Projects

Using Sar Data To Detect Windthrows In The Amazon Rainforest

MSc Dissertation, The University of Manchester, 2024

In my master’s dissertation, I investigated the detection of windthrow events in the Amazon rainforest using Sentinel-1 SAR data, comparing its effectiveness to Sentinel-2 optical data. I processed and analysed radar backscatter and multispectral images from the Google Earth Engine platform, applying time-series analysis to identify changes before and after windthrow events. My research explored the strengths and limitations of SAR data in monitoring forest damage, particularly in complex terrains with high vegetation density, and highlighted the potential for integrating multi-source satellite data to enhance detection accuracy.

Statistical Analysis of Urban Road Traffic Using Time-Series Data from High-Resolution SAR Satellites

BSc Dissertation, Wuhan University, GNSS Research Center, 2023

In my undergraduate thesis, I utilised high-resolution Synthetic Aperture Radar (SAR) satellite data to conduct a six-year time-series analysis of traffic volume and vehicle speed on the Third Ring Road near Yezhihu, Wuhan, China. By applying the Doppler effect principle, I derived a quantitative relationship between azimuth displacement in SAR imagery and actual vehicle speed. Algorithms such as filtering and down-sampling were employed to enhance vehicle detection accuracy. Using TerraSAR-X satellite data, I automated the extraction of road vehicle targets and completed time-series statistics on vehicle counts and speeds, revealing traffic trends for the area. I also analysed the impact of noise in high-resolution SAR imagery and applied a median filtering method for denoising in time-series data, providing a feasibility study for large-scale traffic monitoring in the future.

Remote Sensing Analysis of Reservoirs and Lakes in Hubei Province

Team Leader and Organiser, Wuhan University, University Student Scientific Innovation Competition Project, 2023

In this project, I utilised satellite imagery from the European Space Agency’s Sentinel series and Landsat 8 to extract and analyse water area changes over time. Using platforms such as Google Earth Engine, ENVI, and ArcGIS, I processed multi-temporal optical and SAR data to evaluate and optimise water body extraction algorithms, improving both accuracy and computational efficiency. I built a time-series database documenting years of water area changes in Hubei Province and analysed their correlation with regional climatic factors, such as precipitation and temperature, using multivariate regression analysis. Additionally, I modelled flood inundation areas to assess impacts during flood events, supporting precision relief efforts and agricultural evaluations. This project provided scientific insights into the quantitative relationship between climate fluctuations and changes in lakes and reservoirs, offering valuable references for regional economic development and climate adaptation strategies.

Resort Geospatial Database Development and UAV-based Mapping

Summer GIS Analyst Internship, Tianzi Lake Resort, China, 2022

During my internship as a GIS Analyst for a resort project, I led the design, processing, and application analysis of a 2D and 3D foundational spatial database using ArcGIS. At the project’s outset, I developed a geographic entity feature database encompassing roads, buildings, ponds, vegetation, and car parks, meticulously defining attribute information for each layer. Building on the database, I analysed road network connectivity within the resort and identified optimal routes. For surveillance camera placement, I evaluated visibility zones based on terrain and building distribution, optimising coverage efficiency. Finally, I utilised 3D visualisation techniques to create an interactive 3D navigation map of the area, enhancing the visitor experience and improving customer satisfaction.

3D Laser Scanning and Tunnel Cross-Section Modelling

Summer on-campus internship, Wuhan University, 2022

During my campus-based internship, I led a team of eight in 3D laser scanning using a Leica P50 LiDAR scanner to capture detailed point cloud data of building structures and tunnel cross-sections. Faced with a tight schedule, I independently learnt the equipment in three days, trained my teammates, and ensured a smooth operation. We scanned the tunnel structure, processed the data in Cyclone through registration, merging, and modelling, and successfully generated a 3D tunnel model. I refined cross-section images in AutoCAD and performed multi-station LiDAR data stitching and optimisation to enhance spatial accuracy and visualisation quality, delivering high-quality results under pressure.

Mohan Sheng