---
license: agpl-3.0
pipeline_tag: depth-estimation
tags:
- ONNX
- Indoor
- Small
- Large
- Base
- DepthAnything
- Depth
- UniDepth
- YOLO
base_model:
- Ultralytics/YOLO11
- lpiccinelli/unidepth-v2-vits14
- lpiccinelli/unidepth-v2-vitb14
- lpiccinelli/unidepth-v2-vitl14
- depth-anything/Depth-Anything-V2-Metric-Indoor-Small-hf
- depth-anything/Depth-Anything-V2-Metric-Indoor-Base-hf
- depth-anything/Depth-Anything-V2-Metric-Indoor-Large-hf
---

<p align="center">
  <img src="assets/EdgeCV4Safety_logo.png" alt="EdgeCV4Safety Logo" width="18%"/>
</p>

# EdgeCV4Safety: AI-Driven Contextual Safety System for Industry 4.0/5.0

**EdgeCV4Safety-Models** contains **ONNX** **_depth estimation_** and **_object detection_** models. Specifically **UniDepth v2** and **Depth Anything v2** for depth estimation and **YOLO11** for object detecion.

This repo was created to support the main GH repository [**Edge4CVSafety**](https://github.com/EdgeCV4Safety/EdgeCV4Safety.git). This project implements a **modular and scalable Computer Vision (CV) system** designed to replace traditional physical barriers, enhancing **worker safety** in industrial settings (Industry 4.0/5.0). The core objective is to achieve **contextual control** of machinery based on the dynamic state of the surrounding work environment.

The system continuously monitors a defined workspace. Upon the detection of personnel entering this area, appropriate countermeasures are instantly triggered, influencing machinery behavior to prevent hazardous situations. This compartmentalized architecture promotes high **modularity and scalability**.