Volume Holographic Wavelet Correlators for Classification and Quality Inspection of Industry Products

WY Feng,QS He,YB Yan,GF Jin,MX Wu
DOI: https://doi.org/10.1117/12.370241
1999-01-01
Abstract:Techniques of volume holographic associative storage and wavelet transform are combined to develop a system for classification and quality inspection of industry products. Wavelet transform is introduced to improve the recognition accuracy. The Comparison with the conventional correlation is studied. Multichannel correlation characteristic of the system is adopted to classify products. According to the position of the correlation output with highest intensity, a manipulator can classify the input product. Localization of products is used to solve invariance of the system in application. Variation of the highest correlation intensity is used to inspect quality of products. Curves of each product between correlation intensity and deformed degree are measured in advance. A critical intensity value is setting to judge a product is good in quality or not. In the inspection, the highest correlation output is detected and compared with its own curve. The unqualified product is rejected by the manipulator and the qualified one is accepted and classified. The classification and quality inspection of standard elements are used as an example in our experiment.
What problem does this paper attempt to address?