The method researches of ferrography wear particle ecognition based on artificial immune theory
In this paper artificial immune arithmetic is exploring applied in ferrography wear particle image identify technology. The normal wear state is defined as the system of itself while abnormal state as the non-own. The antigen is equivalent to the pattern to be recognized, while the antibody is equivalent to the identification function or the identification vector. The Affinity is equivalent to decision standard that show the capability of the identification function or the identification vector. Artificial immune arithmetic is applied in ferrography wear particle image identify technology on the basic of representation of biology models and arithmetic derived, including common immune arithmetic, negative selection arithmetic, immune learning arithmetic, clone selection arithmetic and wear particle classifier based on artificial immune system (WPCAIS) is designed. Taking shape factor, the Euler Number and filling degree as eigenvectors, the sliding, cutting, normal, fatigue erosion wear are classified 1 and the simulation test of classifier is taken.
wear ferrography technology wearparticle image recognition Artificial immune systemwear particle character
Gu Daqiang Fan Jun Zhou Lixia
Department of Mechanical Engineering, Zhejiang University, Hangzhou, 310027
国际会议
北京
英文
2007-08-05（万方平台首次上网日期，不代表论文的发表时间）