Material Defects Identificaion Based On Bioinformatics And Imaging Technology

Bioinformatics and imaging technology were widely used in the fields ofbiological sciences, medical sciences and so on. This study proposed six kinds ofdisaggregated indicators of surface defects on materials by extracting large amountsof surface topology of imags using bioinformatical imaging technologies, and thendeveloped a new method for identifying defects on materials. A processing andrecognition system for material classification based on the integrated classifier ofsupport vector machine, BP neural network and Fisher linear discriminant wasestablished.Firstly, this article provided an overview on bioinformatics, and describedseveral intelligent models commonly used in bioinformatics, and the main contentsand targets of the research were discussed. Considering the actual needs of thisinspection system, the basic principles of system design were proposed, and thesystematic composition, running processes, software interface design, function ofeach part and operation process of the inspection system were briefly discussed.Secondly, the storage format and surface characteristics of the research targetwere simply elaborated, and several classic digital image preprocessing algorithmswere discussed. In order to meet the needs of production environment, image processing algorithms based on comparative method were designed, involving thechoice of the reference image, correction information extraction, correction, scanningand feature extraction of the pending picture. The method of image reconstruction wasused to verify the effect of image processing algorithms designed in this research, tomake the algorithms best at last.Thirdly, several classifiers for image intelligent recognition were simplydiscussed, and support vector machine classifier, BP neural network classifier andFisher linear classifier used in this inspection system were elaborated from three keyaspects, such as the applicaion background, principles and structure, as well asprogram design and implementation, and an integrated algorithm was proposed basedon the three algorithms above.Finally, some simulated pictures were built according to the volatile factors ofthe industrial production environment, and the running speed and recognitionaccuracy of this inspection system was verified.The experimental results showed that the materials inspection system establishedin this research was feasible, and the running speed and accuracy met the demands ofindustrial production. At the same time, the system was scalable.