An Adaptive Vision-Based Method for Automated Inspection in Manufacturing

作者：Lin Tsun Kuo^*

来源：Advances in Mechanical Engineering, 2014, 6(0): 616341.

DOI：10.1155/2014/616341

This study proposes a new adaptive vision-based method combining discrete wavelet transform-(DWT-) based feature extraction and support vector machine (SVM) classification for automated inspection in manufacturing. This method involves transforming input optical images into a gray-level space and adaptively segmenting them by using region growing combined with DWT-based feature extraction based on support vector machines (SVMs). A multiclassifier SVM is first used to solve multicase problems in inspection. The SVM can be used to effectively classify samples based on the segmented images combined with the image features and perform superior multiclass classification. The proposed algorithm can select the most suitable features for the inspection from many features. The method achieves high-performance inspections and produces more favorable results than existing methods do.

• 出版日期2014