A security camera that cries wolf is worse than no camera, because it trains you to dismiss it. The usual response is to turn the sensitivity down until the alerts stop, which also turns off the alerts you wanted. The better approach is to understand what is actually triggering the camera, because almost every cause has a specific fix and they are not interchangeable. Here they are in the order worth trying.
Fix one: move the camera before you touch a setting
More false alerts are caused by where a camera points than by how it is configured. Four placements are reliably bad. A camera that can see a public road or sidewalk will alert on every vehicle and pedestrian. A camera aimed at a tree, shrub or hanging basket alerts on every gust of wind. A camera looking at a surface in direct sunlight alerts as cloud shadows cross it, and in the morning and evening it alerts as the sun moves. And a camera near a heating vent, extractor, wood stove or clothes-dryer vent will trigger an infrared motion sensor with warm air alone — nothing visible moves at all.
Re-aiming is free and usually decisive. Point cameras across an approach rather than down it, keep the horizon and the sky out of frame where you can, and get vegetation out of the lower third of the picture. Our camera placement guide covers the angles worth covering; this is the same advice read backwards, from the false-alert end.
Fix two: draw activity zones properly
Nearly every camera app lets you mask part of the frame so motion outside it is ignored. This is the single highest-value setting on the page and most people either skip it or draw it too generously. Three rules make it work. Exclude everything you do not own— the road, the sidewalk, the neighbor's drive, the bus stop across the street. Exclude the top third of the frame in most outdoor scenes, because sky, trees and rooflines produce motion that is never a security event. And draw the zone around the approach rather than the destination: the strip of path somebody has to cross to reach your door, not the door itself, because by the time they are at the door the useful warning has passed.
Note the thing that catches people out: on many cameras the activity zone filters notifications, not recording. The camera may still be recording continuously, which is what you want — you are suppressing the alerts, not the footage.
Fix three: understand what kind of detection you actually have
This is where most of the remaining noise lives, and it is a hardware question rather than a settings one. There are three mechanisms in play and they fail differently.
Pixel-change motion detection compares frames and flags differences. It is cheap, it runs on anything, and it cannot distinguish a person from a cat, a car, a shadow or rain. PIR (passive infrared) detection, used by almost every battery camera to avoid burning power on constant analysis, looks for a moving heat source — which is why battery cameras produce fewer shadow and rain alerts but are reliably fooled by warm air from a vent and by direct sun warming a wall. Person, pet and vehicle detection runs a classifier over the image and only alerts when it recognises the category you asked for. That is the one that genuinely solves this problem.
The commercial catch: on several of the biggest brands, that classification is the feature the subscription pays for. Wyze puts its smartest AI events behind Cam Plus; Ring gates smart alerts behind its plan. On the three cameras below it runs on the camera itself at no recurring cost, which is why they are the picks on a page about this specific problem. Camera subscription costs lays out what the paid route adds up to.
Fix four: turn infrared off if the camera is behind glass
A camera indoors pointed through a window is a legitimate and often excellent setup, but its infrared LEDs reflect off the glass and wash the whole frame out — and that blown-out, flickering image generates motion events all night. Two fixes: push the lens flush against the glass so no light can get between camera and window, and turn the camera's infrared illumination off entirely, relying on a porch or street light. The same applies to a camera under an overhang with a white soffit above it, which bounces infrared straight back.
Fix five: deal with the insects, because they are probably the cause at night
If your camera is quiet all day and alerts constantly between dusk and dawn, the answer is almost always insects. The camera's own infrared LEDs attract moths and spiders, which then cross the lens inches away and look like a large fast-moving object. Spiders in particular build webs directly across the housing, and a web strand moving in the breeze triggers alerts all night, every night. The fixes are mundane and they work: wipe the housing and clear webs monthly, mount the camera a foot or two away from the wall rather than flush against it, and if the camera supports an external infrared illuminator, use one positioned away from the lens so the insects gather there instead.
Fix six: tune sensitivity last, and tune it per camera
Sensitivity is the setting everyone reaches for first and it should be last, because lowering it reduces all detections rather than the wrong ones — you are trading away real events, not noise. Once placement, zones and detection type are right, a modest reduction cleans up the remainder. Do it per camera rather than globally: a doorbell camera watching a three-foot porch and a driveway camera watching forty feet need completely different settings, and a single site-wide value is wrong for both.
Fix seven: require two signals before you call it an alarm
The most robust fix is not a camera setting at all. If you also run sensors, make the alarm condition motion plus something else — a contact sensor on the door, a second sensor in the hallway — rather than motion alone. A cat triggers one sensor; a person coming through a door triggers two. That logic is what separates a system that you trust from a stream of notifications you swipe away, and it is covered in smart home security automations. If you have not reached the sensor stage yet, how home security sensors work explains why PIR behaves the way it does.
When the camera is the problem
There is a point at which no amount of tuning helps. If the camera only offers raw motion detection, has no activity zones, and gates person detection behind a subscription you do not want to pay, you are managing a limitation rather than fixing a fault. The three cameras below all do category detection on the camera itself with no plan, which is the specification to buy on if false alerts are the reason you are shopping. For the broader field see best no-subscription security cameras, and if the camera is also dropping offline rather than over-alerting, why is my security camera offline is the page for that.
