Security Camera AI Sensitivity Settings: How to Reduce False Alerts

Quick Answer: What AI Sensitivity Setting Should You Use?

What Does AI Sensitivity Mean?

AI sensitivity is a configurable threshold that affects how readily the camera accepts a detected object or event as significant enough to trigger an action.

The exact meaning varies between manufacturers.

For example, Reolink’s current smart-detection documentation describes a sensitivity scale from 1 to 100 on supported cameras, with higher numerical values representing higher sensitivity. Its NVR documentation describes lower settings as requiring a higher similarity to the selected object class, while higher settings can accept objects with lower similarity.

Therefore, do not assume that a sensitivity value such as 70 means the same thing on every brand.

AI Sensitivity vs Motion Sensitivity

This is one of the most important distinctions when configuring a security camera.

A camera can therefore have a high conventional motion sensitivity while using a lower AI person sensitivity.

That distinction is particularly useful when you want continuous recording but only want notifications for selected AI events.

Why Higher Sensitivity Can Increase False Alerts

AI classification is not binary in the real world.

A camera may encounter objects that resemble the target but do not clearly match it.

Examples include:

• Shadows resembling people
• Reflections resembling vehicles
• Animals partially hidden by vegetation
• Images or posters containing people
• Moving branches
• Objects seen only briefly
• Very distant people
• Objects heavily affected by darkness or glare

When the sensitivity threshold is increased, more borderline situations may qualify as detections.

Reolink similarly describes higher smart-detection sensitivity as allowing objects with lower similarity to the selected class to trigger detection.

Why Lower Sensitivity Can Miss Real Events

Lowering sensitivity can reduce nuisance alerts, but it may also make the system less responsive to borderline events.

This can happen when:

• The person is far away
• The person is partially hidden
• The object is very small
• The lighting is poor
• The subject moves quickly
• The subject appears only briefly
• The viewing angle is difficult
• The image is affected by rain, fog or glare

Reolink explicitly notes that lower sensitivity can reduce false alarms but may miss some events, while higher sensitivity increases detection of smaller movements or changes.

The Sensitivity Trade-Off

The objective is not to select the highest or lowest setting. The objective is to find the lowest sensitivity that reliably detects the events you actually care about.

Do Not Change Sensitivity Before Checking the Detection Zone

This is one of the most common configuration mistakes.

Suppose a driveway camera generates 50 unnecessary alerts from a public road.

Lowering vehicle sensitivity may reduce the alerts, but it can also make the camera less responsive to vehicles entering your driveway.

If the road is irrelevant, the better solution is usually to exclude it with a detection zone.

Our guide to security camera AI detection zones explains how to configure the monitored area before using sensitivity as a correction mechanism.

Reolink likewise recommends combining appropriate detection areas with sensitivity settings when reducing unwanted alarms.

Use Object-Specific Sensitivity Whenever Possible

A major advantage of modern AI cameras is that some systems allow different thresholds for different object categories.

For example:

These values are examples, not universal recommended numbers. Your actual starting point should depend on the camera and scene.

Person Detection Sensitivity

Person detection is usually one of the most important AI settings for home security.

Use higher sensitivity when:

• The camera monitors a critical entrance
• People may appear far away
• The area has few nuisance triggers
• Early warning is important

Use lower sensitivity when:

• The camera faces a busy sidewalk
• Many distant people are irrelevant
• Vegetation creates false classifications
• The camera is generating frequent nuisance alerts

Vehicle Detection Sensitivity

Vehicle sensitivity is especially important for driveway cameras.

A camera facing a road may see dozens or hundreds of vehicles that are irrelevant to your security objective.

Before lowering vehicle sensitivity, ask:

Can I exclude the road with a detection zone?

If yes, that is usually the first adjustment.

If vehicles inside the relevant zone are still generating false alerts, then reduce vehicle sensitivity gradually.

Pet and Animal Detection Sensitivity

Pet detection can be more difficult than person detection because pets may be smaller, move unpredictably and frequently become partially hidden by furniture, vegetation or other objects.

If your backyard camera repeatedly identifies vegetation or other objects as animals, try:

• Reducing animal sensitivity
• Using a detection zone
• Adjusting minimum object size
• Changing the camera angle
• Improving nighttime illumination

For deeper technical information, see how accurate AI pet detection is.

Package Detection Sensitivity

Package detection presents a different problem because packages may be small and stationary.

A package may also appear only after a delivery person leaves it.

Therefore, package detection may require:

• A suitable camera angle
• Enough image detail
• A relatively close detection area
• Appropriate object-size settings
• Correct sensitivity

Simply increasing sensitivity may not solve the problem if the package is too small in the image.

Object Size and Sensitivity Must Be Tuned Together

Sensitivity and object-size filtering are complementary controls.

Consider a camera overlooking a large driveway.

If tiny distant objects produce alerts, you have two possible approaches:

1. Lower sensitivity.
2. Set an appropriate minimum object size.

The second approach may be preferable because it specifically filters objects that are too small without necessarily reducing sensitivity for properly sized targets.

Supported Reolink systems provide minimum and maximum object-size settings for smart detection, specifically to filter objects outside the desired size range.

Alarm Delay Can Be Better Than Lowering Sensitivity

Some cameras provide an alarm delay or tolerance setting.

This can require an object to remain detected for a specified period before an alert is generated.

For example, if a camera supports a 2-second alarm delay:

A very brief detection lasting less than the configured threshold may not generate an alert.

Reolink specifically documents alarm delay as a way to reduce AI false alarms, alongside sensitivity, object-size filtering and detection zones.

This can be useful when the camera frequently produces alerts from very brief or borderline events.

Sensitivity vs Alarm Delay

Day vs Night Sensitivity

A camera may require different sensitivity settings during the day and at night.

At night, the image can contain:

• More noise
• Infrared reflections
• Headlights
• Strong contrast
• Moving shadows
• Insects near the lens
• Less object detail

Some systems allow sensitivity schedules. Reolink documents time-based sensitivity scheduling for supported motion-detection systems and notes that lower sensitivity at night can help reduce unwanted alerts in some situations.

However, do not automatically lower nighttime sensitivity. If nighttime intrusion detection is the primary purpose of the camera, reducing sensitivity too aggressively could increase missed events.

How to Tune Sensitivity for Day and Night

How Camera Angle Affects Sensitivity

Changing sensitivity cannot compensate for a fundamentally poor camera angle.

For example, a camera mounted toward a moving tree may continuously receive difficult visual changes. Lowering sensitivity may reduce alerts, but it also affects legitimate targets.

Changing the camera angle so that the tree occupies less of the frame may solve the problem without sacrificing sensitivity for people.

Reolink similarly recommends adjusting the mounting angle when environmental movement is causing false alarms.

How Lighting Affects Sensitivity

Strong lighting changes can cause conventional motion detection to trigger even when there is no person or vehicle.

Potential causes include:

• Sunlight entering the scene
• Moving clouds
• Vehicle headlights
• Street lighting
• Reflections
• Sudden transitions between light and shade

Reolink specifically identifies sunlight, shadows, clouds, rain, snow and external light changes as possible sources of unwanted motion alarms.

AI Sensitivity and Detection Accuracy

Sensitivity should not be confused with AI accuracy.

A camera with a highly accurate model can still produce poor results if configured incorrectly.

Conversely, increasing sensitivity cannot make an AI model understand an object that does not contain enough useful visual information.

For example, if a person is only a few pixels high at the far end of a property, increasing sensitivity does not create missing image detail.

In that situation, consider:

• A narrower lens
• A closer camera
• Higher resolution
• Better illumination
• A second camera

See security camera AI detection accuracy for a broader explanation of the factors affecting false positives and false negatives.

Recommended Sensitivity by Camera Location

Important: These are starting strategies, not universal numerical settings. Manufacturers use different sensitivity scales and AI models, so actual values should be calibrated against the camera’s real behavior.

A Practical Sensitivity Calibration Procedure

The best way to configure AI sensitivity is to change one variable at a time.

Step 1: Start at Medium

Use the manufacturer’s middle or default smart-detection sensitivity as the initial baseline.

Step 2: Configure the Detection Zone

Remove irrelevant roads, vegetation and neighboring areas before evaluating sensitivity.

Step 3: Select the Required AI Categories

Enable only the object types you actually need.

Step 4: Test a Real Person

Walk through the monitored area at the normal distance and speed expected in real use.

Step 5: Test the Farthest Important Point

Walk at the maximum distance where a person should still generate an alert.

Step 6: Test the Common False-Alert Sources

Observe:

• Trees
• Shadows
• Vehicles
• Animals
• Reflections
• Rain
• Nighttime movement

Step 7: Change Sensitivity by a Small Amount

Do not move directly from minimum to maximum.

Step 8: Run the Camera for Several Days

One test cannot represent all weather and lighting conditions.

Step 9: Compare Missed Events and False Alerts

Keep a simple log.

Choose the configuration that produces an acceptable balance rather than simply minimizing the number of alerts.

How to Calculate Your False-Alert Rate

A simple practical measurement is:

False Alert Rate = False Alerts ÷ Total Alerts × 100

Suppose your camera produces 80 person alerts and 8 are actually false:

8 ÷ 80 × 100 = 10%

You can compare this result before and after changing sensitivity.

Also track missed events because reducing false alerts is not useful if important detections disappear.

How to Find the Best Sensitivity Setting

Think of the process as an optimization problem.

Sensitivity Settings for High-Security Areas

Some locations justify accepting more alerts.

Examples include:

• Restricted gates
• Remote property entrances
• Warehouse perimeters
• Side gates
• Areas where unauthorized entry is rare but important

In such locations, higher sensitivity may be appropriate because a missed intrusion can be more costly than an occasional nuisance notification.

For low-priority areas, lower sensitivity may produce a better user experience.

Sensitivity Settings for a Front Door

A front door is usually a relatively controlled environment.

Start with:

• Person detection enabled
• Package detection if supported and needed
• Detection zone covering the approach
• Moderate person sensitivity
• Appropriate object-size range

If the camera sees a busy sidewalk, use a zone before significantly lowering person sensitivity.

Sensitivity Settings for a Driveway

A driveway often needs both person and vehicle detection.

Recommended sequence:

1. Exclude irrelevant road traffic.
2. Enable person detection.
3. Enable vehicle detection.
4. Set a reasonable object-size range.
5. Start at moderate sensitivity.
6. Test daytime and nighttime traffic.
7. Increase sensitivity only if legitimate events are missed.

Sensitivity Settings for a Backyard

Backyards often have more environmental movement.

Potential nuisance sources include:

• Trees
• Leaves
• Pets
• Insects
• Shadows
• Wind

Use detection zones aggressively and then tune person and pet sensitivity separately where the camera supports it.

Sensitivity Settings for Perimeter Cameras

Perimeter cameras present a different problem because targets may be far away.

Reducing sensitivity too much can cause missed detections because people occupy relatively few pixels.

Instead of simply increasing sensitivity, first consider whether the camera has:

• A sufficiently narrow field of view
• Enough resolution
• Good nighttime illumination
• Suitable mounting height
• A clear view of the approach path

For perimeter security, AI intrusion detection and tripwire/line-crossing analytics may also be more appropriate than ordinary motion alerts.

AI Sensitivity for PTZ Cameras

PTZ cameras require additional attention because the camera can move.

A sensitivity configuration that works at one preset may not work at another.

When testing a PTZ camera, evaluate:

• Each important preset
• Different zoom levels
• Auto-tracking behavior
• Nighttime presets
• Different target distances

Some professional analytics platforms allow analytics scenarios to be configured for different views or operating conditions.

What If Your Camera Has Only One Sensitivity Slider?

Some cameras provide only one overall sensitivity control.

In that situation, prioritize adjustments in this order:

1. Camera position
2. Detection zone
3. Object-type selection
4. Object-size filtering, if available
5. Overall sensitivity
6. Alarm delay, if available

A single sensitivity slider gives you less control, so physical positioning and detection-zone configuration become even more important.

What If Sensitivity Settings Do Not Solve the Problem?

If the camera still produces poor results at both high and low sensitivity, the problem may not be sensitivity at all.

Check:

• Camera angle
• Target distance
• Lens field of view
• Image resolution
• Nighttime lighting
• Motion blur
• Reflections
• Weather
• AI model limitations
• Firmware/software version

For a detailed explanation of these factors, see security camera AI detection accuracy and false alerts.

Common Sensitivity-Setting Mistakes

1. Setting Everything to Maximum

Maximum sensitivity can create alert fatigue and make important notifications harder to notice.

2. Lowering Sensitivity Instead of Fixing the Zone

If a road is producing unwanted vehicle alerts, exclude the road rather than making vehicle detection less responsive everywhere.

3. Changing Multiple Settings at Once

If you change sensitivity, object size and the detection zone simultaneously, you will not know which adjustment caused the improvement.

4. Testing Only During the Day

Nighttime conditions can behave very differently.

5. Ignoring Missed Events

A configuration that produces almost no false alerts may simply be missing legitimate activity.

6. Assuming the Same Number Means the Same Thing Across Brands

A sensitivity value of 70 on one manufacturer’s system cannot automatically be compared with 70 on another.

7. Using Sensitivity to Compensate for Poor Camera Placement

If the target is too small or poorly illuminated, sensitivity cannot manufacture missing image information.

Practical AI Sensitivity Decision Path

Best Practices for AI Sensitivity

• Start with the manufacturer’s default or medium setting.
• Configure the detection zone first.
• Use separate sensitivity for different object types when available.
• Use object-size filtering for tiny or oversized unwanted objects.
• Use alarm delay for very brief unwanted detections where supported.
• Test day and night independently.
• Change one setting at a time.
• Keep a record of false alerts and missed events.
• Do not assume a numerical sensitivity value is universal.
• Fix camera placement problems before relying on sensitivity.

AI Sensitivity Buying Checklist

• ☐ Does the camera provide adjustable AI sensitivity?
• ☐ Can person sensitivity be adjusted separately?
• ☐ Can vehicle sensitivity be adjusted separately?
• ☐ Can pet/animal sensitivity be adjusted?
• ☐ Are package alerts configurable?
• ☐ Does the camera provide detection zones?
• ☐ Does it provide object-size filtering?
• ☐ Does it provide alarm delay or tolerance?
• ☐ Can sensitivity be scheduled by time?
• ☐ Can alert schedules be configured separately from recording?
• ☐ Does AI continue operating locally if the internet is unavailable?
• ☐ Can you review recordings to verify false and missed events?

Frequently Asked Questions

What should I set my security camera AI sensitivity to?

Start around the middle of the camera’s available range and test it. Then reduce sensitivity if false alerts are excessive or increase it if legitimate events are being missed. There is no universal number because manufacturers use different scales and AI models.

Does higher sensitivity mean better security?

Not necessarily. Higher sensitivity can increase detection of borderline events but can also increase false alerts. The best setting balances detection reliability with manageable alert volume.

Why does my camera send too many person alerts?

Check the detection zone first. If irrelevant areas are already excluded, reduce person sensitivity gradually and consider object-size filtering or alarm delay if those controls are available.

Why does my AI camera miss people?

The target may be too small, poorly illuminated, partially obscured or moving too quickly. Check camera placement and image quality before simply increasing sensitivity.

Should I use different sensitivity for people and vehicles?

Yes, when the camera supports separate controls. A driveway, for example, may require relatively sensitive person detection while using a lower vehicle threshold to ignore distant road traffic.

Should nighttime sensitivity be lower?

Sometimes. Nighttime image noise, infrared reflections, headlights and other environmental effects can increase false alerts. However, if nighttime intrusion detection is critical, lowering sensitivity too much can increase missed events. Test the actual scene.

What is the difference between sensitivity and object size?

Sensitivity changes how readily the AI accepts a detection, while object-size filtering excludes targets that fall outside a configured size range. They can work together.

What is alarm delay?

Alarm delay requires an object or event to remain detected for a specified period before triggering an alert, where supported. It can reduce alerts from very brief detections.

Can sensitivity eliminate false alerts completely?

No. False alerts can also originate from camera placement, lighting, reflections, weather, moving vegetation, image quality and AI classification limitations.

Should I lower sensitivity if trees are causing false alerts?

Usually check the detection zone and camera angle first. If trees are outside the area you actually need to monitor, excluding them is preferable to lowering sensitivity for the entire image.

Can AI sensitivity improve detection at long distances?

Increasing sensitivity may help the system accept weaker or less certain classifications, but it cannot create image detail that is not present. A better lens, closer camera position or improved resolution may be necessary.

How often should I adjust AI sensitivity?

Configure it initially, then revisit it when the environment changes, such as after moving the camera, changing its field of view, adding lighting or experiencing seasonal changes in vegetation and weather.

Final Verdict

AI sensitivity is a tuning control, not a measure of camera quality.

The highest setting is not automatically the most secure, and the lowest setting is not automatically the most accurate. The correct setting depends on how large the target appears, how difficult the lighting is, how much irrelevant activity surrounds the target and how costly a missed event would be.

The best configuration sequence is:

Camera placement → detection zone → object type → object size → alarm delay → sensitivity → real-world testing.

If your camera is generating too many alerts, do not immediately turn down sensitivity. First determine why the alerts are occurring.

If the problem is an irrelevant area, change the zone. If the problem is tiny objects, change object-size filtering. If the problem is brief detections, use alarm delay. If the problem is borderline AI classification, adjust sensitivity.

Finally, evaluate the result by tracking both false positives and missed events. A camera that sends fewer notifications is not necessarily performing better if it is also failing to detect genuine security events.

For the most reliable AI configuration, combine this guide with AI detection zones and AI detection accuracy and false-alert management.

AI Security Camera Buying Guides & Technical Articles

Best AI Security Cameras by Detection Type

• Best Security Cameras With Person Detection in the USA
• Best Security Cameras With Vehicle Detection in the USA
• Best Security Cameras With Pet Detection in the USA
• Best Security Cameras With Package Detection in the USA
• Best Security Cameras With Facial Recognition in the USA
• Best License Plate Recognition Security Cameras in the USA
• Best security camera with intrusion detection

AI Detection & Classification Guides

• AI vs Motion Detection in Security Cameras: What’s the Difference?
• Person vs Vehicle Detection in Security Cameras: How AI Classification Works
• Pet Detection in Security Cameras: How Accurate Is AI?
• Security Camera AI Detection Accuracy: What Affects False Alerts?

AI Processing & Architecture

• Edge AI vs Cloud AI Security Cameras: Which Is Better?
• Security Camera AI vs NVR AI: Where Should Video Analytics Run?
• Do AI Security Cameras Need the Internet?

AI Detection Performance & Configuration

• Does AI Detection Work at Night? Security Camera AI in Low Light
• Security Camera AI Detection Zones: How to Set Them Correctly

Advanced AI Detection Features

• Security Cameras With Facial Recognition: How It Works and Privacy Issues
• Security Cameras With License Plate Recognition: What You Need to Know
• Security Cameras With AI Intrusion Detection: How It Works

AI Storage, Bandwidth & Subscription Guides

• AI Security Cameras Without a Subscription: What Features Do You Get?
• AI Security Camera Subscription Costs: What Features Are Behind Paywalls?

AI Security Camera Privacy

• AI Security Camera Privacy: What Happens to Your Video?

Related Security Camera article

• Best AI Security Cameras in the USA