项目名称: 基于矢量非局部均值模型的PolSAR相干斑抑制技术及其效果评估体系研究
项目编号: No.61302160
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 无线电电子学、电信技术
项目作者: 牛朝阳
作者单位: 中国人民解放军信息工程大学
项目金额: 24万元
中文摘要: 相干斑不仅降低了PolSAR图像质量,更阻碍着目标极化信息的精确提取,所以如何有效抑制相干斑并对其效果进行科学评估,已经成为PolSAR信息处理技术的研究热点之一。针对该方向已有研究工作存在的若干关键问题,本项目重点开展以下三个方面内容的研究:1)对标量非局部均值理论进行矢量化扩展,结合PolSAR的多维数据形式建立矢量贝叶斯非局部均值模型,丰富非局部均值理论的思想内涵,突破非局部均值理论在PolSAR相干斑抑制中应用的理论基础;2)基于中低、高分辨率PolSAR数据的统计特性推导相似性度量函数,将矢量贝叶斯非局部均值模型分别与乘性噪声模型和多维噪声模型相结合,研究不同分辨率PolSAR相干斑抑制的新算法,有效提升PolSAR相干斑抑制技术水平;3)从三个角度探索新评估指标和改进评估方法,构建面向极化散射特性的评估方案,为PolSAR相干斑抑制效果评估提供更为合理的参考依据。
中文关键词: 极化合成孔径雷达;矢量非局部均值;相干斑抑制;效果评估;极化散射特性
英文摘要: Speckle does not only reduce the quality of PolSAR images, but also obstruct the accurate extraction of target polarization information, so suppressing speckle and scientifically evaluating its effectiveness, has become one research focus of PolSAR information processing technology. To address some key issues of existing research work, the project focuses on the following three aspects: 1) expands the nonlocal means model from scalar into vector, and establishes the vectorial form of Bayesian nonlocal means model with PolSAR data vector, which can rich ideological connotations of the nonlocal means theoretical and provide theoretical basis for applications of nonlocal means theory in PolSAR despeckling; 2) derives the similarity measure function based on statistical properties of PolSAR data of low and high resolution, combines vectorial Bayesian nonlocal means model with the multiplicative noise model and the multi-dimensional noise model respectively, and studies new despeckling algorithms for PolSAR data of different resolutions, in order to effectively develop the PolSAR despeckling technology; 3) explores new assessment indicators and improves assessment programs from three aspects to building a new evaluation system based on polarization scattering characteristics, which will provide the effect evaluation
英文关键词: Quad-polarization Synthetic Aperture Radar;Vectoral Non-local Means;Speckle Reduction;Effect Evaluation;Polarimetric Scattering Characteristic