摘要高光谱遥感技术在环境监测、农业生产、地质勘探等领域应用广泛。高光谱遥感具有很高的光谱分辨率,能够有效识别具有诊断性光谱特征的地物。利用高光谱遥感图像,通过光谱分析技术进行地物识别是目前遥感信息处理领域的研究热点。但是高光谱图像波段多、数据量大,已有的识别算法往往运行效率较低,因此有必要利用高性能计算技术进行算法的并行优化研究,提高算法的处理效率。33016
本文基于光谱分析技术研究了高光谱遥感图像地物识别算法,分析了高光谱遥感探测和地物识别的相关概念和技术原理,基于GPU(Graphics Processing Unit)和CUDA(Compute Unified Device Architecture)架构设计了高光谱遥感地物识别快速处理流程和方法,充分利用GPU的并行计算能力优化PCA(Principal Component Analysis)、PPI(Pixel Purity Index)、SAM(Spectral Angle Mapping)等算法的设计与实现,并基于Visual C++开发了相应的软件系统。实验结果表明,本文的方法和系统有效提高了遥感地物识别的精度和效率,达到并行优化的目的。
关键词 高光谱遥感 地物识别 GPU CUDA 并行优化
毕业论文设计说明书外文摘要
Title Parallel optimization and system implementation of Hyperspectral Remote Sensing object recognition
Abstract
Hyperspectral remote sensing technology is widely used in the field of environmental monitoring, agricultural production, geological exploration. Hyperspectral remote sensing has a high spectral resolution, which can effectively identify the features of the diagnostic spectral features. Using hyperspectral remote sensing image, through spectral analysis technology for ground object recognition is currently a hot spot in the field of remote sensing information processing. But for hyperspectral image,there is a lot of bands and a large amount of data, the efficiency of existing recognition algorithms is low. Therefore, it is necessary to use high performance computing technology to research algorithm parallel optimization study, improving the processing efficiency of the algorithm.
In this paper, we analyzes the hyperspectral remote sensing image recognition algorithm based on spectral, analyzes the concepts and technical principle of hyperspectral remote sensing detection and object recognition, based on GPU (graphics processing unit) and CUDA (Compute Unified Device Architecture) architecture design of hyperspectral image recognition fast processing flow and method, take full advantage of GPU parallel computing ability of optimized PCA(Principal Component Analysis)、PPI(Pixel Purity Index)、SAM(Spectral Angle Mapping)algorithm design and implementation.Experimental results show that the method and system of this paper can effectively improve the accuracy and efficiency of remote sensing ground object recognition, and achieve the purpose of parallel optimization.
Keywords Hyperspectral remote sensing objects recognition
GPU_CUDA Parallel optimization
目 次
1 引言 1
1.1研究背景 1
1.2国内外研究现状 1
1.2.1 高光谱遥感技术的研究现状 2 高光谱遥感地物识别并行优化和系统实现:http://www.751com.cn/jisuanji/lunwen_29855.html