Utilizing Multiple Subpixel Shifted Images in Subpixel Mapping With Image Interpolation

被引：32

作者：

Wang, Qunming ^{[1
]}

Shi, Wenzhong ^{[1
]}

机构：

[1] Hong Kong Polytech Univ, Dept Land Surveying & Geoinformat, Kowloon, Hong Kong, Peoples R China

来源：

IEEE GEOSCIENCE AND REMOTE SENSING LETTERS | 2014年 / 11卷 / 04期

关键词：

Image interpolation; remote sensing; subpixel mapping (SPM); subpixel shifted images; superresolution mapping; MARKOV-RANDOM-FIELD; IDENTIFICATION; CLASSIFICATION;

D O I：

10.1109/LGRS.2013.2279138

中图分类号：

P3 [地球物理学]; P59 [地球化学];

学科分类号：

0708 ; 070902 ;

摘要：

In this letter, multiple subpixel shifted images (MSIs) were utilized to increase the accuracy of subpixel mapping (SPM), based on the fast bilinear and bicubic interpolation. First, each coarse spatial resolution image of MSI is soft classified to obtain class fraction images. Using bilinear or bicubic interpolation, all fraction images of MSI are upsampled to the desired fine spatial resolution. The multiple fine spatial resolution images for each class are then integrated. Finally, the integrated fine spatial resolution images are used to allocate hard class labels to subpixels. Experiments on two remote sensing images showed that, with MSI, both bilinear and bicubic interpolation-based SPMs are more accurate. The new methods are fast and do not need any prior spatial structure information.

引用

页码：798 / 802

页数：5

共 25 条

[1] Markov-random-field-based super-resolution mapping for identification of urban trees in VHR images [J].