Urban Mining

What if our cities were understood as resources in their own right, composing a new kind of material repository, a hyper-local material library.

3D Scanning and Target Recognition Detection of Interior Materials

This process combines 3D scanning, photogrammetry, and historical document research to construct virtual architectural models. Machine learning image recognition techniques were then used to identify material types.

House Scanning and Deconstruction Strategies

After collecting architectural data, different demolition methods will be chosen based on the existing material availability and the specific material requirements for the new construction.

Datasets for Target Recognition and Semantic Segmentation

Machine learning is employed for material identification and quantity assessment. Target recognition is used to determine material types, while semantic segmentation is used to estimate the proportion of materials within the building.

The Process of House Dismantling

Based on the pre-inspection before demolition, the materials inside the building are selectively removed and catalogued.

Material Library

The data of the demolished materials is uploaded to the online material library, providing the user with a visualisation of the material information.

Material Recoding Strategies

The building is first treated as a box with different resolutions and then materialised and functionalised. According to the different sizes and functionality of the materials, the building is finally assembled with reversible disassembly design.

Small Material Voxelisation Definitions

This process deconstructs how a closed space can be formed by defining the modules of the different dimensions of the voxel, and how the voxel is connected. The voxel is thus divided into different functions and different dimensions.

Multi-Scale Voxels Iteration

Based on the difference in the integrity of the building materials obtained from the different demolition methods, the building elements are divided into three different sizes to incorporate different sizes of demolished materials.

Methods of Assembling Materials

In the process of assembling reused materials, uniform-sized cut pieces or a variety of different-sized pieces can be used.

Building Recoding Generation with Hashtags

Buildings can be constructed by selecting material modules based on different hashtags selected by users.

Material Autonomy Community Function Definition

Land is divided into a 1m x 1m grid and categorised into different basic functions. These functions are combined to create complete 3m x 3m spaces. The space is then materialised by connecting grid units using a wave function collapse algorithm.

Resident Participatory Preference Scoring Community Layout Generation

The system's adjacency mechanism delineates the specific functions of the community site when residents make design choices.

The Process of Community Optimisation

Layout optimisation of the community is based on factors such as plot ratio, solar radiation, community green space ratio, and community public spaces.

The Materialisation of Community Buildings

Material Package Assembly Tutorial in Platform

The system provides the user with a range of building assembly modules with corresponding assembly guidance, and the embodied carbon of the materials will be used as a reference for the purchaser's carbon credit.

Platform Interactive Construction

Time-Lapse of the Building Construction Process

Material Locale