Research Of Nsst Image Mosaic Algorithm Based On Compressed Sensing In Surface Defect Detection Of Parts

The traditional defect detection of parts surface usually adopts artificial methods to detect the defects on the surface of the products.Not only is the speed slow,but also the reliability of product defect detection cannot be guaranteed effectively,and it cannot meet the urgent demand of the enterprises for increasing product quality.With the continuous development of science and technology,various defect detection methods based on machine vision continue to emerge.Due to the large amount of image sampling data and the strong noise interference at the scene,these defect detection methods are difficult to meet the real-time requirements for defect detection.In the image mosaic research field,the quality of the finished large-field image is closely related to the effectiveness of the image representation method.As a multiscale analysis tool,the NSST transform is different from the traditional shear wave transform.The NSST transform does not perform the downsampling operation when decomposing images at multiple scales,and the NSST transform can overcome the pseudo-gibbs phenomenon when reconstructing the image.Decomposition of multi-scale image by NSST transform,the high frequency coefficients of large amount of data,and the sparsity of the relatively large,according to its characteristics,this paper introduces the compression perception theory,and studies a NSST image mosaic algorithm based on CS.Which not only improves the quality of fused images,but also achieves the real-time requirement of image mosaic system.The main work of this paper is as follows:(1)The research background and significance of image mosaic technology and the research status of compressed sensing at home and abroad are summarized.Finally,the research contents and organization arrangements of this paper are given.(2)First,the image mosaic process is introduced in detail.Secondly,the methods of image registration and fusion are summarized,and the characteristics of each method are summarized.(3)This paper introduces the basic theory of Shearlet,the implementation method of NSST,the application in image fusion,and the evaluation method of image quality is given,and verifies the feasibility of the application of NSST in image fusion through simulation experiments.(4)Aiming at the large amount of high-frequency coefficient data generated by NSST multi-scale decomposition and the high frequency coefficient's large sparsity,this paper proposes an image mosaic algorithm based on CS and NSST.The results show that without prior knowledge of source images,the effective fusion of images can be achieved.Although the quality of fused images will be sacrificed very little,the computational complexity of the algorithm will be greatly reduced.It provides a way of thinking for the application of the image mosaic system with high real-time requirement.(5)The hardware system of the quick defect detection device for the sleeve-type parts was built,and the image preprocessing system and the splicing system were designed.The image preprocessing system performs graying,denoising,and other processing on the parts images acquired by the image acquisition device,to provide the basis for the next image splicing,and the image splicing system realizes the splicing of the pre-processed part images.Using the image mosaic algorithm based on CS and NSST is proposed in this paper,the stitching image is effectively fused,which provides the support for the image splicing technology for the rapid detection of the next part surface defect.