| Inspecting quality of zonal steel include inner Inspecting quality and external inspecting quality. On the drastic competitive conditions, good external quality is not only the representation of enterprise visualize, but also is precondition of winning market. Inspecting disfigurement system of zonal steel surface has much applied value, which is monitoring equipment. In this years, micro-electronics technology, computer technology, automation technology and photo-electronics technology develop rapidly, and artificial intelligence and theory of nerve network have deepen and applied, which make inspecting disguise of zonal steel surface more and more intellectual, high precision, high credibility, high velocity.This Paper adopt image processing and image analyze and nerve network to recognize the disfigurement of zonal steel surface intelligently. It mainly includes the following aspects:l.When compared the common the methods, The methods of zonal steel image processing can be confirmed researching the method of image processing;2. When researching the pick-up of image-feature, This paper pick up theparameters of whole structure configuration and put forward a method of rapid polygon approach; After analyzed the parameter of the whole structure, The thresholds can be found;3.When researched sort technology of the network of nerve, A network of nerve named ART which can recognize the known disfigurement and unknown disfigurement is designed;4. The image processing procedure and image analyzing procedure and the network of nerve procedure are designed adopting Visual C++; The flat of experiment is built to gain the effect of design.In inspecting disfigurement, adopting image processing and nerve network is a new method, it has much merit to other method, it is high precision and high efficiency. In system adopting nerve network recognize the disfigurement, it can recognize the inexistence disfigurement of stylebook. In the following practice, the method will widely spread. |
PDF Full Text Request