The Application Of Beijing-1 Image Fusion And Its Information Extraction From Water Body

Quick and accurate extraction of water body information through satellite remote sensing images has become an important means for the investigation of water resources, macro-monitor of water resources, surveys of urban greening, and wetland protection. With the factors like high cost and long cycles of study, traditional sensors have restricted RS applications. Small satellites, taking advantage of low cost, short cycles of study, means flexibility, and prone launch, etc, has become the research of cutting-edge for the spacecraft development. Considering Beijing-1 small satellite as a study object, this paper emphasizes on theoretical analysis and experimental study with three aspects of fusion algorithm and its implementation over images, image quality evaluation after integration, as well as water information extraction with RS technology.Pixel-level fusion methods including HPF transform, color space method, PCA transform, wavelet transform, and wavelet-based IHS transform, etc., are carried out for summary, analysis and algorithm achievement. Different results of image fusion from Beijing-1 are obtained through these methods. Methods of subjective and objective evaluation are used for the fusion images to take single indicator quality evaluation. On the basis of the above-mentioned study, the introduction of artificial neural network of effective simulation is integrated on the single indicator. In addition, varieties of classification techniques on RS images are used to extract information of water body after Beijing-1 image fusion.In accordance with complementary of evaluation methods for the image fusion quality, a complicated evaluation system on RS image quality is built up including three evaluation processes, five evaluation indicators, and an index complex calculation method.The experimental results show that 4-meter panchromatic images and 32-meter multi-spectral images from Beijing-1 can be fused. Maintaining the spectral information, variety of algorithms enables to improve the image resolution. However, due to the over-large scale of the resolution as well as the differences between fusion algorithms, the image quality is quite different after image fusion. After three levels of quality evaluation, the wavelet-based IHS transform algorithm is more suitable for Beijing-1 images than other algorithms.