How harisudhan005/mineguard-ai works
MineGuard AI is presented as an open-pit mine safety dashboard combining sensor readings and slope-image crack analysis into a risk assessment, with alerts, weather context, and an AI assistant. The documented workflow is to upload sensor CSV data and an image, view the analysis on the dashboard, and optionally ask the assistant about it. The repository also contains a YOLO-labeled mine-crack dataset and several historical UI/backend snapshots. The main backend excerpt confirms upload analysis, weather lookup, vision inference, and Ollama chat; other UI capabilities are documented but not all implementation details are shown.
Main entry points: app.py Flask analysis API · script.js Dashboard UI · postchatbot/script.js Assistant UI · Mine-crack-datasets-1 Labeled crack dataset
Components 16
Dashboard experience
- Dashboard UI
script.js - Alert management
Snapshot/postAlertUi/script.js - Assistant UI
Snapshot/postchatbot/script.js
Risk analysis
- Flask analysis API
app.py - Sensor risk model
app.py - Crack detection
app.py - Weather context
app.py - Risk response enrichment
app.py
Analysis inputs
- Labeled crack dataset
Mine-crack-datasets-1 - Dataset configuration
Mine-crack-datasets-1/data.yaml - Sensor CSV upload
- Slope image upload
External services
- WeatherAPI
- Ultralytics YOLO
- Ollama Phi-3
Other components
- Mine operator
Connections 19
- Mine operator → Dashboard UI: uses
- Mine operator → Sensor CSV upload: uploads CSV
- Mine operator → Slope image upload: uploads image
- Dashboard UI → Flask analysis API: requests analysis
- Flask analysis API → Sensor CSV upload: reads CSV
app.py - Flask analysis API → Slope image upload: saves image
app.py - Flask analysis API → Sensor risk model: predicts risk
app.py - Flask analysis API → Crack detection: infers cracks
app.py - Crack detection → Ultralytics YOLO: runs inference
app.py - Flask analysis API → Weather context: fetches weather
app.py - Weather context → WeatherAPI: requests conditions
app.py - Flask analysis API → Risk response enrichment: builds response
app.py - Flask analysis API → Dashboard UI: returns risk data
- Dashboard UI → Alert management: updates alerts
- Mine operator → Assistant UI: asks questions
- Assistant UI → Flask analysis API: sends question
- Flask analysis API → Ollama Phi-3: posts chat prompt
app.py - Flask analysis API → Assistant UI: returns reply
- Labeled crack dataset → Dataset configuration: configured by