Shape Defect Recognition Model And Software System Based On Convolution Neural Network

It is the key step to identify the shape defects of the straightener when it is working.Science and technology are developing rapidly now,and the application of artificial intelligence is rising rapidly.It is recognized because it can make up for the shortcomings of traditional manual operation.Therefore,aiming at the problem of traditional shape defect recognition,in order to improve the recognition accuracy,this paper proposes a new shape defect recognition model,On this basis,the corresponding system is developed for easy operation.The key lies in the recognition of complex shape defect patterns and whether it can improve efficiency and accuracy in practical application.In order to solve these problems,this paper has done the following research:(1)In the preprocessing stage,by comparing the existing image denoising methods,the method with the best experimental effect is selected for denoising.Three commonly used methods in image processing are compared,and the best one is selected to process the image in this paper.By improving and optimizing the existing network model,this paper constructs a new model structure to adapt to the shape defect identification.In order to make the model suitable for the research content,when building the model,we need to make statistics on the existing convolutional neural network models and compare their performance,and then improve and optimize them.Compared with the basic model,the structure of the improved model is more in line with the research content of this paper.Through experimental comparison,the effect of the improved model is27.5% and 13.2% higher than the original basic model and the other two commonly used models VGGNet and Inception,which meets the requirements of plate shape defect identification.(2)Based on the above model,the corresponding model software system is developed to identify the defects in the plate.The system is divided into four modules: convolution neural network shape recognition model module,data management module,preprocessing and parameter management module,system help and maintenance module.And the demand analysis,overall design and detailed design of these modules are carried out respectively.(3)The development and experimental detection of the software system are completed.By inputting the plate image with defects to detect its functions,the results show that the system can effectively identify the types of plate defects,and the recognition accuracy reaches 91.28%,which can meet the operation requirements,and the operation accuracy is improved by about 16% compared with the conventional straightener.