The Research On Target Recognition Techniques Under Complex Background Based On Edge

The techniques of target detection and recognition in the complex background have become a new direction in the field of target recognition in recent years.At present,there is no general method to deal with the problem of target recognition in different complex backgrounds.There are some algorithms that can deal with complex background target recognition,has not been applied because of complexity.While in the real visual inspection system,the disturbance information occurring in the natural background or in the motion state when a image is captured often makes the background of the target become more complex.In order to further study the target recognition in the actual complex background,this paper analyses and summarizes the application status of the machine vision detection technology and the research status of the target recognition in the complex background at home and abroad.Aiming at the two different types of complex backgrounds in the actual system and using the edge information of the target as the recognition characteristic,an in-depth theoretical research and an experimental exploration of the target detection and recognition technology have been proposed.The main research work is as follows:(1)Through the depth understanding of the composition and the calibration of the actual vision detection system,this paper used the classical calibration method Zhang Zhengyou calibration method to calibrate the visual system under the Matlab software environment.The calibrated result showed that the influence on the image acquisition accuracy by distortion in the vision system was eliminated after the calibration.(2)Aiming at the complex background with noise occurring in the state of motion and based on PCNN algorithm,a new edge detection method based on improved PCNN algorithm was proposed according to the actual situation.Combining with Labview program and Matlab program,this new method was applied to the actual vision detection system.Finally,the robustness and accuracy of the method were verified by experiments.(3)Aiming at the complex background of the change of ambient light,an edge detection method based on image depth information was proposed and the simulation experiment was carried out under Labview software.Conclusion was draw that this method had good robustness to the complex background edge extraction under natural light.(4)On the basis of theoretical research and combined with two practical research topics,a set of visual on-line detection system was developed.And the on-line detection experimentalplatforms for relay cover and tire belt were designed.The verification work of target recognition theories under the two complex backgrounds with noise points and ambient light changes was completed,and the expected detection result was achieved.