Research And Implementation Of Industrial X-ray Image Enhancement Algorithms

X-ray inspection technology is a kind of non-destructive testing,which plays an important role in industrial flaw detection and security check.However,due to the influence of X ray detection hardware system and the nature size and other factors of the workpiece under inspection,the quality of the X-ray image will decline and the visual effect will become worse,causing certain difficulties in the process of X ray defect detection.Therefore,it is necessary to perform enhancement processing on the radio graphic image to meet the needs in the industry.In the image enhancement process,it is necessary to include as much detail as possible in the original image,while improving the image contrast and reducing noise within the visible range of the image.The paper expounds the current situation of industrial X-ray image enhancement technology at home and abroad,and studies several widely used image enhancement algorithms,and applies them to ray images,and analyzes and summarizes the enhanced effects and existing problems.In view of the low contrast and many details of industrial ray images,two improved image enhancement algorithms are proposed based on the existing algorithms,and the improved algorithms are simulated and verified.The research contents are as follows:1.Widely used image enhancement algorithms,such as adaptive histogram equalization,dynamic histogram equalization,limited contrast adaptive histogram,reverse sharpening mask algorithm,and Retinex algorithm,ect,applied to the X-ray image enhancement to analyze its effect.2.A multi-scale Retinex ray enhancement algorithm based on sigmoid function is proposed.By studying the single-scale Retinex algorithm and the multi-scale Retinexalgorithm,this paper adopts a S-type function model that is more in line with human eye characteristics,instead of the logarithmic model in Retinex,this method can avoid pixel overflow caused by logarithmic function,at the same time,before the image enhancement,the ray image is pre-processed,and a global weighting factor is introduced to suppress noise and improve contrast.Compared with the traditional algorithm,the proposed algorithm not only can significantly enhance the detail information,but also improve the contrast of the ray image and effectively suppress the noise.3.A ray image enhancement algorithm in the non-down sampled contourlet wave(NSCT)domain is proposed.The traditional contourlet wave transform will have pseudo-Gibbs effect when the image is enhanced,and the NSCT change has the characteristics of translation invariance,which can avoid this phenomenon.Therefore,the algorithm takes advantage of the NSCT transform and performs NSCT transform on the ray image to obtain the high frequency and low frequency coefficients.In the part of high frequency coefficient,aiming at the defects of general threshold value,an adaptive threshold value is proposed.Combining the adaptive threshold value with the coefficient discrimination criterion,the high frequency coefficient is further divided into three categories: strong edge,weak edge and noise.The low-frequency part is directly enhanced by linear transformation,and finally all the enhanced coefficients are transformed by inverse NSCT.The experimental results show that NSCT domain transform can enhance the detail of the ray image and has better visual effect.