Abstract
Detailed information on the spatial-temporal change of impervious surfaces is important for quantifying the effects of rapid urbanization. Free access of the Landsat archive provides new opportunities for impervious surface mapping with fine spatial and temporal resolution. To improve the classification accuracy, a temporal consistency (TC) model may be applied on the original classification results of Landsat time-series datasets. However, existing TC models only use class labels, and ignore the uncertainty of classification during the process. In this study, an uncertainty-based spatial-temporal consistency (USTC) model was proposed to improve the accuracy of the long time series of impervious surface classifications. In contrast to existing TC methods, the proposed USTC model integrates classification uncertainty with the spatial-temporal context information to better describe the spatial-temporal consistency for the long time-series datasets. The proposed USTC model was used to obtain an annual map of impervious surfaces in Wuhan city with Landsat Thematic Mapper (TM), Enhanced Thematic Mapper (ETM+), and Operational Land Imager (OLI) images from 1987 to 2016. The impervious surfaces mapped by the proposed USTC model were compared with those produced by the support vector machine (SVM) classifier and the TC model. The accuracy comparison of these results indicated that the proposed USTC model had the best performance in terms of classification accuracy. The increase of overall accuracy was about 4.23% compared with the SVM classifier, and about 1.79% compared with the TC model, which indicates the effectiveness of the proposed USTC model in mapping impervious surfaces from long-term Landsat sensor imagery.
Highlights
Impervious surfaces are mainly artificial structures that do not allow water to infiltrate into the ground surface
These subareas (i.e., Jiangxia, Huangpi, and Xinzhou districts) represent three different cases of classification results derived by the support vector machine (SVM) classifier; that is, the overestimation of impervious surface, the underestimation of impervious surfaces, and the invalid area where the class labels were not determined
The comparison of three classification results derived by the SVM classifier, the temporal consistency (TC) model, and the uncertainty-based spatial-temporal consistency (USTC) model was shown visually (Figures 4–6) and numerically (Table 2)
Summary
Impervious surfaces are mainly artificial structures that do not allow water to infiltrate into the ground surface. They include rooftops, driveways, sidewalks, parking lots, and roads, and have been recognized as a key indicator of urban environments [1]. Impervious surfaces have been successfully monitored with remotely-sensed imagery over a wide range of spatial scales and using a variety of data sources [8]. Among all of the remotely sensed imagery used for impervious surfaces mapping, Landsat sensor data are perhaps the most popular because of the unparalleled temporal span together with its relatively fine spatial resolution and spectral coverage
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
