How to Troubleshoot Inaccurate Unifi Person Detection
One of the most valuable features of the Unifi Protect system is its ability to distinguish between general motion and the specific presence of a person. However, when this feature doesn't work as expected, it can lead to a flood of false notifications or, even worse, missed events.
If you're finding that your camera is incorrectly identifying people—or failing to identify them at all—this guide will provide you with the steps to diagnose and correct the problem.
1. Optimise Camera Placement and View
The performance of the detection AI is highly dependent on the quality of the video it has to analyse. The placement of your camera is the most critical factor for success.
- Angle and Height: Install your camera between 2.5 and 3 metres (around 8 to 10 feet) off the ground, with a clear downward angle. A camera looking straight at the horizon will struggle more than one with a slightly top-down view of the person's profile.
- Avoid Backlighting: Try not to point the camera directly at a strong light source, like the rising or setting sun. A heavily backlit silhouette is very difficult for the AI to identify as a person.
- Clean the Lens: A smudged or dirty lens can soften the image, making it harder for the algorithm to distinguish shapes. Regularly wipe the lens with a soft microfibre cloth.
2. Meticulously Configure Smart Detection Zones
This is the most important setting you can adjust. A Smart Detection Zone tells the camera where to look for people. Movement outside the zone is ignored.
- Be Specific: Go into your camera's settings and carefully draw the detection zone to cover only the areas you care about. For example, trace your walkway and porch, but exclude the public pavement, the street, and any swaying trees or bushes.
- Refine the Zone: If you get a false alert, check the recording to see what caused it. Was it a shadow from a tree at the edge of your zone? Adjust the zone to be slightly smaller to exclude that area. This is an iterative process.
3. Check for Environmental False Positives
Sometimes, the AI can be tricked by movement or objects that mimic the shape or motion of a person.
- Moving Foliage: Trees, large bushes, and even tall grass swaying in the wind are common culprits for false person detections. Ensure these are excluded from your detection zone.
- Shadows and Light Changes: Fast-moving shadows from clouds on a sunny day, or car headlights sweeping across a wall at night, can sometimes be misinterpreted. While harder to control, adjusting camera angle can help.
- Reflections: Be mindful of reflections in windows or puddles of water, which can distort shapes and trigger false alerts.
4. Keep Your Unifi Protect System Updated
Ubiquiti continuously works to improve its detection algorithms. These improvements are delivered through firmware updates.
- Update Everything: Make it a habit to check for and apply updates for your Unifi OS Console (Cloud Key, NVR, etc.) and for your cameras. An update can often lead to a noticeable improvement in detection accuracy overnight.
5. Adjust Sensitivity and Reset
If the above steps don't resolve the issue, there are a couple more things to try.
- Adjust Motion Sensitivity: While the AI is smart, it often relies on an initial motion trigger. If this sensitivity is too high, it may force the AI to analyse non-events. Try lowering it slightly.
- Reboot the System: A simple reboot of both the camera and the main Unifi Protect console can clear temporary software bugs that might be impacting performance.