| In recent years, artificial immune is another hot spot of artificial intelligence research after the artificial neural networks, fuzzy systems and evolutionary computation. Artificial Immune has intelligent behavior with a simulated biological characteristics, the results of the immune algorithm theory and application keeps on emerging, but it and visual detection technology combining image processing is still in the initial stage, has great space for development. Based on the study of artificial immune algorithm is proposed to the product automatic detection system of the improved negative selection algorithm, the system can identify the product images for image identification and classification provides a new method and means.Combined with image recognition and the immune system has a similarity, based on the negative selection algorithm is proposed the product image recognition and classification systems. First, the negative selection algorithm is to be part improved, combined with r-continuous matching rules and the characteristics of Hamming matching rules, proposed a new matching rule with a variable degree of matching theoretical and experimental results show that the method conducive to obtaining better matching threshold, generate a more suitable number of mature detectors and improve the detection accuracy and recognition rate. Secondly, the proposed and designed a product automatic detection system of the improved negative selection algorithm and algorithm implementation. The system needs to build the product image measurement system, access to product's original image, an image filtering and edge feature extraction to obtain a valid edge eigenvalues. Improved the negative selection algorithm can not adapt the diversity of products, it need to improve the negative selection algorithm consists of a state extended to a variety of state, to adapt the product testing requirements, therefore, based on the improved negative selection algorithm established an adapted multi-type of product detection system. With the traditional method of image recognition compared to artificial immune algorithm, identify the image of the artificial immune algorithm will help improve the image quality of the recognition rate and detection. |
PDF Full Text Request