Research On Key Algorithms Of Bearing Positioning System Based On Machine Vision

With the transformation of science and technology and the growing demand for material culture,the era of intelligence has arrived.High efficiency and high precision have become the basic requirements of all walks of life.As a kind of fixed rigid body and industrial parts bearing radial load,the importance of bearing in our daily production is self-evident.At present,many processes in the processing of bearings still rely on manual operations,and such traditional technologies are not well adapted to the needs of modern development.This paper studies the key algorithms of bearing positioning system based on machine vision to achieve more accurate detection and can greatly improve the production efficiency of enterprises.The main work of the thesis is as follows:(1)A hand-eye interpolation calibration algorithm for a multilayer workpiece is proposed.Due to the stacked placement of the workpiece,the conventional workpiece calibration technology needs to be carried out layer by layer,and the problem of the camera bracket lifting and lowering during the falling process of the camera may affect the accuracy of the final calibration.The improved multi-layer automatic interpolation workpiece calibration algorithm is based on the two-handed two-eye calibration data in the middle layer,and automatically interpolates and calculates the workpiece pose coordinates of the middle layers,which solves the need for layer-by-layer calibration due to spatial relative position changes.The problem is that the number of calibrations is reduced and the time is saved.The improved hand-eye interpolation calibration algorithm for multi-layer workpieces is measured to an accuracy of ±1 mm,which meets the production requirements.(2)An adaptive image fusion algorithm based on Mahalanobis distance discrimination is proposed.Due to the complicated industrial field environment(light,temperature,etc.)and the assembly error of the imaging system itself,and due to the similarity of the image of the batch workpiece to be tested,the traditional image fusion algorithm will lose detail,the color difference is obvious,and there is splicing.Sewing and other issues.In this paper,by evaluating the optimization of the overlapping regions of each camera,the appropriate selection of each lens imaging point set is automatically selected to make the fused image more realistic,which greatly reduces the image fusion error caused by the missing details of the spliced image.(3)Study the template matching algorithm.In this paper,the gray-based template matching algorithm is studied in detail and applied to the bearing workpiece positioning detection.The image smoothing processing is designed before the template matching algorithm.Several commonly used image filtering algorithms are compared and compared.The image processed by the median filter smoothing can well meet the detection requirements of the system.The gray template matching algorithm is compared with the shape template matching algorithm.The results show that the smoothed gray template matching algorithm has higher detection accuracy in the practical application of the bearing positioning system,reaching 99.99%,and it is in the corresponding industry.Robustness in production environments is better.The bearing positioning system studied in this paper applies ten sets in the actual production site,which can run stably for 24 hours,saves the enterprise cost,improves the production efficiency,and has strong anti-interference ability,stable and reliable.The technical content of this paper can be applied to a variety of small workpiece production areas,thereby improving the production efficiency of the enterprise.