Thesis Project
Digital Reconstruction of Landscapes in Mangya
This thesis project focuses on the digital reconstruction of three major landscapes in Mangya, exploring how large-scale natural environments can be translated into cohesive, immersive 3D worlds. Using procedural workflows and terrain-generation tools, the project recreates the region’s distinctive geological features and spatial structures while maintaining visual consistency and realism. The work emphasizes landscape formation, material transitions, and atmospheric lighting, demonstrating a scalable pipeline for environmental reconstruction applicable to game cinematics and real-time environments.
Week 1
01/11/2026
This week, I summarized the research I did during the winter break. I focused on an article discussing the historical and geological background of Mangya city and three articles which provide examples for the digital reconstruction. The link below leads to the research document.
https://docs.google.com/document/d/1R-Zx7M8U5rufRAUccS8NjyUsy2WtZ7YMDHHa7-9gh3Q/edit?tab=t.0
Week 2
01/18/2026
This week, I explored the workflow for recreating terrains. I chose to start with real-world elevation data as a foundation and then build additional detail using Houdini’s procedural tools. I tested heightmaps from several different sources and ultimately selected data from USGS Earth Explorer, as it is a reliable and authoritative source for elevation data.
Elevation map and Celestial map from USGS Earth Explorer
Terrain from heightmap in Houdini
Week 3
01/25/2026
This week, I further developed the terrain details of the solitary mountain and began testing the material workflow. I generated a flow mask and used it as an input in the landscape material to blend different surface types. Areas outside the mask are covered with rock, while masked regions are blended with sand. This approach can be extended using additional masks, such as sediment, to introduce greater material variation and detail.
Detailed terrain in Houdini
Reference
Flow masks
