Development Of Full-size Measurement System For Power Battery Module Based On 3D Vision

New energy vehicles are the direction of a new round of technological revolution and industrial transformation.As the "heart" of new energy vehicles,the size accuracy of the power battery has an important impact on its compression resistance,explosion-proof performance and sealing performance.Therefore,the realization of fast online accurate measurement of the full size of the battery module in production is the top priority in its production technology.This paper uses machine vision technology to build a full-size detection platform for power battery module(i.e.,the length,height,thickness,flatness of mounting surface and aperture of mounting hole),including servo motor control module,image acquisition module and image processing module,to achieve online accurate detection of the full size of power battery module.The main contents of this paper are as follows:Firstly,the laser triangulation imaging technology and image filtering method are used to establish the three-dimensional point cloud model of the power battery module.According to the requirements of the detection accuracy and the characteristics of the overall dimension and structure of the battery module,the laser triangulation method is adopted as the three-dimensional non-contact measurement method of the power battery module,and the full-size measurement system of the power battery module is calibrated,and the three-dimensional imaging system of the battery module is completed.For the original image of the battery module,the median filter and Gaussian filter are used for filtering and denoising.Secondly,the plane fitting algorithm is used to realize the fitting of the mounting surface of the battery module,and the algorithm for measuring the height and flatness dimensions of the battery module is completed.A plane fitting algorithm of power battery module based on data optimization is designed.This method uses PCA method to reduce the dimension of the battery module image data again,and then uses the least square algorithm to determine the optimal threshold and eliminate the outliers in the point cloud data for several iterations,and finally obtains a more ideal estimation model.Thirdly,the edge detection algorithm is used to realize the fitting of battery module profile and mounting hole,and the measurement algorithm of battery module length,width and mounting hole diameter is completed.A variety of edge detection operators are used to process the battery module image.After comparing and analyzing the results obtained,Canny operator is selected as the edge detection operator of the module.According to the requirements of size accuracy,an improved edge detection algorithm for battery module edge is designed on the basis of sub-pixel accuracy to achieve high-precision edge detection.The feature detection algorithm is used to detect the geometric shape of the contour and the mounting hole,and the fitting of the linear contour of the battery module edge and the mounting hole circle is completed.Finally,the visualization software platform of the full size measurement system of the battery module is developed by applying the size measurement method of the battery module studied,the size measurement experiment is carried out,and the error generated by the measurement results is analyzed,and the improvement measures are proposed according to the corresponding error,and the development of the full size measurement platform of the power battery module is completed.