Researches On Image Lossless Compression And Denosing Techniques

At present, the digital image technology has been widely used in areas such asvideo communications, computer, radio and television . It made a high demand at ahuge amount of video data on the storage space and transmission channel , but alsosuch as aerospace, medical and other fields of the video images the qualityrequirements of lossless or near lossless. Therefore, the video image compressioncoding technology research under the current environment has become increasinglyimportant and growing role in its remarkable, but also has become more compellinginfluence factor in the development of high-tech fields. Therefore, in order tosuppress noise and improve image quality, ease of further processing, research fastand effective image lossless compression method is still a current research focus.The image acquisition and transmission process is vulnerable to variety offactors, makes the collected images often contain noise. In order to follow-up imageprocessing, such as compression, it is essential to the image denoising processing.The noise removal has become a very important step in image processing. Therefore,in order to suppress noise and improve image quality, ease of further processing,research fast and effective de-noising method is still a current research focus.This article focuses on the image lossless coding and denoising technologies.In-depth researches have been made in the key technologies and applications, duringwhich new algorithms and system design are proposed. The main contribution of thispaper summarized as follows:1. The basic principles of image and video compression are studied, with emphasison the theories and research status of lossless coding. The existing standards andalgorithms are analyzed and summarized, based on which we analyze andsummarize the classical image denoising algorithms under rhe framework oflossless compression syetems. Meantime, the development and application ofvideo compression systems are summarized. We also compared the mainstreamsystem design methods and pointed out the invididual advantages anddisadvantages.2. An improved lossless image coding algorithm is proposed by combining integerwavelet transform and SPIHT. This section focuses on the integer waveletmethod and a more reasonable balance between the coding speed and efficiency. Good compression ability can be obtained for the image data by adjusting thecorresponding SPIHT coding structure using the coefficient distribution of theinteger wavelet. Experimental results show that when testing different kinds ofimages, the average bit rate obtained using the proposed method is lower thanother methods, and the coding time is significantly reduced. Besides, theproposed method can chieve better compression performance under smallerstorage space, so that it is suitable for hardware implementation.3. A new adaptive lossless video coding algorithm is proposed. A new method foradaptive forecasting model selector design is given. The proposed method usesthe redundant information of time, space and transform domain, and achievelossless compression by backward adaptive model with reduced transmission ofboundary information. The proposed method uses the adaptive prediction modeselector to replace the extra bits prediction model so that it controls thecomputational complexity better. Compared with the existing algorithm, theproposed scheme shows better performance and the compression rate improvessignificantly.4. An image noise reduction method is proposed combining the Non-subsampledContourlet Transform (NSCT) and Gaussian mixture model. The algorithm isbased on NSCT transform. It builds the Gaussian mixture image model (GSM)and uses the Bayesian estimation to obtain the de-noising model. Secondly, anadaptive image denoising algorithm is proposed by combining theNon-subsampled Contourlet Transform and SURE guidelines. This methodestablishs an estimate of the MSE based on SURE, and achieves noise removalusing the adaptive adjustment of Countour details with different scale anddirection after linear threshold image decomposition.5. To solve the problem of Gibbs artifacts, an image denoising method is proposedby combining model of total variation firstly. The original image is firstdecomposed using the Nonsubsampled Pyramid Filter and the decomposedimage model is built based on total variation model. Then the image areproducing the preliminary denoised image after reconstruction. Afterwards, adetail compensation image is obtained using the difference between thepreliminary denoised image and the original image. Finally, the denoised imagecan be obtained by adding the compensation image to the reconstructedimage.Then an image denoising method is proposed by combining NSCT andadaptive model of local total variation. The method filtered image using adaptive local total variation model. Then image noises can be removed whilepreserving the image details. Experimental results show that the above methodscan effectively remove the image noise and Gibbs artifacts, during which imagedetails are preserved so that better image quality and higher PSNR values areobtained.6. For The requirements of lossless compression in current field of wide range ofembedded system, this paper designed and implemented a general-purposeembedded image lossless compression system used in space and a losslessimage compression system. System first general-purpose lossless compressionalgorithm combines relevant to TMS320DM642 core processor which canrealize a variety of lossless compression algorithms , and can be used in avariety of industrial fields. Another combined with CCSDS technology whichused FPGA (STRIX-II) as the core processor to image lossless compression canbe applied to space image processing system. The main features of the systemare embedded design, hardware, real-time compression, small size, low powerconsumption, which can be widely used in satellite remote sensing, aerial flightsand other fields.