A Spatial Resolution Enhancement Method For Cascading Hyperspectral Remote Sensing Images

In order to obtain high spectral resolution,hyperspectral imaging system design must ensure that the photosensitive element having a larger field of view,so as to ensure adequate imaging can instantly amount of light and the integration time,but this will inevitably lead to image spatial resolution decreases.For this "inherent"technical problems of hyperspectral data,a wide range of remote sensing image fusion approach was used to enhance the spatial resolution of hyperspectral remote sensing images,the nature of which was based on the principle of PCA fusion because it very meet the high dimensionality of hyperspectral data.However,when the existence of the current PCA fusion method is used,there are two problems:First,PCA image fusion technology need to get the high spatial resolution imaging at the same scene.But in most cases,because of technology and economic constraints,we can't get high-resolution high-resolution images.Secondly,due to the first principal component contains the main spatial structure and spectral information of the hyperspectral image,so when it is replaced by the high spatial resolution image,the spectral distortion occurs.For these two issues,this paper PCA study image fusion technology,super-resolution image reconstruction techniques,and proposes a cascade hyperspectral remote sensing image fusion method.The main contents of this paper are:(1)The PCA fusion algorithms and principles are detailed derived,and experiments of PCA imaging fusion were taken by use of Rapid Eye and Hyperion hyperspectral imaging data.The existing problems of the algorithm were analyzed;(2)We studied the image super-resolution reconstruction of compressed sensing framework,and made the Matlab programming based on discrete cosine transform matrix and compressed sensing single image super-resolution reconstruction algorithm.At last,we made the simulation and evaluation for experimental results;(3)we proposed cascade hyperspectral remote sensing image fusion method using PCA fusion algorithm and compression super-resolution image reconstruction algorithm,give full play to each other's strengths,completed a single super-resolution reconstruction based only one low-resolution hyperspectral remote sensing images.Through experimentation and evaluation,this method can prove sufficient to ensure that the original hyperspectral imaging spectral information is not lost and its spatial resolution enhancement.