The Research On The Wavelet Interpolation Algorithms Which Are Applied In Remote Sensing Image

With the advancing of remote sensing technology, the carrier of remote sensing device has developed from camera to satellite, astro-ship, or space shuttle, which can provide high resolution data of earth fast with multiply platform. The remote sensing has been put into good use in agriculture, industry, trade and military affairs. At the same time , some practical application problems have emerged, such as a great deal of data, broad area detection with high resolution, and so on. Remote sensing image interpolation will be beneficial to solving those problems.The remote sensing image, which is always of bigger size, plentiful information, and higher entropy, requires that the interpolation algorithm can preserve the information of source image as much as possible, increase the image resolution, and be appropriate for people to observe, and so forth. Firstly, the paper overviewed the requirements of the former interpolation algorithms to the interpolation function, several interpolation algorithms, then evaluated the interpolation effect from three aspects, which are Fourier analysis, the quality of interpolation and the computation complexity. However, those interpolation algorithms all have flat effects and don't take the global correlation into consideration. The wavelet transform processes the high frequence information and the low frequence information simultaneously, which is like the human visual system. For satisfying the requirements of remote sensing image interpolation, taking the characteristic of wavelet transform into account, this paper presented wavelet bilinear interpolation algorithm. The fractal interpolation algorithm can preserve the texture information of image. So, this paper use the fractal interpolation algorithm as resemble transform based on the wavelet transform(named wavelet fractal interpolation algorithm). It is found that the two algorithms both increased SNR, level resolution and plumb resolution compared with former interpolation algorithms, for example, bilinear interpolation. Moreover, they are appropriate for human to observe.This paper also discussed the theory of wavelet transform including the theory of filter banks. These theories are the base of the two algorithms, wavelet bilinear interpolation algorithm and wavelet fractal interpolation algorithm, and the precondition of wavelet construction. For speeding up the wavelet bilinear interpolation algorithm, taking the characteristic of orthonormal symmetry wavelet into account, This paper speeded up the wavelet bilinear interpolation algorithm about 43% by polyphase matrix.There are still some places which need improvement. For example, speed upthe wavelet transform more, construct a new wavelet which is appropriate to remote sensing image, and so on.