The lug of the traditional flexible battery is a thin and flexible multilayer structure. Common welding methods include laser welding, ultrasonic welding, etc. In the welding process, due to clamping, knife closing and other processes, wrinkles, damage, folding and other defects are easy to occur.
At present, the defect detection after polar ear welding in China mainly depends on manual experience and visual inspection judgment, lacks corresponding and systematic defect detection procedures, is greatly affected by human subjective factors, and is difficult to ensure the accuracy and efficiency of detection. With the help of X-ray detection technology, the defect detection of lithium core lug after welding can be realized quickly and accurately.
The general process is as follows: first, conduct X-ray inspection on the lithium battery core lug after welding, conduct image pre-processing on the collected image based on the image processing of X-ray machine software, and use the image difference method to extract the corresponding defect features to achieve defect identification and detection; Finally, the effect of the whole visual inspection system is verified, and the feasibility and rationality of the machine vision inspection method are verified.
A) The purpose of image preprocessing in X-ray testing equipment is to remove noise, interference features and other irrelevant data and information in the image, strengthen useful information in the image, and improve the accuracy and stability of subsequent feature extraction. During the acquisition process, due to environmental factors such as dirt and lens pollution, the image will produce irregular noise of varying degrees. Therefore, without destroying the important features of the image, it is necessary to use image filtering to remove irrelevant noise. Common image filtering algorithms include average filtering, median filtering and Gaussian filtering.
B) The purpose of image enhancement and segmentation image enhancement of X-ray testing equipment is to enhance the grayscale of the filtered image, highlight the useful feature information in the image, enhance image recognition, and weaken or eliminate other irrelevant feature information.
C) After the image preprocessing after lithium battery welding, the image features of the weld imprint area will be highlighted to the maximum extent, so as to facilitate the identification and detection of the number of solder joints, the position of solder joints and the multi-layer lug. For the defect detection after lithium core welding, the key is to extract the quantity feature and location feature of solder joints. First, the linear gray level enhanced image is used to locate the features of the welded lug, and then the threshold segmentation image is used to find the central coordinates of each welding point and the contour edges of the multi-layer lug, and according to the central coordinates of each welding point, the central coordinates of the weld imprint area are fitted to find the relative position of the weld imprint area.
