Research On Industrial X-ray Image Enhancement And Algorithm Acceleration Based On Gradient Field

With the progress of society,people's living standard is improving day by day,and safety awareness is becoming stronger and stronger.Therefore,safety inspection plays an increasingly important role in life.X-ray detection technology is a non-destructive safety testing technology,this technology is widely used in medical,industrial fields.In order to improve the effectiveness of X-ray nondestructive testing,X-ray imaging technology has been developed continuously and for a long time.However,due to the hardware problems of X-ray nondestructive testing system and the complex and diverse features of the objects to be tested,the image quality usually has the problems of low contrast and low illumination.Therefore,in order to improve the effect of X-ray nondestructive testing,it is necessary to improve the gray contrast of the sensitive region of the X-ray image,the overall brightness of the X-ray image and the quality of the X-ray image.This paper describes the development history of X-ray nondestructive testing(NDT)and the current situation of using digital image processing technology to solve the existing problems of X-ray images.By analyzing the characteristics of X-ray images,various image enhancement technologies are studied,and then the algorithm modeling is carried out for the feature defects of X-ray images.In order to improve the practicability of the algorithm,this paper studies a variety of acceleration methods to accelerate the algorithm to achieve the purpose of real-time processing.Its specific research contents are as follows:1.The development history of X-ray detection system and the recent development of X-ray image enhancement using digital image enhancement technology at home and abroad are described.A variety of digital image enhancement techniques are studied,such ashistogram correction,grayscale transformation,image decomposition and reconstruction,and partial differentiation.2.Analysis of the characteristics of X-ray image,combining the advantages of a variety of image enhancement processing technology and the characteristics of the X-ray image modeling,the first to use pixel of image pixel gray value correction technique is compressed,then use the gray level transformation technology to normalized processing of images,and then use the gradient field reconstruction technology to enhance the detail of the image.In order to further improve the generalization ability and processing effect of the algorithm,the probability density of image grayscale value is used to modify the image grayscale,and the image quality is improved by image fusion in different stages.3.In order to improve the practicability of the algorithm,this paper studies a variety of ways to accelerate the algorithm.This paper studies how to use the CPU(Central Processing Unit)for parallel Processing to accelerate the algorithm.By using CPU acceleration,the algorithm can run quickly on a low-configuration computer.Considering the rapid development and wide application of high-performance computers,this paper studies the parallel acceleration of the algorithm by using the GPU(Graphics Processing Unit),which greatly improves the Processing speed of the algorithm and makes the algorithm meet the requirements of real-time enhancement,that is,it can process no less than 50 frames of images per second.