Digital Remote Sensing Image Analysis: Enhancement, Compression, Transformation, Classification, and Target Detection
| Autor: | Meng-Jye Hu, 胡孟杰 |
|---|---|
| Rok vydání: | 2004 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 92 In this paper, we discuss several applications of remote sensing image analysis. Unlike the traditional color image, remote sensing image is acquired by the sensor comprises a number of bands, each of which represents the intensity of an imaged scene that is received by a sensor at a particular wavelength. In other words, it is either a multispectral or a hyperspectral image. The applications of remote sensing image we discussed in this paper including image enhancement technology, spectral domain transformation, image compression, classification, spectral unmixing, and subpixel target detection. Our discussion is focus on the unusual and unintuitive characteristics of hyperspectral image comparing with multispectral image for data compression and classification, the application of spectral domain linear transformations, especially the segmented based transformation, and the performance of the different target detection algorithms. The experiment results shows that spectral domain linear transformation can be used to solve the problem that is caused by the high dimensional data set and the segmented based transformation can greatly improve the compression ratio and classification accuracy of hyperspectral imagery. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|