A Robotic Visual Servoing System With Stereo Vision Based On Parallel Processing

During the rapid developing of machine vision and hardware, visual servoingsystem is the combination between visual information and robot control,that maderobot have the intelligent switching capacity with the external environment. This isone of the main direction of developing of robot. Visual servoing system through thedetection of the object image get visual information in order to control the movementof the robot. The servoing system approach which based on the position use the modalof robot, work space can be achieved and objective to estimate objectives in cartesiancoordinates in space relative to the location of cameras.Visual tracking which basedon computer image processing and analysis or comprehend of dynamic image becomeconcern topic. Image processing confront the complexity of the challenges. Parallelprocessing technology play a important role in this.The passage introduced the content of the research of visual servoing system,servo method and existing problems; the developing of robot vision and objectivecapture method are summarized. It is present analysis of the stereo vision system,the existing problem and development trend, which are related with machine vision,introducing the sorts of parallel processing system, and performance evaluation.The math models which include stereo vision servo system is established.Through the method of image processing, realize the strategy of servo system con-trol which based on parallel processing. Based on servo control strategy of imageprocessing layer, parallel processing layer and operational control layer, establishedrobot servoing system. At the same time, researches of cartesian coordinates trans-form relations, the relation of transform matrix between coordinates robot calibrationwith world coordinate system. In addition, through research of stereo vision plat-form structure, establishing the model of stereo visual, designing stereo visual plat-form model, which applied to visual servoing system. Camera parameters calibrationmethod is researched, and completion of the camera calibration.Objection detection and localization method based on visual servoing systemis studied in detail. According to the differently tasks of servoing system determinethe corresponding objects detection method. Putting the adjustable image processing strategies achieve accuracy of the priorities of computer speed, precision and robustimage processing. Two goals against the current image of a moment ago and the re-sults of the objective test results, the use of a two-stage stereo matching algorithm todetermine the best stereo matching set. According binocular stereo vision range ofmathematical models and options matching points of the current moment and objec-tives before a moment of transformation between the odd relationship. The forecastfor future goals and robot path planning calculations.It studies on the method of tracking and prediction of moving objectives, design-ing a robot path planning strategy. In the time predicting of moving objectives, usingmultiple observations lag estimates of steady-state Kalman Estimators model predic-tion method to achieve the estimation of moving object state and design binocularplatform control strategy and robot pain planning strategy according to the estima-tion results. Also given a binocular platform movement's security protection programand achieved fast, reliable and safe method of tracking moving targets. In additionintroduces a robot path planning method, the robot path real-time correction.In view of the cycle does not reduce the servo period improve positioning ac-curacy and the identification of the purpose, a three-dimensional robot visual servosystem of parallel image processing strategy. According to the design of parallelprocessing steps, a parallel set of rules. While the design of an adaptive load bal-ancing algorithm(ALCB), through the use of LC measure to adjust the parameters ofa processor, the establishment of an appropriate group of target detection and locationparameter combinations of the various processor QoS optimized to meet the machinehuman visual system in Servo Servo cycle, servo accuracy and robustness of the bal-ance between. The communications system, a combination of coding solutions in thevideo transmission disorder phenomenon, improve the robustness of the system. Inaddition, parallel performance is researched from period, speedup and efficiency ofthis strategy. While the system scalability analysis, and gives the treatment measures.Finally, the establishment of a set of related hardware tracking experimental sys-tem. Through multiple experimental group measured system performance indicatorsand achieved the desired results, test the effectiveness of the algorithm and real-time.