通信专业文献及翻译--基于jpeg2000技术，融合dsp和fpga的可纠错图像传输系统(编辑修改稿)

通信专业文献及翻译--基于jpeg2000技术，融合dsp和fpga的可纠错图像传输系统(编辑修改稿)内容摘要：

et contained in the bit stream is read, and the data and parameters of each codeblock are extracted, and fed as inputs to the EBCOT decoder. EBCOT Right after the bitstream syntax parser, the subsequent stage in the JPEG2020 depression chain is the entropy decoder (EBCOT). From an algorith mic point of view, EBCOT is a blockbased bitplane encoder followed by a reduced plexity arithmetic coder (MQ). It subdivides each wavelet subband into a disjoint set of rectangular blocks, called codeblocks. Then the pression algorithm is independently applied to every codeblock. The samples of every codeblock are arranged into socalled bitplanes. To decode a codeblock, EBCOT always starts from the most significant bitplanes, and then moves towards the least significant ones. The pressed information of every codeblock is then arranged in several quality layers, to create a scalable pressed bitstream. Conceptually, each quality layer monotonically increases the knowledge of samples magnitudes, . increases the quality of the reconstructed image. Formally, EBCOT is made of three main steps, namely Significance Propagation (SP), Magnitude Refinement (MR), and Clean Up (CL). Each of the above steps can resort to four decoding primitives, namely Zero Coding, Sign Coding, Magnitude Refinement Coding, and Run Length Coding. The bitplane visiting order follows the sequence SP 中北大学 2020 届毕业设计说明书 第 6 页 共 15 页 MRCL: it is worth noticing that every sample of a given codeblock is processed in just one of the three steps. As far as putational plexity is concerned, CL demand s the largest effort during the decoding of the most significant bitplanes. As SP steps are applied, an increasing number of samples bee significant, and are inserted in a list of MRready samples. Progressively, the load required by MR steps grows, making the decoder efficiency directly dependent on the MR and CL optimization level. During the development of the EBCOT decoder block, particular care has been posed on the design of agile data structures, particularly suited to DSP optimized C code of MR and CL steps. Uniform scalar dequantizer According to, the quantization method supported by JPEG2020 is called scalar uniform. Uniform scalar dequantization can be simply acplished by means of a single multiplication for each wavelet coefficient. Inverse wavelet transform The discrete wavelet transform can be evaluated by means of a convolutionbased kernel, or a liftingbased kernel, this latter being the default transform kernel employed in JPEG2020. It has been demonstrated that the lifting scheme may run up to twice as fast as convolution. The wavelet transform has to be performed on both image rows and columns, in order to obtain a separable twodimensional subband deposition: JPEG2020 performs first the columnwise, and then the rowwise filtering. The default filter used for lossy pression is the wellknown DB(9,7): since it does not have rational coefficients, particular care ought to be posed to the effects of fini te precision representation. Due to the use of a fixed point TI TMS320C6201 DSP, a detailed study of internal data representation has been performed. Experimental re sults shows that excellent perceptive quality can be achieved recurring to 9 fractional bits for filter coefficients. In order to optimize the dyna mic range around zero, a DCshift is foreseen by the standard, as the DC ponent could lead to an excessive growth of the dynamic range of lowpass subband coefficients. Moreover, the lowpass filters can keep the samples in a fixed range, provided that a unitary DC filter gain is guaranteed. The joint 中北大学 2020 届毕业设计说明书 第 7 页 共 15 页 effect of DC ponent suppression and unitary gain ensures that range requirements are fulfilled during the whole wavelet transform. Adaptive ReedSolomon packet de coding Adaptive ReedSolomom packet decoding The deinterleaving RS decoder has been mapped on the FPGA device。 it is split into two functional subblocks: the first is the deinterleaver, the second is the RS decoder. The former collects packets received from the work, filling the columns of a matrix, and transferring them, row by row, to the RS decoder core. The latter performs the decoding process in five calculation steps, namely syndromes, erasure locator polynomial, erasures evaluator polynomial, error values, and cor rection. The RS core has been designed and developed taking into account, as much as possible, the issue of modularity. Since RS codes are strongly based on Galois Fields (GF) inner operations,。