How keshav-077/ai-driven-border-security-and-strategic-management-system works
This repository documents a border-surveillance prototype: camera feeds are analyzed for weapons and people, with separate face identification and cross-camera re-identification capabilities. Training notebooks and dataset configuration files support the detection models; result artifacts record evaluation. The README describes the integrations, but no source excerpts were available to verify runtime call ownership or detailed wiring.
Main entry points: trainingpersons.ipynb Person model training · trainingweapons&Mini Project.ipynb Weapon model training · data.yaml Dataset config · Faces.ipynb Face detection
Components 15
Camera Inputs
- IP/Webcam feeds
Video Analysis
- Weapon detection
trainingweapons&Mini Project.ipynb - Person detection
trainingpersons.ipynb - Face detection
Faces.ipynb - Face classification
Faces.ipynb - Person re-identification
trainingpersons.ipynb
Model Training
- Person dataset config
dataperson.yaml - Weapon dataset config
dataweapon.yaml - Dataset config
data.yaml - Person model training
trainingpersons.ipynb - Weapon model training
trainingweapons&Mini Project.ipynb - YOLO weights
yolov8n.pt
Evaluation
- Person results
Resultspersontraining/results.csv - Weapon results
Resultsweapon/results.csv - Final results
finalresults/results
Connections 16
- IP/Webcam feeds → Weapon detection: video input
- IP/Webcam feeds → Person detection: video input
- IP/Webcam feeds → Face detection: video input
- IP/Webcam feeds → Person re-identification: camera streams
- Face detection → Face classification: face regions
- Person dataset config → Person model training: configures data
- Dataset config → Person model training: configures data
- Weapon dataset config → Weapon model training: configures data
- Dataset config → Weapon model training: configures data
- Person model training → Person detection: trained model
- Weapon model training → Weapon detection: trained model
- YOLO weights → Person detection: model weights
- Person model training → Person results: records metrics
- Weapon model training → Weapon results: records metrics
- Person results → Final results: evaluation artifacts
- Weapon results → Final results: evaluation artifacts