Journal of Remote Sensing / 2021 / Article / Alg 1

Research Article

Shearlet-Based Structure-Aware Filtering for Hyperspectral and LiDAR Data Classification

Algorithm 1

ShearSAF for HSI and LiDAR feature extraction and classification.
1: INPUT: raw HSI data , LiDAR data ;
2: OUTPUT: the classification map ;
3: BEGIN
4: , , , , ;
5: using SNIC on to obtain the over-segmentation superpixel map with regions;
6: using KPCA on to obtain the information-focused HSI data ;
7: using equation (22) and (24) to obtain the texture and area information , , and ;
8: using equation (33) to calculate the merging cost ;
9: conducting progressive region merging procedure on and obtain the final merging map , where the number of regions in is calculated by equation (39);
10: using equation (34) to obtain distance map ;
11: using equations (35) and (36) to obtain scale map ;
12: using equation (37) to obtain structure-aware filter for each spatial pixel ;
13: using equation (38) to obtain convolution feature from and ;
14: using RF classifier on to achieve classification map ;
15: END