The Research Of Image Fusion Technique Based On HJ1A/1B Satellite

In recently years, as the development of remote sensing technology, many earth satelliteswith different types of sensors were launched in succession, which provide a large amount ofdata resources for our resource investigation, environmental monitoring, disaster warning andso on. However, there exist big differences in spatial resolution, spectral characteristics andtime period of the remote sensing data. In order to use the data efficiently, it needs to removethe redundant information and integrate the advantages of these data.Remote sensing image fusion technology emerged in recent years, which has furtherapplication. Remote sensing image fusion is to fuse the images which are based on the samearea at different period and from different satellite sensors. The purpose of the remote imagefusion is to overcome the limitations of image in the spatial resolution and spectral resolution.The HJ-1satellite multispectral image and Landsat8satellite panchromatic image of thesame region were selected as the data source and remote sensing image fusion algorithm wasstudied in the article. Verification image is divided into two groups: the region of remoteimage is Dahuofang reservoir，Fushun and Hunnan new district of Shenyang. The principleand technological process of traditional fusion algorithm were analyzed. Basing on the fusionalgorithm by wavelet transform, one improved algorithm was proposed. The improvedalgorithm based on the high frequency component. Through the both subjective and objectivecomprehensive evaluation of the experimental images, we conclude that the improvedalgorithm is better than the original algorithm on the spectral characteristics and spatialinformation. Aimed at reducing the spectrum distortion degree of the existing wavelettransform combined with HSI fusion algorithm, this paper also puts forward an algorithm,which increases the amount of information and retain much more spectral informationcompared with original method.