为提高废旧铅酸蓄电池回收效率和金属回收质量,提出采用自动化手段分类废旧铅酸蓄电池.针对该识别分类问题,首先将主成分分析法、线性判别分析法应用于废旧铅酸蓄电池的X射线图像的特征提取.通过支持向量机对提取的训练集图像特征向量进行训练,分别对测试集的图像数据进行分类实验,并对比了不同数量的训练集和不同特征空间维度下各种方法的识别率.实验结果表明,主成分分析法、线性判别分析法可用于废旧铅酸蓄电池X射线图像的识别,并且随着训练集与测试集样本量的增加,二次线性判别分析法表现出较为稳定的识别率.%In order to improve recycling efficiency of used lead-acid and recycling quality of metal, the automated method to classify used lead-acid was proposed. According to the recognition classification problem, first, used PCA and LDA to extract feature of X-ray image about used lead-acid batteries. Then, trained the extracted image feature vectors of the training set by SVM and identified test set in the classification experiment, finally, compared the recognition rate in different numbers of train data and different feature space dimensions. The result shows that PCA and LDA could be used to recognition of X-ray images about used lead-acid batteries, and with the increasing of number of test data and the train data, the Secondary LDA showed more stable than others in recognition.
展开▼
