Research On The Noise Elimination And Contrast Enhancement Algorithm Of X-ray Image

Due to its own existence of many kinds of superior performance, the application ofSynchrotron Radiation is becoming more and more widely. Otherwise, the research of thecorresponding imaging technology and the related X-ray image processing technologyare also becoming more and more deeply. It seems that the X-ray images obtained fromthe Synchrotron Radiation will often get some problems, such as a low grayscale contrast,or an uneven grayscale distribution. Moreover, an imponderable and uncontrollable noisewill pollute the images invariably, which will cover the sample information on theimages. All these problems abovementioned will cause that the detail information of thesamples on the image are difficult to be observed and analyzed.Using the traditional methods of contrast enhancement to process the SynchrotronRadiation X-ray images, maybe it can enhance the contrast of images, but can’t solve theproblem of noise pollution; or maybe it can reduce image noise, but get another problemof that the effect of contrast enhancement is not ideal, and can’t improve the problem ofuneven distribution of image gray level. To solve these problems, a kind of suitablealgorithm for X-ray image contrast enhancement is researched in this paper.Homomorphic filter is set as the core of this algorithm, and combining with imagedenoising pretreatment and gray level adjustment, to realize the processing aim of imagenoise reduction and its contrast enhancement, and finally improve the image quality. Inthe procedure of realizing the image contrast enhancement algorithm, a number ofresearch work are finished, and some innovative achievements as below are acquired:1. Some different sample images and their histogram are utilized to analyze thecharacteristics of the image obtained from Synchrotron Radiation, such as low contrast, noise pollution, uneven distribution of grayscale and so on. And a methodbased on the image histogram is put forward to analyze the type of image noise, thehistogram is computed from the part of the image which is without sample, by whichthe type of noise is determined as Gaussian noise.2. In this paper, a noise reduction method based on wavelet transform is put forward, anda threshold function based on hard threshold method, soft threshold method andXiao-P Zhang’s threshold function is improved to process the wavelet coefficients;Moreover, the neighborhood average method is applied to process the waveletcoefficients, and a new noise estimator based on the neighborhood average methodof wavelet coefficients is presented, which compared with the general noiseestimation that the new one can obtain a noise estimator closer to the real value.3. After image denoising, the grayscale adjustment and homomorphic filtering is utilizedto enhance the contrast of the image. In this paper, the proposed algorithm can notonly realize the image contrast enhancement, and also reduce the noise of the image,and make the distribution of the grayscale becomes well-distributed.In this paper, it is mainly utilizing the software simulation and the experimentalsample test to analyze the performance of algorithm in this paper, the results show thatthe algorithm can enhance the low contrast X-ray images with uneven-distributedgrayscale, make the sample in the image displays more clear, and the gray leveldistributes more uniform; Simultaneously, the proposed algorithm has obvious effect toreduce the noise, and can keep the image edge and prevent the detail structure fromdestroyed.