Automatic Recognition Model Of Steel Sheet Surface Defect Based On Deep Learning And Application In Quality Management Of Steel Sheet Surface

Industrial enterprises,especially steel enterprises,as the backbone of the industry,are the foundation of the country’s prosperity,the soul of national rejuvenation,and the foundation of industrial development.In recent years,with the support of relevant policies and the recovery of the steel plate market,the steel manufacturing industry has achieved rapid development in recent years.However the iron and steel manufacturing of our country is of large field but not excellent,massive on scale but not strong.The specific performance is that China’s steel plate production is high,a significant gap still exists between us and european countries and the most direct and primary factor reflecting the quality of steel plates is surface quality of steel plate.Therefore,improving the quality management level of steel sheet surface has become an urgent problem for steel enterprises.In view of this,this paper has carried out the following research on the quality management of steel sheet surface.Firstly,through the deduction of quality and quality management theory,the concept of steel surface quality management is defined,and the quality management methods,objectives and principles of steel surface quality are clarified.From the current management situation,the problems existing in the quality management method of steel surface are analyzed and explained the significance of this paper.Then,for the purpose of improving the quality management level of steel plate surface,an automatic recognition model of steel plate surface defects is constructed.Using the deep learning-based target detection algorithm Faster rcnn as the basis of the automatic surface recognition model algorithm for steel sheet surface defects,we optimize some functions and overall structure of the algorithm,and do a comparative experiment of model algorithm to verify the automatic recognition model algorithm of steel sheet surface defect.The difference between the target detection algorithm and the accuracy and speed of the surface defect detection of the steel sheet.Finally,the automatic identification model of steel surface layer defect is used to realize the alarm function for steel surface quality monitoring,and determination of steel plate and the retrospective management for surface quality control of steel plate.Finally,the steel sheet surface quality management based on steel sheet surface defect automatic recognition model is put forward.Ultimately,the automatic identification model of steel surface layer defect is used to realize the alarm function for steel surface quality monitoring,steel surface layer judgment and backtracking management for steel sheet surface quality control.At length,the steel sheet surface quality management based on steel sheet surface defect automatic recognition model is realized.The research aims to realize the tool revolution of steel sheet surface quality management technology from manual quality management to model algorithm quality management based on data.The purpose of this study is to build on data,computation and algorithms.In order to realize the tool revolution of surface quality management technology of steel plate from manual quality management to model algorithmic quality management calculation power and algorithm,and realize the decision-making revolution of steel enterprise quality management from experience decision to data and algorithm decision-making.Through the steel plate surface quality management technology empowerment to promote the closed-loop control of the steel surface quality,to achieve high-quality steel supply in China’s steel enterprises,significantly enhance the quality advantage of China’s steel industry in the international competition.