the person detection on my outdoor cam is awful, it alerts me for every car and shadow

Diagnosis

The Google Nest Cam Outdoor Wired utilizes a combination of pixel-change detection and an onboard AI model to differentiate between humans and other moving objects. When you receive constant alerts for cars and shadows, it is usually because the camera's "Field of View" (FOV) includes high-contrast areas or high-traffic zones. Shadows from swaying trees or the reflection of headlights on a wet driveway create rapid changes in pixel luminosity that the AI often misinterprets as the movement of a human figure, especially if those shadows fall across a vertical surface like a wall. Furthermore, the camera's default motion sensitivity is often tuned too high for environments with street traffic. If the camera is angled such that the street occupies more than 20% of the frame, the AI will frequently trigger "Person" alerts because it detects a vertical silhouette (like a car pillar or a pedestrian on a distant sidewalk) within the motion-heavy area of the street. To confirm this, review your Event History; if the "Person" alert triggers exactly when a car passes a specific shadow-casting tree or lamp post, you are dealing with a lighting-trigger issue rather than a software glitch.

How to Fix It

To stop the false positives, you must isolate the AI's focus using a combination of Activity Zones and specific event tuning. Follow these precise steps:

1. Open the Google Home app and select your Nest Cam Outdoor Wired from the device list.
2. Tap the Settings (gear icon) in the top right corner and navigate to Events.
3. Select Seen Events and ensure that "People" is toggled ON, but check if "Animals" or "Vehicles" are also on; if you only want people, disable everything else to reduce noise.
4. Go back to the main camera settings and select Activity Zones.
5. Tap Add a Zone. Instead of one large box, draw a tight, precise polygon that covers only your porch, walkway, or driveway. Crucially, ensure the zone boundary does not touch the edge of the street or any areas where tree shadows fall.
6. Save the zone and give it a specific name (e.g., "Front Porch").
7. Navigate to Video settings and adjust the Motion Sensitivity slider. Move it from the default "Medium" to "Low." This forces the AI to require a more distinct, larger human shape before triggering an alert, effectively filtering out small shadows and distant car movements.
8. Restart the camera by unplugging the power cable for 10 seconds to ensure the new zone parameters are fully initialized in the local cache.

If That Didn't Work

• Adjust the Physical Tilt: If the camera is angled too low, it captures too much of the foreground pavement, which is where shadows are most volatile. Tilt the camera up by 5-10 degrees. This shifts the horizon line and reduces the amount of street surface in the frame, limiting the AI's opportunity to misidentify car reflections as people.
• Install a Motion-Shielding Trim: If a specific tree branch or bush is causing the shadow triggers, use a small piece of black electrical tape or a plastic shroud on the very edge of the lens to physically block that specific area of the FOV. This prevents the sensor from ever seeing the problematic movement.
• Check for IR Glare: At night, the infrared (IR) LEDs can reflect off white fences or glossy cars, creating "blooms" that the AI sees as movement. If this is happening, slightly rotate the camera 5 degrees to the left or right to change the angle of reflection.
• Update Firmware via Power Cycle: Sometimes the AI model fails to update. Perform a hard factory reset by holding the reset button on the back of the camera for 10-15 seconds until the status light pulses. Re-add the camera to the Home app to ensure you are on the latest version of the person-detection algorithm.

Heads Up

• Zone Overlap: Never create overlapping Activity Zones; this can cause the AI to double-trigger and increase the frequency of false alerts.
• Lighting Consistency: Avoid placing the camera directly facing a bright street light or the sun. Extreme backlighting creates silhouettes that the AI almost always identifies as "People" regardless of your settings.
• Cleaning the Lens: Dust and spider webs on the lens can reflect IR light at night, creating "ghost" movements that trigger the motion sensor. Clean the lens weekly with a microfiber cloth.
• Network Latency: If your upload speed is below 2Mbps, the AI processing in the cloud may lag, causing alerts to trigger several seconds after a car has already left the frame, making it seem like the camera is "hallucinating" people.