Spatio-Temporal Super-Resolution Land Cover Mapping Based on Fuzzy C-Means Clustering

Xiaohong Yang,Ming Zhong,Xiaodong Li,Yihang Zhang,Zhong Xie,Feng Ling

doi:10.3390/rs10081212

Abstract

Super-resolution land cover mapping (SRM) is a method that aims to generate land cover maps with fine spatial resolutions from the original coarse spatial resolution remotely sensed image. The accuracy of the resultant land cover map produced by existing SRM methods is often limited by the errors of fraction images and the uncertainty of spatial pattern models. To address these limitations in this study, we proposed a fuzzy c-means clustering (FCM)-based spatio-temporal SRM (FCM_STSRM) model that combines the spectral, spatial, and temporal information into a single objective function. The spectral term is constructed with the FCM criterion, the spatial term is constructed with the maximal spatial dependence principle, and the temporal term is characterized by the land cover transition probabilities in the bitemporal land cover maps. The performance of the proposed FCM_STSRM method is assessed using data simulated from the National Land Cover Database dataset and real Landsat images. Results of the two experiments show that the proposed FCM_STSRM method can decrease the influence of fraction errors by directly using the original images as the input and the spatial pattern uncertainty by inheriting land cover information from the existing fine resolution land cover map. Compared with the hard classification and FCM_SRM method applied to mono-temporal images, the proposed FCM_STSRM method produced fine resolution land cover maps with high accuracy, thus showing the efficiency and potential of the novel approach for producing fine spatial resolution maps from coarse resolution remotely sensed images.

Highlights

Land cover and its change are one of the most prominent landscape signs to record the characteristic and dynamic processes of the Earth’s surface [1,2]
A visual comparison of the results shown in Figure 6 indicates that the proposed FCM_STSRM method prevailed over the two other methods at all zoom factors
The proposed FCM_STSRM method is an image-based Super-resolution land cover mapping (SRM) method that can be directly applied to original coarse spatial resolution multispectral images

Summary

Introduction

Land cover and its change are one of the most prominent landscape signs to record the characteristic and dynamic processes of the Earth’s surface [1,2]. A land cover map is produced by the hard classification method, which assigns each pixel in the remotely sensed image to a certain land cover class. The remotely sensed image pixels are often mixed with several different land cover classes, the pixels located in the land cover patch boundaries in coarse spatial resolution images. Pixel based hard classification methods always ignore the mixed pixel phenomena and only use the statistical characteristics of the pixel spectra to classify a mixed pixel as a specific land cover type, such that the resultant land cover map cannot accurately represent the land cover information. Soft classification can extract the area percentage of each land cover class in mixed pixels and obtain more accurate land cover information than hard classification. Soft classification does not give the spatial distribution of land cover classes at the subpixel level and the spatial resolution of classification result does not really improved

Methods

Results

Discussion

Conclusion