Satellite Images Enhancement Using Super Resolution Wavelet and Interpolation Technique with Edge Extraction and Sparse Matrix
In this research letter a super-resolution algorithm using wavelet and Lanczos interpolation technique along with nearest neighbourhood interpolation and sparse matrix is proposed. A good quality of pixels is generated in satellite image by using this novel algorithm. The Sharper images can be generated by high frequency sub bands. The Interpolation method is used between satellite image and HF sub bands through the discrete wavelet transform which preserves edges. The Sparse image is obtained by applying sparse mixing weights to low resolution (LR) image. The Lanczos interpolation is a sinc filter which reduces arti-facts in an image. The High resolution (HR) image is generated by using inverse DWT to LH, HL, HH and modified LL sub band. The proposed method is carried out on Google earth satellite images. The quantitative parameters such as PSNR (peak signal to noise ratio), RMSE (root mean square value), CC (correlation coefficient), EME (Enhancement measurement) are measured for satellite images. The proposed technique i.e., super resolution using wavelet along interpolation and sparse matrix has got better values compared to DWT in terms of quantitative parameters.