项目名称: 基于方向变换与稀疏优化的遥感图像压缩理论与方法
项目编号: No.61201327
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 电子学与信息系统
项目作者: 陈波
作者单位: 中国人民解放军国防科学技术大学
项目金额: 27万元
中文摘要: 基于方向变换的遥感图像压缩具有高保真的优势,但存在方向变换的相关理论仍不完备,方向变换计算复杂度较高的缺陷。针对这些问题,本课题利用稀疏优化可高效、准确表示图像几何正则性的优势,主要研究: (1)方向变换新机理与方向滤波器构造,(2) 针对遥感图像压缩的方向变换系数的快速稀疏优化方法,(3)方向变换系数编码算法以及低复杂度、低存储实现技术,力图解决现有星载压缩中方向性理论的缺陷以及复杂度高的问题,为解决星载图像高质量实时传输瓶颈提供理论与技术支撑。项目成果还可以进一步拓展到各类具有几何正则特性的高维信息的压缩及其它处理过程中。
中文关键词: 双树复数小波;图像压缩;方向变换;稀疏性;
英文摘要: Remote images compressed by the directional transform based method have less distortion, but the theory on on the directional transform is still not perfect and the transform costs too high computation. To solve these problem, and noticed the advantage of sparse optimizing, that it can present the geometric regular in images,we plan to study on, (1) the new directional mechanism of the transform and the construction of the directional filter, (2) fast saprse optimizing of directional transform coefficents of remote sensing images, (3) the coding of directional transform coefficents and the implementation with low complexity and low memory. The purpose is to propose a new directional transform theory and the corresponding feasible remote sensing image on-board compression method, which can be further used to process any high dimension data with geometric regular.
英文关键词: dual tree complex wavelet;image compression;directional transform;sparsity;