Zosi Person Detection Not Working Correctly? Let's Fix It
The person detection feature on Zosi cameras is a powerful tool designed to provide you with the most important security alerts, but it can be frustrating when it doesn't work as intended. Whether you're receiving alerts for moving cars and swaying trees, or worse, the camera is failing to detect an actual person, this guide will help. We provide professional, easy-to-follow solutions to fine-tune your camera's settings and dramatically improve its detection accuracy.
1. Update Firmware and App Settings
Before adjusting anything else, it's crucial to ensure your system's software is up to date. AI algorithms are constantly being improved, and a simple update can often solve the problem.
- Update Camera Firmware: Open the Zosi Smart App, select your camera, and navigate to its settings. Look for an option like "Device Information" or "Firmware Update." If an update is available, proceed with the installation.
- Enable the Correct Detection Mode: In the camera's alarm or detection settings, make sure you have specifically selected "Human" or "Person" Detection. If you only have general "Motion Detection" enabled, the camera will alert you to any movement it sees.
2. Calibrate Detection Sensitivity and Zones
This is the most effective step for reducing false alarms. It involves telling your camera where to look and how sensitive it should be to movement.
- Adjust Sensitivity: A sensitivity level that is too high is the most common cause of false alerts. In the detection settings, you'll find a sensitivity slider or level selector (e.g., Low, Medium, High). If you're getting too many alerts, lower the sensitivity by one level. Test it, and adjust again if needed.
- Set Up Activity Zones: This is a vital feature. An activity zone lets you draw a specific area in the camera's field of view to monitor. By doing this, you can instruct the camera to ignore irrelevant areas. For example, draw a zone around your doorway and path, but exclude the public pavement or a busy road. This will stop cars or pedestrians from triggering alerts.
3. Optimise Camera Placement and Environment
The physical installation of the camera plays a huge role in the AI's ability to work correctly. The algorithm needs a clear view to accurately identify a human shape.
- Camera Angle: Avoid positioning the camera too high with a steep downward angle. This "top-down" view can make it difficult for the AI to recognise a person's form. A height of 7-10 feet (2-3 metres), angled slightly down, is often ideal.
- Lighting Conditions: Ensure the area you want to monitor is well-lit, especially at night. Strong backlighting (e.g., pointing directly at the sun) or deep shadows can confuse the camera's sensor and lead to missed detections or false alarms.
- Remove Obstructions: Make sure there are no obstructions, like tree branches or flags, moving constantly in the foreground of your detection zone. These can trigger repeated, unnecessary alerts.
4. Reboot the Camera
Finally, a simple reboot can often resolve temporary software glitches. Power the camera off, wait for 30 seconds, and then power it back on. This can refresh its internal processes and may improve detection performance.
By systematically working through these steps, you can significantly reduce incorrect alerts and trust that your Zosi camera will notify you of what truly matters.