Image Decompression Process Optimization Based On JPEG2000 Technology On GPU Research And System Design

In the aerospace field,the amount of data transmitted is large and real-time requirements are high.Image compression on the aircraft has certain practical significance.In the aerospace field,due to the limited amount of data that can be cached by aerospace equipment such as aircraft,in addition to the high compression performance requirements,strict requirements are imposed on the compression rate.At present,JPEG2000 image compression standard is widely used in various fields of study for its image compressing algorithms at home and abroad.JPEG2000 algorithm has a certain degree of parallelism.It can use multi-core CPU to achieve acceleration on CPU.However,in most cases,because of space resources and power consumption requirements,using more CPU cores can no longer meet more realistic requirements.In recent years,with the continuous development of the GPU and the continuous improvement of the CUDA programming framework,the GPU is used to implement the JPEG2000 image compression standard,which resolves the conflict between the computational capabilities and space resource limitations and power consumption.The hardware structure of the GPU uses a large number of integrated transistors as the operation unit and also has good parallelism.This paper aims to implement the JPEG2000 image compression standard on GPU.It studies the GPU's parallel computing capabilities,improves the basic swarm intelligence optimization algorithm,and accelerates experiments on GPUs to verify the parallelism of GPU in matrix operations.Effects,and then implemented the JPEG2000 image compression algorithm in the development environment of the VS2013 equipped with the Qt design library,and developed a software platform to understand pressure and fast vision,and achieved decompression and restoration of compressed data.The specific research content of this article includes:(1)GPU hardware architecture and programming framework.First of all,the parallel processing technology of GPU is studied.The current GPU architecture on the market is compared.The advantages and disadvantages of GPU and CPU are analyzed.The specific implementation methods of CPU and GPU work together are studied.Finally,the understanding is profound.CUDA programming framework and the implementation of the programming framework.Provides a theoretical basis for implementing parallel operations on the GPU below.(2)Implementation of swarm intelligence optimization algorithm on GPU.For the characteristics that the particle swarm optimization algorithm is fast and easy to mature,and the bacterial connoisseur algorithm has slow convergence and high precision,the algorithm is improved by combining the advantages of the two algorithms,and the improved algorithm is verified by the classic test function.Advantages: Finally,using GPU-rich hardware resources,the three algorithms are accelerated on the GPU,and the advantages of the GPU in matrix operations are also verified.(3)Parallel acceleration of JPEG2000 image compression on the GPU.Firstly,the method of compressing and decompressing JPEG2000 images is realized by using the hardware resource of GPU.Then the result is compared with the result of CPU and ADV212,and the accelerated performance of image compression on GPU is verified,and parallel optimization is performed.(4)JPEG2000 image decompression and fast-viewing software platform design.In order to verify the results of GPU decompression and test the ADV212 decompression card performance,the JPEG2000 image decompression and fast-viewing software platform was developed.The software platform supports image decompression and data storage on the GPU,and supports the ADV212 decompression board.The configuration and operation of the card and the fast-viewing function of the compressed image enable the testing of the parallel acceleration of the GPU and the interface interconnection of other devices in the satellite system.The research results of this paper have practical significance for on-site compression and decompression of aerospace system data.