Land Cover Change Mapping at the Subpixel Scale With Different Spatial-Resolution Remotely Sensed Imagery
Land Cover Change Mapping at the Subpixel Scale With Different Spatial-Resolution Remotely Sensed Imagery
复制标题
DOI：
10.1109/lgrs.2010.2055034
复制
发表时间：
2011
影响因子：
4.8
通讯作者：
F. Ling;Wenbo Li;Yun Du;Xiaodong Li
中科院分区：
文献类型：
--
作者：
F. Ling;Wenbo Li;Yun Du;Xiaodong Li
当使用粗分辨率遥感图像进行变化检测时，提取子像素尺度的土地覆盖变化（LCC）信息非常重要。虽然源自软分类技术的分数图像可用于子像素 LCC 检测，但无法提供每个粗分辨率像素内变化的子像素的空间分布。本文提出了一种亚像素 LCC 映射（SLCCM）算法，旨在通过比较前者的高分辨率土地覆盖图和后者源自粗分辨率图像的分数图像，预测双时图像之间亚像素尺度的 LCC 空间模式。所得到的子像素LCC图由空间相关原理和每个混合像素中的LCC规则确定。利用模拟图像和真实图像对所提出的算法进行了评估，结果表明了所提出的方法对于SLCCM的有效性。
Extracting land cover change (LCC) information at the subpixel scale is important when coarse-resolution remotely sensed images are used for change detection. Although fraction images derived from soft-classification technologies can be used for subpixel LCC detection, the spatial distribution of changed subpixels within each coarse-resolution pixel cannot be provided. This letter presents a subpixel LCC mapping (SLCCM) algorithm, aiming to predict the spatial pattern of LCC at the subpixel scale between bitemporal images through comparing the former high-resolution land cover map and the latter fraction images derived from the coarse-resolution image. The resulting subpixel LCC map is determined by the spatial dependence principle and an LCC rule in each mixed pixel. The proposed algorithm was evaluated with simulated and real images, and the results showed the effectiveness of the proposed method for SLCCM.