遥感图像质量评价方法研究与软件设计

发布时间：2017-12-28 09:38

本文关键词：遥感图像质量评价方法研究与软件设计　出处：《浙江大学》2017年硕士论文　论文类型：学位论文

【摘要】：遥感技术的应用越来越广,在农业、水利、军事等各方面都有广泛的应用。遥感图像的质量直接影响到各领域的监测、预估或侦察的准确性,另外遥感图像质量的优劣还可以反映遥感成像系统的性能,因此遥感图像质量评价对于遥感应用方面起着非常重要的作用。本文主要从以下几个方面展开研究:1)遥感图像的降质因素研究。从遥感图像的成像原理和成像链路系统展开讨论,分析了几种主要导致图像降质的影响因素,表现为噪声、模糊和云覆盖。2)遥感图像的单一质量评价方法研究。针对模糊度评价,深入分析了基于图像直方图集中度的算法和基于灰度梯度共生矩阵的算法;对于噪声评价,深入研究了基于局部标准差的算法和基于小波域统计特性的算法;对于云检测,深入分析了基于模糊C均值的聚类算法与基于多阈值的算法。针对模糊度、噪声和云检测的各个算法进行了仿真实验,通过实验结果的比较,分析了算法的优缺点和适用场合。3)遥感图像的多失真评价方法研究。针对模糊度和噪声的评价算法的抗干扰性能不足,提出了适合多失真的评价算法,同时提取较直观的空间域和多尺度的频率域自然场景统计特征,计算高斯概率密度分布函数的主要参数作为特征值,利用支持向量回归得到综合评价值。4)遥感图像质量评价软件设计。给出了遥感图像质量评价软件的总体架构和总体流程,并且完成了模糊度评价、噪声评价,云检测模块和综合判决模块。详细介绍了这些模块的主要算法及算法流程,介绍了遥感图像评价软件的界面操作,展示了算法运行结果。
[Abstract]:The application of remote sensing technology is more and more widely used in agriculture, water conservancy, military and other aspects. The quality of remote sensing image directly affects the accuracy of monitoring, prediction or reconnaissance in all fields. In addition, the quality of remote sensing image can also reflect the performance of remote sensing imaging system. Therefore, the quality evaluation of remote sensing image plays a very important role in remote sensing applications. This paper mainly studies from the following aspects: 1) the study of the degradation factors of remote sensing images. From the theory of remote sensing image and the link system of imaging, the factors that cause the degradation of image quality are analyzed, including noise, blur and cloud coverage. 2) study on the single quality evaluation method of remote sensing image. According to the fuzzy evaluation, in-depth analysis of the image histogram concentration algorithm and gray gradient co-occurrence matrix algorithm based on noise; for evaluation, in-depth study of the local standard deviation and the algorithm based on wavelet domain algorithm based on statistics; for cloud detection, in-depth analysis of the clustering algorithm based on fuzzy C means based on multi threshold the algorithm. The simulation experiments are carried out for the algorithms of fuzzy degree, noise and cloud detection. The advantages and disadvantages of the algorithm and the applicable situation are analyzed through the comparison of the experimental results. 3) research on the multi distortion evaluation method of remote sensing image. Aiming at the shortage of the anti-jamming performance evaluation of ambiguity and noise algorithm, is proposed for evaluating distortion algorithm, frequency domain feature extraction and natural scene statistics in spatial domain more intuitive and multiscale, calculate the main parameters of Gauss probability density distribution function as a feature, using support vector regression to get the comprehensive evaluation value. 4) the design of remote sensing image quality evaluation software. The overall framework and overall process of remote sensing image quality evaluation software are given, and fuzzy evaluation, noise evaluation, cloud detection module and comprehensive decision module are completed. The main algorithm and algorithm flow of these modules are introduced in detail, and the interface operation of remote sensing image evaluation software is introduced, and the operation results of the algorithm are displayed.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP751

【相似文献】