Ongoing Projects


Inter- and intra-annual dynamics of urban impervious surfaces in the Pearl River Delta using deep learning networks from synergized optical and SAR data

Principal Investigator: Prof. Zhang Hongsheng

Start Date: 2018
End Date: 2019
Type of Grant: GRF
Project details

Although urban areas occupy less than 1% of the world’s land surface, they contribute over 90% of the global economy. By 2014, already 50% of the world’s population was living in urban areas, and this figure is increasing (United Nations, 2015). Urban impervious surfaces, such as buildings, parking lots and transportation networks, have been widely recognized as an important indicator of urban expansion and its related environmental issues, such as public health, urban heat islands and various environmental pollutants. In the long run, and especially for areas that are subject to a very strong urbanization process, such as the Pearl River Delta (PRD), a comprehensive assessment of impervious surfaces dynamics will be vital and urgent in the assessment and management of these effects. Unfortunately, it has been increasingly difficult to accurately assess the urban impervious surfaces in this area because of its wide geographical extent. PRD has witnessed the world’s most dramatic urbanization process since 1980s, thanks to the implementation of the reform and opening up policies of the Chinese Government. More specifically, PRD is a special metropolitan area and includes three different urban planning and development policy bodies, Mainland China, Hong Kong and Macau. The increase of urban impervious surfaces in PRD has led to a wide range of environmental issues such as urban flooding and air and water pollution. These effects have been threatening the environmental equilibrium and human health in the whole PRD region. Therefore, a continuous and timely monitoring of the dynamics of impervious surfaces in PRD using remote sensing has been identified by both local and central governments. However, PRD is located in a subtropical climate zone characterized by rainy and cloudy weather throughout the year, and thus requires the most advanced remote sensing technology to overcome weather-related difficulties. The ultimate objectives of this research are to: 1) estimate and validate the urban impervious surfaces extents by manifold learning and deep learning jointly considering optical and synthetic aperture radar (SAR) satellite data; and 2) investigate and model the inter- and intra-annual dynamics of the urban impervious surfaces in PRD exploiting available time-series data analysis techniques. The outcomes of the research will: 1) include a novel methodological framework for the use of manifold learning and deep learning for accurate impervious surface estimation from optical and SAR data; and 2) provide a comprehensive understanding of the inter- and intra-annual dynamics of the urbanization in the PRD region.


Development of a Novel, Fine Prediction System of Storm Surge Inundation for Hong Kong Coastal Area

Principal Investigator: Prof. Pan Jiayi

Start Date: 2017
End Date: 2018
Type of Grant: ITF
Project details

With so many typhoon passages each year, the Hong Kong coastal area is strongly influenced by extremely strong wind events during the typhoon season. The typhoon related storm surges can induce disastrous inundations in the Hong Kong coastal area, resulting in serious casualty and property loss. Thus, rapid and precise prediction of the storm surge-induced inundations is of great importance to mitigation of inundation disasters in the coastal area of Hong Kong. Because of the complexity in Hong Kong coastal topography and elevation and the complicated interaction between storm-generated strong ocean currents and the coastal elevation, however, there has been no effective inundation prediction system that could be applied to the Hong Kong coastal area.

This project aims at developing a precise and fine prediction system for storm surge inundation in the Hong Kong coastal area with the satellite remote sensing technology on the platform of open source geographic information system (GIS), incorporating the enhanced Hong Kong coastal digital elevation model (DEM) database to provide an operational inundation forecast system for relevant users. The system will play an importance role in Hong Kong oceanic disaster mitigations.


Mangrove Species Discrimination in Hong Kong with Synergistic Use of High Resolution Optical and SAR Satellite Data

Principal Investigator: Prof. Zhang Hongsheng

Start Date: 2017
End Date: 2018
Type of Grant: GRF
Project details

Mangrove forests are ecosystems of considerable ecological, biological and socioeconomic significance as: 1) vital habitats for a wide variety of animal and plants species, 2) important sources of carbon for detritus-based food webs in adjacent coastal waters, 3) a means of reducing the erosion of shorelines, and 4) buffers against the impact of storm waves and floods. However, mangrove forests have been significantly eroded over the past century due to various human activities such as agriculture conversion, urbanization and tourism. Hong Kong benefits from a number of mangrove forests which continue to play important ecological and socioeconomic roles to local communities. Unfortunately, mangrove forests cover has rapidly decreased in the past few decades due to increasing agricultural exploitation and infrastructural development such as new highways and airport as well as associated construction. Given the importance of mangrove forests, timely and accurately monitoring at large scale is imperative for their conservation and restoration. Mangrove forests are usually located in inaccessible regions and often temporarily inundated, which makes it difficult to carry out large-scale and long-term monitoring with traditional field survey approaches. Therefore, satellite techniques have been widely employed for large scale monitoring. However, the coarse spatial and spectral resolution of early satellite imaging restricted the detail and accuracy of the results. With the latest high resolution satellite sensors, synthetic aperture radar (SAR) data has been increasingly reported to be sensitive to biomass and structural attributes of vegetation and thus holds great potentials for mangroves monitoring. Nevertheless, the potentials of SAR for discriminating mangrove species are still underexplored, especially with fully polarimetric SAR data. The general aims of this research are to: 1) investigate the comprehensive potentials of multi-frequency polarimetric SAR data with single, dual, compact and full polarizations for improving the discrimination of mangrove species in Hong Kong, and 2) develop effective and efficient methods for synergizing polarimetric SAR data and high resolution optical data at both feature and decision levels. The outcomes of this proposed research will not only include a comprehensive method for investigating a diversity of polarimetric features by evaluating their effectiveness and responses to various mangrove species, but also a general framework for incorporating high resolution optical and multi-frequency polarimetric SAR features for mangrove species discrimination. The implementation of this project can serve as a general model to promote polarimetric SAR techniques for mangrove forests monitoring, and finally, to support the conservation and restoration of mangrove forests.


Improving the estimation of impervious surfaces using optical and polarimetric SAR data in humid subtropical urban areas

Principal Investigator: Prof. Zhang Hongsheng

Start Date: 2016
End Date: 2018
Type of Grant: GRF
Project details

Impervious surfaces (IS), as the major land surface modification of rapid urbanization in many metropolitans in the past decades, have been widely recognized as imperative indicator for not only urban environmental issues (e.g. water quality, aerosols, urban heat island and climate change) but also socio-economic issues (e.g. population distribution and urban planning) from local, regional to global scales. Pearl River Delta (PRD) is such a typical metropolitan with large area of IS, which have been leading to various environmental and socio-economic problems, and thus urgently require accurate estimation of IS for timely monitoring and understanding its urbanization processes. Given its importance, numerous methods have been developed for IS estimation (ISE) from satellite data. However, most of them were tested in temperate continental areas and with only optical satellite data. Only a few of them considered the case in subtropical humid areas like PRD, where climatology (e.g. rainy and cloudy) and phenology are special and different from temperate regions and thus cause great difficulties for remote sensing studies. To overcome the difficulties, Synthetic Aperture Radar (SAR) data has been synergized with optical data in subtropical regions in the past several years, but only single polarization SAR data were employed and the accuracy is far from satisfactory. Multi-polarimetric SAR data have been frequently reported to be able to provide much richer information in urban areas, while its potential application in ISE is still underexplored and many challenging problems remain unaddressed. The general aims of this research are to: 1) Investigate the full potential of multi-polarimetric SAR data, including dual, compact and fully polarimetric SAR data, for improving ISE in subtropical urban areas; and 2) Develop effective fusion methodology for synergizing optical and polarimetric SAR data at both feature and decision levels from city to metropolitan scales. The outcomes of this proposed research will not only provide a comprehensive feature extraction framework for polarimetric SAR data, but also evaluate the effectiveness of these polarimetric SAR features by their contributions to ISE at both feature and decision levels. Additionally, as a validation of the proposed methodology, PRD will be employed for case studies from city scale and metropolitan scale. The implementation of this project can serve as a general model to promote the applications of optical and polarimetric SAR data in accurate ISE of other humid subtropical urban areas in the world.



Copyright © 2016 All Rights Reserved. Institute of Space and Earth Information Science, The Chinese University of Hong Kong
About Us | Contact Us | Disclaimer | Privacy Policy | Sitemap

Top