Detection Of Crack Defects In Metal Materials Based On Eddy Current Pulse Thermal Imaging

Timely detection of crack defects in metal materials is an important measure to ensure the safety of industrial production and the personal safety of workers.Therefore,nondestructive testing of key components of metal equipment,locating the location of crack defects and extracting the characteristic information of crack defects are of great practical significance to ensure the safe operation of equipment and improve industrial production efficiency.In this paper,based on the eddy current pulse thermal imaging testing technology,the eddy current thermal imaging testing experimental platform is built,and the eddy current thermal imaging crack detection identification model is established,which realizes the nondestructive testing and quantitative analysis of metal parts crack defects.This paper focuses on the feature extraction of infrared thermal image of metal materials and crack defect recognition,and solves the key problems of crack defect detection and quantitative analysis.The main contents of this paper are as follows:(1)The experimental platform of eddy current pulse thermography for crack detection was established.Based on the principle of eddy current heating,the platform heats the metal parts and obtains the infrared thermal image,which is uploaded to the host computer for analysis through remote transmission.The acquisition part of the system hardware consists of infrared dot matrix temperature measurement module,stm32f407vet6 chip as the control module,and esp-32 s as the Wi Fi module for remote wireless transmission.The collected data is stored and sent to the upper computer software through wireless transmission.The upper computer software transforms the transmitted temperature signal into infrared thermal image for further analysis,and finally realizes the remote real-time monitoring of metal material crack defect state.(2)In this paper,a multi feature extraction method for crack detection in infrared thermal image is proposed.Extracting the features of infrared thermal image of metal materials is the key to establish the health status recognition model of metal materials.In this paper,the color features and texture features of infrared thermal image of metal materials are extracted,which are used to construct the multi feature set of infrared thermal image and input into the crack defect recognition model of elm(extreme learning machine).The experimental results show that the proposed method can accurately identify the infrared thermal image with cracks.(3)Elm based infrared thermal image crack recognition model is affected by the random selection of extreme learning machine parameters,which leads to the instability of the performance and accuracy of the recognition model.In order to solve the above problems,a crack detection method based on structure Clustering Optimization Extreme Learning Machine(sco-elm)infrared thermal image is proposed.The K-means clustering number is automatically determined by affinity propagation(AP)clustering algorithm,which is used to optimize the structure and parameters of elm classification model.That is,the K-means clustering number,clustering center and clustering radius are used to determine the number of hidden layer nodes of extreme learning machine model and the expansion width of hidden layer node center of activation function.Experiments show that this method can effectively improve the recognition performance of elm classification model.(4)Aiming at the quantitative analysis of crack defects,an infrared image feature extraction method based on OTSU and Canny operator is proposed.Firstly,the infrared image with crack defects is classified and identified and denoised to enhance the feature information in the image;Then Otsu algorithm is used to separate the crack region from the background region,and Canny operator is used to extract the complete edge of the crack region.Finally,the perimeter and area of the crack edge are obtained.The experimental results show that the method can accurately locate the crack edge region of infrared image and extract the crack edge feature information of infrared image.In this paper,the infrared thermal image of metal materials is taken as the research object,and the experimental platform of infrared thermal image nondestructive testing is built,The results obtained are of theoretical significance for the detection and quantitative analysis of crack defects in metal materials.