Skip to content

Latest commit

 

History

History
52 lines (36 loc) · 1.56 KB

File metadata and controls

52 lines (36 loc) · 1.56 KB

System Design

System Goal

Provide operator-facing long-range movement awareness on top of existing surveillance hardware without requiring replacement of the installed camera fleet.

Core Design Principles

  • Work with existing RTSP / ONVIF / recorder infrastructure
  • Focus on movement credibility before semantic certainty
  • Keep tuning practical for real deployment conditions
  • Support both live surveillance and offline validation
  • Make performance measurable

High-Level Components

Source Manager

Handles the active source inventory and coordinates camera, recorder, and offline inputs.

Source Worker

Runs per-source capture and timing logic.

Motion Engine

Runs the detection pipeline:

  1. grayscale preprocessing
  2. optional contrast enhancement / smoothing
  3. motion detection
  4. cleanup and filtering
  5. blob extraction
  6. tracking
  7. confirmation

UI Layer

Provides multi-source operator monitoring, visual overlays, source controls, and alert visibility.

Alert and Event Layer

Routes beeps / voice alerts, snapshots, and motion event history.

Performance Logging

Captures per-source timing and runtime metrics to support deployment analysis.

Why This Architecture Works

This design separates concerns cleanly:

  • source handling is independent from detection logic
  • the detection engine can be tuned without changing the rest of the app
  • operator UX remains responsive
  • performance can be analyzed per stream

That is important for real-world multi-camera deployments where CPU, stream choice, and scene conditions vary from site to site.