AXIS Image Health Analytics - User manual

AXIS Image Health Analytics is an AI-based application that analyzes and learns the behavior of the scene to detect changes in the image. Use the application with compatible Axis cameras to detect image degradations and tampering attempts.

The application can detect the following changes in the image:

• Blocked image

• Redirected image

• Blurred image

• Underexposed image

With the event management system in Axis devices, or with third-party software, you can trigger actions based on the detections, for example notify staff to clean the cameras, or alert security staff.

For a complete list of Axis cameras that include AXIS Image Health Analytics, see axis.com/products/axis-image-health-analytics#compatible-products.

AXIS Image Health Analytics analyzes and learns the behavior of the scene to detect image degradations and tampering attempts. When the application detects a change in the image, it sends an event. Before you set up AXIS Image Health Analytics and create rules for events, consider the following:

Scenes with little or no variation

Scenes without any distinct features or details make it harder to detect changes in the image. This is typical for scenes where the background has little or no variation, like a field of grass, a blank wall, or a dark room.

For more information about scene requirements, see Scene suitability.

Sudden light changes

Turning on or off a light creates a sudden change in the scene. For cameras without built-in IR or sufficient illumination, an image that suddenly turns dark can cause the application to send events both due to an underexposed and a blocked image. For cameras with built-in IR, there is a risk that a sudden loss of light will make the application send an event during the switch to night mode.

Turning on a light can also change the appearance of the scene completely, for example by causing shadows.

Objects obstructing the view

The application can detect when something blocks the camera view, but it can't determine the intention. For example, trucks or buses that happen to stop in front of the camera can obstruct the view, which makes the application send an event for a blocked image.

If the camera view is completely blocked, the image will no longer be in focus and could also go dark. In this case, the application could send events for a blocked, blurred and underexposed image at the same time.

Spider web in front of the camera

Spider web in front of the camera is a common issue, especially during nighttime when the web can cause reflections in the image for cameras with IR illumination.

The application can help monitor issues with spider web in front of the camera by sending an event for a blocked or blurred image, but if spiders keep moving across the view it could cause the application to send the event several times.

Turning off the application

When you turn off the application, and then turn it on again, it relearns the behavior of the scene. For more information about relearning the scene, see Learning and relearning the scene.

Privacy masks

If you add a privacy mask to a camera view that is already running AXIS Image Health Analytics, the mask could make the application send an event due to a blocked image. To relearn the behavior of the scene with the privacy mask in the image, restart the application. However, no detections will be made in the masked area. For more information about relearning the scene, see Learning and relearning the scene.

1. In the camera’s web interface, go to Analytics > AXIS Image Health Analytics.

2. Click Start to start the application.

3. Click Open to open the application in a new browser tab.

4. Select the image changes that you want the application to detect.

   All detection settings are turned on by default.

5. • Blocked image: Sends an event when the view is obstructed.

   • Redirected image: Sends an event when the original orientation of the view changes.

   • Blurred image: Sends an event when the image is blurred.

   • Underexposed image: Sends an event when the image is too dark.

6. To learn more about the detection settings, see Detections and events.

7. You can adjust the sensitivity slider for each detection setting.

   A higher value makes the application more sensitive to changes in the image. For more information, see Sensitivity to image changes.

8. Go to Validation period to set a time threshold in seconds for each detection setting.

   The validation period is relevant if you only want the application to send events when the image change has occurred for a set time. For more information, see Validation period.

To learn more, check out our guide Get started with rules for events.

This example explains how to set up a rule in the camera’s web interface that sends an email when the image has been obstructed for more than 60 seconds, for example by a parked truck.

Before you start:

• Create an email recipient in the camera’s web interface.

1. In AXIS Image Health Analytics:

2. Turn on Blocked image.

3. Set Validation period to 60 seconds.

4. In the camera’s web interface:

5. Go to System > Events and click + Add a rule

6. Type a name for the rule.

7. In the list of conditions, under Application, select Image Health Analytics – Block.

8. In the list of actions, under Notifications, select Send notification to email.

9. Select a recipient from the list.

10. Type a subject and message for the email.

11. Click Save.

AXIS Image Health Analytics can detect the following changes in the image:

• Blocked image: The application detects when the image, or camera view, is obstructed, for example if something covers the camera.

• Redirected image: The application detects when the original orientation of the view changes, for example if someone moves the camera.

• Blurred image: The application detects when the image is blurred, for example due to raindrops on the lens.

• Underexposed image: The application detects when the image is too dark, for example if a light source in the scene suddenly disappears.

By default, the application is set to detect all of these changes. When a detection occurs, the application sends an event. You can set up rules based on these events in the device’s event management system or in the VMS. You can also set up rules based on the scene suitability.

Apart from sending an event when a detection occurs, the application also sends an event when the detected image change goes back to normal.

When you turn on AXIS Image Health Analytics, it immediately starts to analyze and learn the behavior of the scene. If the scene doesn’t contain any distinct features or details, it will be harder for the application to detect any changes. In this case, the application signals that the scene is unsuitable.

The application analyzes the scene continuously. This means that the suitability can change, for example if you change the camera’s zoom level or move or redirect the camera.

AXIS Image Health Analytics learns the behavior of the scene to detect when something changes. If you turn off the application, and then turn it on again, it will relearn the behavior of the scene. The application will also relearn the scene when it changes from unsuitable to suitable.

If your intention is to change the camera view, turn off the application and adjust the camera view before you turn it on again.

You can increase or decrease the application’s sensitivity to changes in the image. A higher sensitivity means that less of the image needs to change for the application to send events, which in most cases increases the number of sent events. A lower sensitivity on the other hand requires more of the image to change, which likely will decrease the number of sent events.

• Blocked image: The sensitivity is affected by how much of the image that is blocked or obstructed, compared to what the image looked like when learning the scene.

• Redirected image: The sensitivity is affected by how much of the image that is still visible, compared to what the image looked like when learning the scene.

• Blurred image. The sensitivity is affected by the level of blurriness in the image, compared to what the image looked like when learning the scene.

• Underexposed image: The sensitivity is affected by the level of underexposure in the image.

In AXIS Image Health Analytics, you can set a validation period for each detection setting. The validation period acts as a time threshold, and is relevant if you only want the application to send events when the detected change has occurred for a set time.

Example

In the scene, there is a risk of delivery trucks making temporary stops in front of the camera, which can block the camera view. You only want to be notified if the camera view has been blocked for two minutes, so you set the validation period in AXIS Image Health Analytics to 120 seconds.

This example explains how to set up a rule in AXIS Camera Station to alert the operator and record video when AXIS Image Health Analytics detects an image change.

Before you start

• You need:

• an Axis network camera with AXIS Image Health Analytics set up and running, see Get started.

• a computer with AXIS Camera Station installed

1. Add the camera to AXIS Camera Station

2. In AXIS Camera Station, add the camera. See the user manual for AXIS Camera Station.

1. Create a device event trigger

2. Click and go to Configuration > Recording and events > Action rules and click New.

3. Click Add to add a trigger.

4. Select Device event from the list of triggers and click Ok.

5. In the Configure device event trigger section:

6. • In Device, select the camera.

   • In Event, select one of the options for AXIS Image Health Analytics.

   • In Trigger period, set an interval time between two successive triggers. Use this function to reduce the number of successive recordings. If an additional trigger occurs within this interval, the recording will continue and the trigger period starts over from that point in time.

7. In Filters, set active to Yes.

8. Click Ok.

1. Create actions to raise alarms and record video

2. Click Next.

3. Click Add to add an action.

4. Select Raise alarm from the list of actions and click Ok.

5. Note

   The alarm message is what the operator sees when an alarm is raised.

6. In the Alarm message section, enter an alarm title and description.

7. Click Ok.

8. Click Add to add another action.

9. Select Record from the list of actions and click Ok.

10. In the list of cameras, select the camera to use for recording.

11. Select a profile and set the prebuffer and postbuffer.

12. Click Ok.

1. Specify when the alarm is active

2. Click Next.

3. If you only want the alarm to be active during certain hours, select Custom schedule.

4. Select a schedule from the list.

5. Click Next.

6. Enter a name for the rule.

7. Click Finish.

AXIS Image Health Analytics helps you monitor common image health issues, either caused by external factors or originating in the camera itself. When you encounter alarms based on AXIS Image Health Analytics, follow this step-by-step guide based on the type of alarm to find out which actions to take.

If you need more help, go to axis.com/support.