高光谱图像压缩的方法研究

发布时间：2018-04-18 15:51

本文选题：高光谱图像 + 图像压缩　；参考：《南京理工大学》2014年硕士论文

【摘要】：作为一种融合了电磁学、光学、信号处理等多学科交叉领域的新兴学科,高光谱遥感成像已经被广泛的应用于军事、地质、医学、农业、海洋生物、大气和环境监测等领域。然而高光谱图像庞大的数据量给存储和传输带来了巨大的压力,严重制约了高光谱成像技术的应用发展,因此找到一种高效的压缩方法是高光谱遥感成像技术发展中一个亟须要解决的课题。本文对高光谱图像压缩的方法进行了研究。首先,对高光谱图像数据的存储格式以及空间和谱间相关性进行了研究。其次,对基于线性预测的高光谱图像压缩的理论进行了详细分析,并对光谱图像数据的编码方法进行了介绍。在对高光谱图像进行压缩之前,通过采用一些预处理手段来提升整个高光谱图像压缩的性能,结合高光谱图像本身的特性,提出了一种基于自适应波段分组和小波变换混合优化编码的高光谱图像压缩算法,通过实验结果比较发现,该算法可以实现高光谱图像的高效压缩。然后,对基于压缩感知理论的高光谱图像压缩方法进行了研究。介绍了压缩感知的理论,分析了压缩感知理论的关键要素,并对不同的重构算法进行了压缩实验以及相应的结果分析。最后,对基于编码孔径成像技术的高光谱图像压缩方法进行了分析研究。介绍了编码孔径光谱成像技术的原理,通过建立的编码孔径成像技术理论模型对高光谱图像进行了实验研究以及结果分析。
[Abstract]:Hyperspectral remote sensing imaging has been widely used in military, geological, medical, agricultural, marine life, atmospheric and environmental monitoring and other fields, as a new interdisciplinary subject, such as electromagnetism, optics, signal processing and so on, and hyperspectral remote sensing imaging has been widely used in military, geology, medicine, agriculture, marine life, atmosphere and environment monitoring and so on.However, the huge amount of data of hyperspectral image brings great pressure to storage and transmission, which seriously restricts the development of hyperspectral imaging technology.Therefore, finding an efficient compression method is an urgent problem in the development of hyperspectral remote sensing imaging technology.The method of hyperspectral image compression is studied in this paper.Firstly, the storage format and spatial and spectral correlation of hyperspectral image data are studied.Secondly, the theory of hyperspectral image compression based on linear prediction is analyzed in detail, and the coding method of spectral image data is introduced.Before compressing the hyperspectral image, the compression performance of the whole hyperspectral image can be improved by using some preprocessing methods, combining with the characteristics of the hyperspectral image itself.A hyperspectral image compression algorithm based on adaptive band grouping and wavelet transform hybrid optimization coding is proposed. The experimental results show that this algorithm can efficiently compress hyperspectral images.Then, the hyperspectral image compression method based on compression sensing theory is studied.This paper introduces the theory of compression perception, analyzes the key elements of the theory of compression perception, and makes compression experiments on different reconstruction algorithms and analyses the corresponding results.Finally, the hyperspectral image compression method based on coded aperture imaging is analyzed and studied.The principle of coded aperture spectral imaging is introduced. The experimental study and analysis of the results of hyperspectral images are carried out through the theoretical model of coded aperture imaging.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP391.41

【参考文献】