| With the development of the remote technology, multispectral imaging technology has been increasingly wide application. Multispectal image is one of the main remote sensing data., which provides two-dimensional geographical spectrum in the intensity , and has high spectral resolutions. However, it is acquired at the price of huge amounts of data. This produces many difficulities in data transportion and storing, so it is very necessary to find a highly effective compression method.Multispectral image has both spatial and spectral redundance. Few in the spectral bands, the overall performance is not evident among spetctral redundance, so some highly effective static image compression methods can be used. When the spectral bands is more, the performance is very distinct in spatial and spectral redundance, spectral transformation can be use in conjunction with 3D encodement methods to compress the multispectral image.Based on the aboving characteristics, we proposes two algorithem: improved SPIHT encoding algorithem and 3D SPIHT encoding algorithem based on histogram transform. The former improve the SPIHT according to its requirement of high memory and long processing time. The latter proposes histogram transform first with the analysis of the multispectral images histogram, then remove the spectral redundance with KLT transform, in the last encode with 3D SPIHT algorithem, experiments show algorithms have better compression performance. In the cameras project joined, multispectral images bands is few, this paper uses the first algorithm in the process of DSP hardware realization, designs the FPGA experiments board to simulation CCD data sources, and achieve a real-time compressed request with better image quality and compression performance.The work of this article has the important function to the aircraft-borne camera recording data compression coding system research, it not only gains the conclusion, but also put forward the beneficial proposal to further work. |
PDF Full Text Request