Defect Detecting System Of The Lithium Battery Pole Piece Coating Based On Machine Vision

With energy conservation and emission reduction, hybrid cars develop rapidly as a kindof new energy vehicles,as the core component of power battery, its quality is especiallyimportant. The surface of lithium battery pole piece is easy to be damaged in coating, rollerpress and so on, and these defects will seriously affect the quality of the battery and theservice life. The existing detection method is mainly non-contact detection method based onmachine vision, usually adopt two CCD cameras respectively detect the two sides of the polepiece through the coaxial lighting, and got a very good detection effect. However, at present,defects detection of power battery pole piece is realized in the space reserved on the coatingmachine or the pole soldering machine, which not only increase the cost of production, butalso reduced the flexibility of the detection.According to the detection requirements of HeFei GuoXuan high-tech power energy Co.,this paper complete the online testing in the existing models of pole soldering machine. Thissystem mainly solve defect detection in the power battery production process using the linearCCD camera and special light source, broke through a semipermeable mirror imaging anddefect intelligent regional identification key technologies, such as regional identification,realize double-sided defect of the online detection on the pole soldering machine. Increase thedetection flexibility, easy to expand its function ,and has important significance to improvethe detection efficiency and reduce the cost. The minimum defect area is 0.2*0.2mm, and thedetection speed is greater than 15m/min, this system can be used in pole piece defectsdetection of all kinds of power battery production industry. This paper illustrated the detectionprinciples firstly, researched the general structure of the system and the main hardwareparameters; Second, we design the light source, analysis and correct the pole piece vibrationand system error; Finally discussed the battery pole piece of image processing flow and thealgorithm design, combining with the practical application proposed a new algorithm to mark goals fast, completed classifier designing, and realized the defects automatic classification andrecognition.