You have a wide-angle camera watching your front yard. A person walks up to your door — you want an alert. A squirrel runs across the lawn — you don't. A car drives by on the street — definitely not.
Even with AI-powered monitoring, analyzing the entire frame can produce unnecessary alerts. The solution? Monitoring zones — tell the AI exactly which areas of the image to focus on.
What Are Monitoring Zones?
Monitoring zones are rectangular regions you draw on your camera's snapshot. When you assign an alert rule to a zone, the AI only analyzes that specific area of the image and ignores everything else.
Think of it like telling a security guard: "Watch the front door. Don't worry about what's happening on the sidewalk."
How to Set Up Zones in SmartCam
- Go to your Dashboard and click Manage on a camera
- In the Monitoring Zones section, click Draw Zone
- Click and drag a rectangle on the camera snapshot
- Name your zone (e.g., "Front Door", "Driveway", "Porch")
- Pick a color to keep zones visually distinct
- When creating an alert rule, select the zone from the dropdown
That's it. The AI now receives instructions like: "Focus ONLY on the region 'Front Door' which occupies 15% to 45% of the image. Is there a person there?"
Real-World Zone Examples
Front Yard Camera
- Zone: "Front Door" (small area around entrance) → Rule: "Is there a person at the door?"
- Zone: "Driveway" (paved area) → Rule: "Is there an unfamiliar vehicle?"
- Zone: "Porch" (covered area) → Rule: "Is there a package or delivery?"
The sidewalk, street, and neighbor's yard are excluded entirely.
Backyard Camera
- Zone: "Gate" → Rule: "Is the gate open?"
- Zone: "Pool Area" → Rule: "Is there a child near the pool?"
- Zone: "Garden" → Rule: "Is there an animal in the garden?"
Indoor Camera (Office/Warehouse)
- Zone: "Server Rack" → Rule: "Is someone touching the servers?"
- Zone: "Emergency Exit" → Rule: "Is the exit door open?"
- Zone: "Reception Desk" → Rule: "Is there someone waiting at reception?"
Zone Best Practices
- Start with 2-3 zones. Don't over-zone. Each zone adds an AI analysis call.
- Make zones slightly larger than the area of interest. Give a little buffer so the AI has context.
- Use descriptive names. "Front Door" is better than "Zone 1" — the AI uses the name in its analysis prompt.
- Different colors help. When you have 3+ zones, distinct colors make the overlay much easier to read.
- Avoid overlapping zones. If two zones overlap, the same event might trigger duplicate alerts.
How It Works Under the Hood
When you assign a zone to an alert rule, SmartCam sends the zone's coordinates to the AI along with the snapshot. The AI receives instructions like:
Focus ONLY on the region "Front Door" which occupies the area from (25%, 30%) to (55%, 85%) of the image. Ignore everything outside this region. Question: Is there a person at the door?
Vision AI models like Google Gemini understand spatial coordinates and can accurately focus on specific regions of an image. This is the same technology used in visual grounding research, adapted for practical security monitoring.
The Impact
Users who set up monitoring zones typically see a 90%+ reduction in false alerts compared to full-frame monitoring. A camera watching your entire front yard might generate 10 alerts a day. The same camera with a zone focused on the front door? Maybe 1-2, and they're both real.
More importantly, every alert becomes trustworthy. When your phone buzzes, you know it matters.
Try zone-based monitoring
Draw zones, set rules, and stop getting alerts you don't care about.
Get Started Free