The Surveillance Technology Of Battery Anomalous Event Based On Image Processing

With the development of the electronic industry, the production of the globalbattery is increasing. The batteries are widely used in electronic products, and vitallyrelated to human lives. But, the safety accident which caused by batteries never stop. Inorder to ensure the safe use of batteries, The United Nations dangerous goods transportproposal test and norms3.3section,188articles defines: batteries which delivered totransport, should pass the battery UN38.3test. However, UN38.3test is time-consuming,labor intensive, objectivity results, threat labor’s safety in test.We proposed surveillance technology of battery anomalous event based on imageprocessing. Use industrial camera as the image detection unit to get the battery image,detect the leak and burning anomalous event in battery test by using image analysistechnique automatically. In leak detecting, we use Speed-Up Robust Features algorithmto mach standard template and the being tested image, then use dynamic threshold todetect the area of the battery leakage. In burning detecting, based on the rhythm and huecharacteristics of fire, we combine the methods of Three Image Difference andContinuously Adaptive Mean-Shift algorithms, detect the fire area in real-time. Thesurveillance technology of battery anomalous event based on image processing assist toremote monitor, guarantee the safety of labor in test process, improve the objectivity ofthe test result, and save the video data of the testing process in disk.The leak and burning detecting algorithms we proposed are tested, which provethat the surveillance technology of anomalous event in battery test based on imageprocessing can detect the anomalous event which happens in lithium battery testaccurately. The accuracy rate of leak and burning are91.0%and98.2%respectively.