基于K-Means的遥感图像分类及其传输系统的研究

发布时间：2018-03-17 07:46

本文选题：无线网络传输系统　切入点：小波变换　出处：《北京邮电大学》2017年硕士论文　论文类型：学位论文

【摘要】：为了保证信息可以不受限于网络、信号、安全等因素进行快速传输,我们与军方合作,设计并开发了一套无线网络传输系统。这套系统通过FPGA射频信号传输加密后的信息,不会受到网络、信号、安全等因素的干扰,不仅可以进行正常通信,还可以用来进行野外救援、作战指挥等。当前系统主要由业务平台和无线平台两个部分组成。基于这个无线网络传输系统,本文研究了遥感图像去噪过程和遥感图像分类过程,将分类后的遥感图像通过本系统进行传输,实现了遥感图像分类传输系统。本文的主要研究内容包括以下四个方面:1)无线网络传输系统中业务平台的设计与实现,无线平台和业务平台间的通信协议SWIP协议的设计,无线平台之间信息传输的实现与传输过程的说明。2)遥感图像包含的噪声类型分析,针对遥感图像中包含的噪声类型,选取了中值滤波和小波阈值去噪法进行去噪处理,分析了传统小波阈值去噪法存在的不足,提出了一种双阈值小波阈值去噪函数。3)分析并比较常用的图像分类算法,选取K均值聚类算法对遥感图像进行分类,针对传统K均值聚类算法在遥感图像分类过程中存在的问题,提出了一种自适应确定分类数并优化初始聚类中心的K均值聚类算法。4)无线网络传输系统应用于遥感图像分类中,实现了遥感图像分类传输系统。本文通过Matlab仿真实验,将峰值信噪比作为图像去噪效果的客观评价标准,对比了传统小波阚值去噪算法与改进的小波阈值去噪算法的去噪效果,实验表明改进的小波阈值去噪法对遥感图像去噪效果更佳。同样的,通过对比实验发现,改进的K均值聚类算法对遥感图像的分类效果更佳。
[Abstract]:In order to ensure that the information can be transmitted quickly without limiting the network, signal, security and other factors, we have designed and developed a wireless network transmission system in cooperation with the military. This system transmits encrypted information through FPGA radio frequency signal. Without interference from network, signal, security and other factors, not only can normal communication be carried out, but also can be used for field rescue. The current system is mainly composed of two parts: service platform and wireless platform. Based on this wireless network transmission system, the process of remote sensing image denoising and remote sensing image classification is studied in this paper. The classified remote sensing image is transmitted through this system, and the remote sensing image classification transmission system is realized. The main research contents of this paper include the following four aspects: 1) the design and implementation of the service platform in the wireless network transmission system. The design of communication protocol SWIP protocol between wireless platform and service platform, the realization of information transmission between wireless platforms and the description of transmission process. 2) the noise type analysis of remote sensing image, aiming at the noise type included in remote sensing image. The median filter and wavelet threshold denoising method are selected for denoising processing. The shortcomings of traditional wavelet threshold denoising method are analyzed, and a double threshold wavelet threshold denoising function .3) is proposed to analyze and compare the common image classification algorithm. The K-means clustering algorithm is selected to classify remote sensing images, and the problems existing in the traditional K-means clustering algorithm in the process of remote sensing image classification are pointed out. This paper presents a K-means clustering algorithm. 4) the wireless network transmission system is applied to remote sensing image classification. The remote sensing image classification and transmission system is realized. In this paper, the Matlab simulation experiment is carried out. The peak signal-to-noise ratio (PSNR) is taken as the objective evaluation criterion of image denoising effect, and the denoising effect of traditional wavelet threshold de-noising algorithm and improved wavelet threshold de-noising algorithm is compared. The experimental results show that the improved wavelet threshold denoising method is better for remote sensing image denoising. Similarly, it is found that the improved K-means clustering algorithm is better for the classification of remote sensing images.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP751

【相似文献】