On Intelligent Detection System Of Automobile Interior Panel Same Color Lock Based On Machine Vision

Machine vision is a comprehensive technology that replaces human eyes with machines.With the continuous progress of technology,machine vision has been widely used in industrial production.Compared with traditional manual detection,machine vision detection has a high efficiency of product detection.In the production process of automobile interior trim panel,manual placement of automobile interior trim panel will cause its position to change greatly,There are different types of automotive interior panels,which leads to the uncertainty of the position to be detected,and it is impossible to complete the identification and detection of specific positions.In addition,Image deformation,false detection and false measurement caused by camera distortion,target false measurement and other problems.With the increase of matching error,the accuracy of detection of the same color lock missing is reduced.In view of the above problems,this paper studies the intelligent detection system of the same color lock of automobile interior panel.First of all,according to the detection requirements of the same color lock of the automobile interior panel,the overall framework of the detection system is given,the functions of each component of the system are analyzed,and the principle and parameters of all the required hardware are analyzed in combination with the influencing factors in the actual detection environment,and the hardware selection is completed.In order to solve the problem of image distortion caused by camera distortion,the radial distortion and tangential distortion of the camera are studied deeply.The distortion elimination algorithm is added to improve the calibration method of Zhang Zhengyou camera,so that the calibration algorithm can be calibrated under the condition of considering lens distortion,obtain more accurate product image information under the premise of obtaining optimized camera parameters,and better solve the problem.The influence of camera distortion is discussed,which makes the precondition and foreshadowing for image template matching.Lot’s of experimental data show that the improved algorithm improves the accuracy by 1.13%,which proves the necessity of considering camera distortion and the feasibility of Zhang Zhengyou’s camera calibration improvement method with the addition of distortion elimination algorithm.In addition,in order to solve the problem of random positions and different types of products to be inspected,products beyond the "field of view" of the visual camera,barcodes,springs,misdetection and misdetection of dropped parts,etc.,improve the template for automotive interior panels The matching speed and accuracy ensure the accurate positioning of the position to be detected,given an account of a geometric edge feature algorithm that adds stop criteria and image pyramid search strategy;the gradient direction is obtained by curve fitting.Accurate location,elimination of redundant target interference,accurate reduction of the scope,shorten the processing time of the algorithm.After the camera calibration and image correction using the improved method of Zhang Zhengyou calibration method,combined with the accelerated geometric edge feature algorithm,the car interior panel images under various conditions such as strong light,rotation,weak light and occlusion were detected 1000 times,and the detection results were analyzed.The results show that the accuracy of detection is as high as 99.8%,and the accuracy of detection can fully meet the production requirements.It is proved that the geometric edge feature matching algorithm is very suitable for the detection of the same color lock missing of the car interior panel.At the same time,it also shows the effectiveness of the combination of this method and the improved Zhang Zhengyou calibration method.Finally,in order to make the system function more complete and the practical operation more simple,the software interface of the detection system is designed and implemented.Based on the MFC framework,the software interface integrating camera module,communication module,image processing module and result display module is developed,which makes the detection system more able to meet the actual needs of production detection.