Research On Intelligent Image Defect Detection Based On Deep Learning

With the rapid development of artificial intelligence today,various industries are saving people from heavy work through continuous innovation technology.The maturity of artificial intelligence also brings a broader development space for traditional industries.Deep learning in industry has been widely used in various industries,such as detecting food safety hazards,extracting the most important related features of disease diagnosis and treatment-oriented in medical images,and intelligent detection of key components in automobile production.These are all intelligent detection through in-depth learning in order to reduce costs and improve efficiency.Compared with in-depth learning,the learning challenge of Intelligent Detection is that it has more room for development,and the accuracy and accuracy of detection need to be gradually improved.In the intelligent detection of forging defects,with the increase of the number and type of defects,problems also follow.In industry,it is necessary for manufacturers to find defects early in the production process in order to save costs and ensure the quality of production.If only traditional detection methods are used in defect detection,different inspectors will have different subjective judgments of the defect,which will lead to inaccurate and labor-consuming defect judgment.However,a large number of manual tests and the impact of the surrounding environment can also lead to misjudgments and missed judgments due to fatigue.Therefore,it is important to find a way for defect detection of industrial forgings to improve the detection accuracy and reduce the detection time after introducing a deep learning method.In view of the above problems in the in-depth learning of defect detection,this topic builds an in-depth learning-based intelligent detection of defects in industrial forgings,the main research contents are as follows:(1)This paper summarizes the current research status of in-depth learning of forging defect detection at home and abroad,classifies in-depth learning defect detection and Xray image defect detection of forging,introduces the classical algorithm and model structure,and summarizes and prospects the application prospects and development trends of in-depth learning technology of forging defect detection.(2)For the in-depth learning of forging defect detection,existing target detection models will face the problems of slower detection speed and lower accuracy.To solve this problem,this paper focuses on the main network of target detection model,follows up the characteristics of the main network,and selects the appropriate network to replace,in order to improve the speed and accuracy of forging X-ray image defect detection.(3)In order to train and detect the X-ray defect image dataset of forgings better in the model,this paper carries out data labeling and image preprocessing according to the characteristics of X-ray defect image of forgings.The X-ray image data collection of forgings used in this paper has the drawbacks of large size and a wide variety of defects.For large size defects,the dataset is cut and split appropriately.Because there are many kinds of defects in the X-ray defect image data set of forgings used in this paper,it is necessary to classify and label the data set to label several prominent defects for model training.In addition,in order to improve the final effect of the model,this paper also uses data enhancement,combined with appropriate image processing,to make the final result more obvious.