Ngo, Duc Khanh
(2024)
Land cover mapping of the Mekong Delta with sentinel-1 synthetic aperture radar.
PhD thesis, University of Nottingham.
Abstract
Synthetic aperture radar (SAR) has great potential for land cover/land use (LCLU) mapping, especially in tropical regions, where frequent cloud cover obstructs optical remote sensing. The use of SAR data derived mapping results plays crucial role in urban and suburban extents characterizations, urban services, rice crop distribution delineation, and land use changes detection. As the Mekong Delta is a significant location ecologically, economically, and socially, food security, forest conservation, natural resource management, and urbanization are a matter of great concern. Urban expansion and conversion wetland areas to aquaculture have impacts on natural forest and coastal ecosystems in the Mekong Delta. Therefore, the use of latest Sentinel-1 C-band SAR data characterizing LCLU including urban expansion, aquaculture development, and productive land and unproductive lands is essential for natural resource management and land use planning.
This thesis demonstrated the use of Sentinel-1 SAR data and Google Earth Engine to map the LCLU of the Mekong Delta. The research in this thesis is divided into three parts: 1) the classification of multi-temporal Sentinel-1A C-band SAR imagery for characterizing the LCLU to support natural resource management; 2) identifying and mapping persistent building structures from coastal plains to high plateaus, as well as on the sea surface; 3) detecting and mapping persistent surface water and seasonal inundated LCLU.
Part 1 of the thesis investigated the classification of multi-temporal Sentinel-1A C-band SAR imagery for characterizing LCLU to support natural resource management for land use planning and monitoring. Twenty-one SAR images acquired in 2016 over Bạc Liêu province, a rapidly developing province of the Mekong Delta, Vietnam were classified. To reduce the effects of rainfall variation confounding the classification, the images were divided into two categories: dry season (Jan–April) and wet season (May–December) and three input image sets were produced: 1) a single-date composite image, 2) a multi-temporal composite image and 3) a multi-temporal and textural composite image. Support Vector Machines (SVM) and Random Forest (RF) classifiers were then applied to characterize urban, forest, aquaculture, and rice paddy field for the three input image sets. A combination of input images and classification algorithms was tested, and the mapping results showed that no matter the classification algorithms used, multi-temporal images had a higher overall classification accuracy than single-date images and that differences between classification algorithms were minimal. The results demonstrated the potential use of SAR as an up-to date complementary data source of land cover information for local authorities, to support their land use master plan and to monitor illegal land use changes.
Part 2 of the thesis developed novel and robust methods using time-series data acquired from Sentinel-1 C-band SAR to identify and map persistent building structures from coastal plains to high plateaus, as well as on the sea surface. Mapping building structures is crucial for environmental change and impact assessment and is especially important to accurately estimate fossil fuel CO2 emissions from human settlements. From annual composites of SAR data in the two-dimensional VV-VH polarization space, the VV-VH domain was determined for detecting building structures, whose persistence was defined based on the number of times that a pixel was identified as a building in time-series data. Moreover, the algorithm accounted for misclassified buildings due to water-tree interactions in radar signatures and due to topography effects in complex mountainous landforms. The methods were tested in five cities (Bạc Liêu, Cà Mau, Sóc Trăng, Tân An, and Phan Thiết) in Vietnam located in different socio-environmental regions with a range of urban configurations. Using in-situ data and field observations, the methods were validated, and the results were found to be accurate, with an average false negative rate of 10.9% and average false positive rate of 6.4% for building detection. The new approach was developed to be robust against variations in SAR incidence and azimuth angles. The results demonstrated the potential use of satellite dual-polarization SAR to identify persistent building structures annually across rural–urban landscapes and on sea surfaces with different environmental conditions.
The final part of the thesis developed a novel method to map persistent surface water and seasonal inundated land cover and land use. The super-intensive shrimp culture in the Mekong Delta region brings substantial profits to the local economy but it poses major challenges to soil and surface water in wetland areas. The use of geospatial data in monitoring the aquaculture areas is necessary but it has been inadequate in aquaculture areas in the Mekong Delta. In this study, a new algorithm was developed to address the problem of detecting LCLU that contains water such as persistent surface water (permanent lake, permanent rivers, persistently denuded unproductive land) and seasonal inundated land cover (rice paddy and aquaculture) in different environmental conditions. The three-dimensional (3-D) space of VV-VH polarization of the SAR data and Season space was introduced. This study found that the use of the three-dimensional polarization of the SAR and season space is successfully in detecting rice paddy, aquaculture, and persistent surface water. Therefore, the novel method can be utilized to monitor aquaculture in other wetland regions.
In conclusion, this thesis demonstrated the potential use of Sentinel-1 C-band SAR data to map LCLU across the urban suburban to rural-natural landscape on level terrains. The proposed methods can be used for urbanization monitoring, aquaculture development monitoring, and illegal land use change.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
