A Practical Guide to Eliminating Avigilon False Alerts
Avigilon systems are renowned for their powerful analytics and high-quality imaging. However, even the most advanced systems can generate false alerts if not configured correctly. Constant notifications from irrelevant events like moving shadows, weather, or animals can lead to "alert fatigue," causing users to ignore genuine threats.
This guide provides professional installers and system administrators with a structured approach to diagnosing and eliminating false alerts within the Avigilon Control Center (ACC). By fine-tuning your system's settings, you can dramatically increase its accuracy and ensure that every notification is meaningful.
## The Core Problem: Pixel Motion vs. Video Analytics
The most common source of false alerts is an over-reliance on traditional, pixel-based motion detection. This legacy technology detects any change in pixels within a scene, making it highly susceptible to triggers from non-threatening sources.
The solution is to leverage Avigilon's advanced video analytics. Instead of just detecting pixel changes, Avigilon analytics use sophisticated algorithms to detect and classify specific types of objects, primarily People and Vehicles. This is the single most effective tool you have for reducing false alerts.
## Step 1: Switch from Motion Detection to Analytic Events
If your rules and alarms are currently being triggered by "Motion Detection," your first step is to migrate them to "Analytic Events."
- Access Camera Settings: In the ACC Client, right-click the camera you want to configure and select Setup.
- Disable Motion Detection: Navigate to the Motion Detection tab. If it's enabled, turn it off. This will immediately stop alerts from leaves, rain, and shadows.
- Enable and Configure Analytics: Go to the Analytic Events tab.
- Ensure "Analytics" is enabled.
- Select the appropriate Analytic Scene Mode (e.g., Outdoor, Indoor, Large Area). This helps the camera calibrate its algorithm to the environment.
- Enable Classified Object Detection. This is the key feature that tells the camera to specifically look for people and vehicles.
## Step 2: Fine-Tune Your Analytic Rules in ACC
Once analytics are enabled on the camera, you need to create rules in ACC to trigger alarms based on this more intelligent detection.
- Create a New Rule: Go to Site Setup > Rules > Add.
- Select the Event: In the "Select Rule Events" dialogue, do not choose "Motion Detected." Instead, choose an analytic event. The most common and effective one is "Object detected in an area."
- Define the Parameters:
- Object Type: This is the most critical setting. Select "Person", "Vehicle", or "Person or Vehicle". This immediately filters out all other types of motion.
- Area of Interest (ROI): Draw a precise, tight bounding box around the specific area you want to monitor. Do not draw a box over a busy road if you only care about your driveway. Be specific. Exclude trees, public footpaths, and other sources of irrelevant movement.
- Thresholds (Advanced): For even greater control, you can adjust thresholds like the minimum object size and the confidence level required to trigger an event. For example, you can require an object to be on screen for a minimum duration before an alert is sent.
## Step 3: Address Environmental Factors
Even with advanced analytics, the camera's physical environment can play a role in generating false positives.
- Lighting Conditions: Drastic changes in lighting can sometimes be challenging. Ensure the area is adequately and consistently lit, especially at night. Use cameras with high-quality Wide Dynamic Range (WDR) and adaptive IR illumination to handle challenging light.
- Camera Placement: Position the camera to have a clear, unobstructed view of the detection area. Avoid pointing the camera directly at sources of bright light or through windows with significant reflections.
- Obstructions: Spiders building webs directly in front of the lens are a common issue. Ensure lenses are kept clean. Tree branches swaying in the wind within the detection zone can also cause issues if they are large enough to be momentarily misclassified. Prune back any foliage that could interfere with the scene.
## Step 4: Use Self-Learning Analytics Effectively
Avigilon's self-learning analytics continuously analyse the scene to determine normal movement patterns. You can use this to your advantage.
- Teach by Example: When you receive a notification for an event, you can provide feedback. In the ACC client, you can classify events as relevant or irrelevant. This feedback helps the system learn and improve its accuracy over time.
- Be Patient: It can take a few days or even a week for the self-learning analytics to fully calibrate to a new scene. Allow the system time to gather data and refine its understanding of normal activity.
By shifting from outdated pixel-based motion to object-based analytics and carefully tuning your rules, you can transform a noisy, frustrating system into a highly accurate and reliable security tool.