) neural network method of binocular camera calibration,which establish the accurate relationship between binocular vision in the three-dimensional object point coordinates and between two camera coordinate,and have fast reconstruction and high precision computing.The simulation experiments show that the method is simple,and has higher accuracy and better robustness.Position is the most basic aspect of mobile robot navigation.Based on the proportion of strategy,this paper applies of artificial intelligence methods to guide the low-price positioning servo robot vision.We propose fuzzy neural network location of the binocular vision system which is 5-storey structure of fuzzy neural network using self-organization and supervised learning methods of studying the combination of hybrid algorithm.MATLAB simulation results show that:compared with other models,the method has a good tracking ability and non-linear approximation with high accuracy and response speed.