anomaly-detection-vision

What problem does it solve?

Visual anomaly detection enables autonomous identification of environmental changes in camera scenes, such as moved objects, opened doors, new items, missing objects, or potential intruders during patrols.

Core Features & Use Cases

• Implement pixel-level difference, SSIM, and feature-based comparison to flag scene changes.
• Manage reference images to establish baselines and support periodic updates.
• Temporal filtering and Behavior Tree integration to provide robust, automated surveillance.
• Use Case: Patrol robots, surveillance systems, or warehouse security requiring reliable detection of unusual activities.

Quick Start

Run the anomaly detector by loading a reference image for a waypoint and starting the vision-based anomaly monitoring.