The short answer
Draw tight activity zones that exclude the road, the pavement and anything that moves in the wind, then enable person detection and turn off general motion alerts. Sensitivity should be high within a small zone rather than low across a wide one, because that catches real events while ignoring the rest.
The failure mode of a security camera is not missing an event, it is sending so many alerts that you mute them and then miss the one that mattered. Tuning is therefore not optional, and it takes about twenty minutes done properly.
Step 1: activity zones
Almost every camera lets you mask parts of the frame. This is the most effective single tool available.
Exclude the road and pavement entirely
Passing cars and pedestrians are the largest source of false alerts by a wide margin.
Exclude trees, bushes and washing lines
Anything that moves in wind. A branch at the edge of frame will trigger every windy night.
Exclude reflective surfaces
Windows, parked car bonnets and puddles, all of which produce moving light that reads as motion.
Keep the zone to your approach
The path, the drive, the doorstep. Nothing beyond your boundary.
Draw it tightly rather than generously
A narrow zone with high sensitivity is far better than a wide zone with low sensitivity.
Step 2: use detection types rather than motion
Modern cameras distinguish people, vehicles, animals and packages. Turning off general motion alerts and enabling person detection alone transforms the alert volume.
Where detection runs on the camera rather than in the cloud, it is faster and works during an internet outage. Where it is a subscription feature, this is usually the part of the subscription that is genuinely worth paying for.
Be aware that person detection is not perfect: it misses people occasionally and identifies a wheelie bin as a person occasionally. Tightening zones reduces both.
Configure alerts and recording separately where the camera allows it. Record everything that moves, so the footage exists if you need it, and alert only on people in the zone you care about. That combination gives you evidence without notification fatigue.
Step 3: the physical causes of false alerts
- Infrared reflecting off a nearby wall or surface at night, which produces a bright area that reads as motion. Angling the camera slightly usually fixes it.
- Insects attracted to the infrared illuminator, a very common cause of nighttime alerts. A spray of insect repellent on the mounting bracket, not the lens, helps. So does mounting away from a light.
- Spiders building webs across the lens, which produce large moving shapes. Regular cleaning and repellent on the bracket are the answer.
- Rain and snow lit by the infrared, which appears as a storm of moving particles. Some cameras have a setting for this.
- Direct sun moving across the frame, producing shadows that trigger motion. Check the view at different times of day.
- Heat haze from a vent or a flue, which is invisible to you and not to the camera.
Step 4: scheduling and modes
Most systems support different behaviour at different times or based on whether anyone is home:
- Alerts only when the house is empty, using geofencing.
- Quieter alerting during the day when comings and goings are expected.
- A cooldown period between alerts, which prevents a single event generating twenty notifications.
- Snooze for a period when you are working in the garden and know you will trigger it repeatedly.
Using these makes alerts meaningful again, because an alert at 3am means something different from one at 3pm.
Step 5: notification settings on the phone
Separate from the camera, control how alerts reach you:
- Different sounds for person detection and general motion.
- Critical or time-sensitive notifications for the cameras that matter, so they bypass Do Not Disturb.
- Notification summaries turned off for the camera app, since batching security alerts defeats the point.
- Unrestricted battery use for the app, or alerts arrive late.
Be careful about tuning so aggressively that you miss what you installed the camera for. The goal is fewer, better alerts rather than silence. Test after tuning by walking the approach yourself at different times of day and confirming you get the alert.
Step 6: check what the alerts cost
On battery cameras, every alert consumes battery, and a camera generating fifty alerts a day will need charging weekly rather than monthly. Tightening zones is therefore also the main way to extend battery life, which is a more compelling argument for many people than notification fatigue.
How do I stop alerts for passing cars?
Draw an activity zone that excludes the road and pavement entirely, and enable person detection rather than general motion. Together those remove the large majority of false alerts.
Why do I get alerts at night with nothing visible?
Usually insects attracted to the infrared illuminator, spider webs across the lens, or rain lit by the infrared. Repellent on the bracket and a slight angle change fix most of them.
Should sensitivity be high or low?
High, within a tightly drawn zone. Low sensitivity across a wide zone misses real events while still catching some false ones, which is the worst of both.
Is person detection reliable?
Good but not perfect. It misses occasionally and misidentifies occasionally, and tight activity zones improve both. On-device detection is faster and works during an internet outage.
Do false alerts drain the battery?
Substantially on battery cameras, since each one wakes the camera to record and upload. Tightening zones is the most effective way to extend battery life as well as reduce notifications.
The short version
Draw tight zones that exclude the road, trees and reflective surfaces. Turn off general motion alerts and use person detection. Record everything but alert on little. Deal with insects and spider webs, which cause most nighttime false alerts. Then walk your own approach to confirm you still get the alerts that matter.
