harshaaa69/fire-detection-yolo

How harshaaa69/fire-detection-yolo works

This is a webcam/video fire-detection application with a separate model-training script. Its main runtime flow captures frames, runs YOLO inference, confirms fire across frames with a cooldown, then raises alerts and displays annotated video. The README also documents training, but the training script references a weight filename that differs from the one in the repository; that relationship is omitted.

Main entry points: detect.py Detection Loop · train.py Training Script · best.pt Trained Weights · Alarm.wav Alarm Audio

Components 16

Detection Runtime

Detection Loop detect.py
Frame Capture detect.py
YOLO Inference detect.py
Trained Weights model/best.pt
Fire Confirmation detect.py
Annotated Display detect.py

Alerts and Records

Model Training

Training Script train.py
Base Weights yolo26s.pt

Other components

Operator
Camera or Video
SMTP Service

Connections 15

Operator → Detection Loop: starts README.md
Camera or Video → Frame Capture: provides frames detect.py
Detection Loop → Frame Capture: reads frames detect.py
Detection Loop → YOLO Inference: runs inference detect.py
Detection Loop → Trained Weights: loads detect.py
Detection Loop → Fire Confirmation: confirms fire detect.py
Detection Loop → Annotated Display: renders results detect.py
Detection Loop → Alarm Playback: triggers detect.py
Alarm Playback → Alarm Audio: plays detect.py
Detection Loop → Email Alert: triggers detect.py
Email Alert → SMTP Service: sends via detect.py
Detection Loop → Event Logging: records event detect.py
Event Logging → Event Log: appends detect.py
Operator → Training Script: runs training README.md
Training Script → Trained Weights: produces README.md

More Python diagrams: shivamim/finance_agenticai · swittouck/scarap · techloset/agentic-ai · ttlns/undetected-playwright-python · uzh-rpg/uzh_fpv_open

Drawn by GitDiagram on September 30, 2026 from the repository's file tree, README and source files. Read as Markdown ·