基于随机森林的锂电池电极缺陷粗细分类法

A Coarse and Fine Classification Method for Surface Defects of Lithium Battery Electrodes Based on Random Forest

在锂电池极片生产过程中,传统的人工检测对电极片表面缺陷分类的效率及质量都不高,并且对于企业的生产工序优化不利,由此研究了电池电极表面的缺陷分类问题,提出一种基于随机森林的锂电池电极缺陷粗细分类法.该方法首先根据缺陷图像指定区域的灰度值变化以及依据缺陷形状得到的特定环形区域内的像素分布,将缺陷图像粗糙地划分为亮缺陷数据集和暗缺陷数据集,然后利用Gabor滤波器提取图像的特征,并使用随机森林分类器对这两大类进行进一步的细致分类得到最终的分类结果.实验表明,提出的对锂离子电池电极表面缺陷进行粗细分类法,不仅在分类精度和速度上都取得了良好的效果,同时还具有优越的鲁棒性.

In the production process of lithium battery electrodes,the traditional manual inspection is not efficient in classifying the surface defects of electrodes and the inspection quality is not high,which is unfavourable for enterprises who want to optimize their production process.In this paper,the classification of surface defects on battery electrodes is studied,and a coarse and fine classification method based on Random Forest is proposed.The method first roughly divides the defect image into light defect datasets and dark defect datasets according to the change of gray value of the specified area of the defect image and the pixel distribution in the specific ring area obtained according to the defect shape,and then extracts the characteristics of the image by Gabor filter,and further classifies the two categories in detail using the Random Forest classifier to obtain the final classification results.Experiments show that the coarse and fine classification method for surface defects of lithium battery electrodes proposed in this paper not only achieves good results in classification accuracy and speed,but also has superior robustness.