3D Construction Of Target Characteristics Based On Binocular Vision Technology

Binocular vision technology can recover the3D information from two imageswith disparity. The technology has a lot of virtues, such as large amount ofinformation captured, wide range of applications and non-contact measurement.Studying the binocular vision technology and exactly recovering3D of targetcharacteristic can help robot navigation, industrial positioning a nd aviationconnection,etc. Based on binocular vision technology, this study is mainly about3D information recovering of target characteristic.Camera calibration technology is the basis of binocular vision technology. A newmethod for camera calibration based on concentric circles with lines has beenproposed based on traditional method for camera calibration. Due to using themethod of curve fitting, this method of camera calibration can obtain more accuratereference points and more accurate model of camera. The distortion can becorrected with the parameters of camera calibration, and polar line can be correctedby the method of rotating polar line.In order to improve the accuracy of3D information recovering of target charac-teristic, the noise and brightness difference should be eliminated. The three stagesof image preprocessing can help filtering image noise, balancing image brightnessdifference and enhancing the weakened details. The target and corner characteris-tics can be identified and extracted by studying the technology of feature detection.Stereo matching is kernel for binocular vision technology, corresponding inform-ation of stereo matching can be used to construct the disparity of target characterist-ic. A new method of the dynamic programming algorithm with feature points hasbeen proposed on the basis of traditional dynamic programming algorithm. Thisalgorithm is based on the corner characteristics to optimize the traditional dynamicprogramming and then improve the execution efficiency. The technology of3Dinformation recovering has been studied to gain the arbitrary space coordinates.Based on the theory of binocular vision technology, the experiment of studyingthe static and dynamic characteristics can prove the proposed algorithm ofbinocular vision technology. The results of experiment show that binocular visioncan find the exact position of static target characteristic, dynamic identify and real-time track the dynamic target characteristic and then proved the effectiveness andreliability of the algorithm proposed.