Hyperspectral and high-resolution image fusion based on second generation Bandelet transform

Dimension reduction for hyperspectral image based on the second generation bandelet transform

“…The hyperspectral device used in this research is Image-λ-V10E-PS, as shown in Fig.1, and its basic parameter is shown in Tab The hyperspectral data are multiplexed and transmitted to the PC by a data line, after that, the data is submitted to the ancillary software for pre-processing (i.e., grayscale correction). All the pre-processing is necessary because of the systematic errors exist in this acquisition operations [13] . A hyperspectral data captured from the experiment which having …”

“…This famous transformation can track the geometric regular direction of images and the characteristics of images edge by selecting best Bandelet basis adaptively [13,15] . The original data can be reconstructed by geometry flows and Bandelet coefficients perfectly due to the reversibility of Bandelet transform.…”

“…To finish the intact algorithm, quadtree segmentation is applied in this algorithm [12] . Every band of the hyperspectral image is separated into four sub-bands in every level until the sub-band is lower than the threshold [13] . Besides, down-sample is introduced and applied in the algorithm.…”

Hybrid image denoising method based on non‐subsampled contourlet transform and bandelet transform

“…In recent years, various types of multi-scale geometric transforms have been proposed, such as the ridgelet transform [1], the curvelet transform [2], the contourlet transform (CT) [3,4], and the bandelet transform (BT) [5,6]. Further, scholars have proposed several image denoising methods based on these multi-scale geometric transforms to improve the quality of image denoising [7][8][9][10] and applied them in many applications [11][12][13][14].…”