Research On Precision Casting Crack Detection System Based On Image Processing

Digital image processing technology plays an important role in foundry industry.Based on the advantages of accurate and reliable,non contact damage surface,high efficiency and easy integration,digital image processing technology has been widely concerned.With pertinence to the features of crack images of precision casting,an automatic detection and identification software for cracks of precision casting integrating image acquisition,segmentation,identification,measurements and extractions was designed and developed.Adopting the threshold segmentation,morphological processing,feature identifications as well as feature extraction and proposing a detection method with pertinence to the precision casting crack characteristics.The following works are accomplished in this paper:1.This paper analyzes the present research situation of magnetic particle testing technology and digital image processing technology,points out the shortcomings of the current magnetic particle testing technology,and combines with the detection method of image processing,the overall structure and design scheme of the automatic crack identification system based on image processing technology is constructed.According to the performance requirements of the automatic identification system,the hardware selection and software design flow of the system are analyzed in detail.2.This paper introduces the technology of image enhancement and denoising,and studies the algorithms of histogram modification,color enhancement and median filtering,mean filtering,sharpening filtering,and so on.Finally,the histogram correction method is used to enhance the image,and the median filter is used to denoise the image,which solving the problem that the original image is not clear because of uneven illumination.3.The image segmentation techniques such as edge detection,threshold segmentation and region segmentation are studied.Finally,Otsu is used to segment the image,realizing the separation of the target information and background information.And the image segmentation,image processing,image processing and closed operation are used to realize the identification of the connected region in the image.4.Through the calculation of the geometric features of the connected region in the image,the roundness is selected as the characteristic parameter of the defect,and the defect extraction of the crack image is realized by setting the threshold value of the connected region.5.This paper designs and develops a software for the automatic identification of cracks in precision castings,the morphology and number of cracks were calculated.The system solves the inconvenience of manual detection of workpiece defects in traditional casting industry,and improves the detection efficiency of the workpiece.The software is compiled,and the program is modify the transplant in the MATLAB environment,realizing the application of software in the environment without MATLAB.Finally,the system is tested on the platform without MATLAB environment,the crack image of the precision casting is processed.And the experimental results show that the system can detect the crack defects in the precision casting.The system effectively eliminates the influences on the detection results by the operators due to such factors as the detection environment,fatigue degree,visual resolution and working experience,and meet the requirements of the industrial site for the correctness,validity and practicability of the system.