Research On Surface Defect Detection Technology In Grinding Of SiCp/Al Composites

SiCp/Al composite materials have excellent performance and are widely used in various fields such as communications,land and ocean transportation.Although the SiCp/Al composite material has the advantages of higher strength,hardness and strong wear resistance due to the existence of silicon carbide particles,the characteristics of the silicon carbide particles weaken the processing performance of the material,resulting in SiCp/Al Various defects on the surface of the composite grinding process,such as pits.Various processing defects on these surfaces significantly reduce the surface quality and performance of the material.Carrying out the research on the surface defect detection and processing parameter improvement methods of SiCp/Al composite material grinding is of great significance to promote the application and development of the composite material in various fields.The research on the surface defects of SiCp/Al composites is the basis of the research on defect detection technology.In order to explore the types of surface defects in material processing,this paper carried out a grinding experiment of SiCp/Al composite materials.According to the SEM inspection images of the processed surface,the types and generation mechanism of surface defects were observed and analyzed.The experimental results show that the pits formed on the machined surface caused by the removal of silicon carbide particles in the grinding process are the main defect form of the machined surface,which provides a basis for the research of defect detection technology.The research on the surface defect detection technology of SiCp/Al composite grinding process is the key to obtain the defect characteristic parameters.Based on the research of machining surface defects,this paper collects the image of the SiCp/Al composite grinding machining surface through an ultra-depth-of-field microscope,and uses the labelimg software to generate an extensible markup language file containing the defect location and other information to complete the data Set production.Using the pytorch deep learning framework,combined with the Python language and training the data set based on the YOLOv4 algorithm,a detection model that can predict the location and type of defects is obtained.Train the defect detection model in advance to initialize the weight of the detection model;use the transfer learning idea to improve the training efficiency;add the ECA module to CSPDark Net53;use the K-means clustering algorithm to calculate the anchors suitable for the data set in this article.The results show that the average accuracy of the detection model for defect detection reaches 93.4%,and the detection time for a single image is 47 ms,which can realize real-time detection of machined surfaces.The high-precision detection model provides a guarantee for extracting defect feature parameters.In order to better extract the characteristic parameters of surface defects in the grinding of SiCp/Al composite materials,use the opencv image processing library and python language to perform a series of image processing on the defect area,separate the defect area and the background and detect the defect outline,use Opencv calculates the characteristic parameters such as the area and length of the defect contour.Based on the above research on detection technology and defect feature parameter extraction technology,the relationship between machining process parameters and grinding wheel parameters and defect feature parameters is analyzed,which provides a basis for better selection of processing parameters.