The Application Of Machine Vision In The Inspection Of Welding Defects In The Polar Ear Of Soft Pack Power Lithium Battery

In recent years,the domestic new energy vehicle industry has developed rapidly.As a result,the demand for power batteries continues to increase.As a new power battery,soft pack power lithium battery has broad market demand and good development prospects.In the production of soft-pack power lithium batteries,the quality inspection of the welding seam of the pole ear is particularly important,which is related to the quality and performance of the finished power battery.However,there are very few studies on the detection of welding defects of the pole ears of soft pack power lithium batteries.To this end,the article applies Machine Vision to the detection of welding defects of soft pack power lithium battery pole ears,and provides a feasibility study idea.The difficulty of detection lies in how to accurately extract the weld seam from the high-reflective,low-contrast,and noisy pole welding seam image,and select appropriate features for classification and recognition.The main content of the thesis includes the following four aspects:Firstly,the types of power batteries were introduced.The composition structure of the soft-pack battery and its advantages of high energy density,good safety performance and flexible design are analyzed in detail.The types of weld defects and the formation mechanism of weld defects are systematically explained.Lay the foundation for future experimental analysis.Secondly,the processing methods of polar seam image are studied,including image denoising,contrast enhancement and image segmentation methods.The experiments compare the mean iterative segmentation,the maximum entropy,the segmentation based on particle swarm algorithm,adaptive-threshold and OTSU image segmentation methods.Aiming at the characteristics of high-reflection and low-contrast of polar ear welds,an image segmentation method of polar ear welds based on morphological reconstruction and OTSU was proposed.The research results show that this method is suitable for segmentation of polar ear weld images,and the speed is also improved on the premise of ensuring accurate weld extraction.Then,the feature extraction and classification methods of the ear weld are studied.The research is based on the decision tree method and the template matchingalgorithm for the identification of polar ear welds,and the template matching method is used to classify the polar ear weld defect types based on the prior feature threshold.The results show that the welds with overlapping defects have an impact on the decision tree method,and the feature-based template matching method has better recognition accuracy than the former.Finally,the design scheme and platform construction work of the ear welding seam inspection platform are studied.It mainly includes hardware platform and software platform.According to the characteristics of the high-reflection and low-contrast of the welding seam of the lithium battery of the soft pack,by selecting the appropriate ring light,BASLER camera,Kowa Lens LM35 HC lens,build an inspection platform.The software testing scheme is designed,and the coordination task of MATLAB and Lab VIEW is adopted to realize the testing task.