Upload README.md

README.md

@@ -1,12 +1,11 @@
----
-title: Codex Model
-emoji: 📊
-colorFrom: purple
-colorTo: pink
-sdk: gradio
-sdk_version: 5.46.0
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# 3D Model Reconstruction for GIS
+
+This repository documents a complete workflow for turning sparse, non-overlapping photographs into interactive 3D assets that can be hosted on a [Hugging Face Space](https://huggingface.co/spaces). It focuses on pipelines that are appropriate for GIS and urban-planning research, integrating structure-from-motion (SfM), neural rendering (NeRF), and Gaussian Splatting approaches.
+
+## How to use this repository
+
+1. **Study the end-to-end plan** in [`docs/hf_sparse_to_3d_pipeline.md`](docs/hf_sparse_to_3d_pipeline.md). It explains the tooling choices (COLMAP, Nerfstudio, 3D Gaussian Splatting, VGGT depth), provides step-by-step commands, and highlights GIS-specific post-processing.
+2. **Start from the Hugging Face Space skeleton** under [`hf_space/`](hf_space/). The folder contains a production-ready `gradio` application and dependency manifests that you can adapt to your project and deploy directly on Hugging Face without building a Docker image.
+3. **Iterate on the pipeline** by wiring additional reconstruction backends, adding evaluation hooks, or integrating geospatial metadata exports as required by your PhD project.
+
+The repository is intentionally modular so that you can swap in newer research ideas (for example recent CVPR papers) without rewriting the Space front-end.