Skip to main content
All CollectionsKnowledge Base
FAQs - Technical & Operational
FAQs - Technical & Operational
Updated over a week ago

How is SafeZone-edge installed and licenced?

Initial installation and licencing of the SafeZone-edge product is designed for simplicity. As the application is embedded on a camera or encoder, there is no server-based software to install, configure or maintain.

When deploying the application on the AXIS platform, software installation and licence key activation is performed through the SafeZone-edge Setup Interface.

The SafeZone-edge Setup Interface run on any typical windows laptop connected to the same network as the camera. This offers the installer access to the necessary installation and licence activation files. Once these have been selected and installed on the camera or encoder, SafeZone-edge can be calibrated and configured.

The SafeZone-edge Setup Interface incorporates a multi-camera installation feature which allows the user to install the application on several devices in a single click.

For details on downloading and accessing software installation and licence activation files, refer to the Configuration Guide that is appropriate to your device platform.

Note, this currently only applies to AXIS devices.


How is SafeZone-edge configured for operation?

SafeZone-edge has designed setup around the installer, instead of the more typical industry approach to video analytics setup, which uses multiple configuration screens and parameters.

Configuring and operating SafeZone-edge is extremely simple thanks to the SafeZone-edge Setup Interface which automates and streamlines the task of installation. It provides an optimised workflow allowing the user to carry out installation of the application, calibration of the camera to work with the analytics, definition of detection zones and scenarios, and setup of alarms. It provides a single intuitive toolset to install, activate, calibrate and configure the application – as well as providing access to alarm setup screens.

The typically time-consuming step of calibrating cameras has been automated and is followed by the automatic definition of a minimum detection area and default sterile zone. This creates a basic intrusion detection scenario with very minimal intervention and effort from the installer – and will be sufficient for many security situations. More advanced scenarios can be defined using an installer-friendly tool that is more intuitive than conventional intrusion detection systems.

Unlike the complex and time consuming manual calibration associated with many other VCA products, SafeZone-edge incorporates an advanced auto-calibration feature to significantly simplify and speed-up deployment – particularly when installing across multiple cameras.


How long does it take to install SafeZone-edge and configure it?

The installation and configuration of SafeZone-edge is very simple and is performed through the SafeZone-edge Setup Interface which enables an intuitive and quick SafeZone-edge installation and configuration process - less than 10 minutes per camera.


Who can install and configure SafeZone-edge?

SafeZone-edge does not have to be installed by a video analytics specialist. Any System Integrator or Installer who is used to deploying video surveillance systems can easily install and configure SafeZone-edge using the SafeZone-edge Setup Guide.


Is there any special equipment needed?

SafeZone-edge runs directly in the compatible Axis camera or encoder so there is no server-based or PC-based processing capacity to deploy. A PC is required to run the SafeZone-edge Setup Interface which aims at managing and configuring SafeZone-edge on multiple Axis devices. Any typical PC or laptop will be sufficient. The minimum specification for this PC is:

  • Windows XP, Windows 7, Windowds 8 or Windows server 2008 R2

  • Processor Core2 duo or better

  • 2GB RAM

  • Hard Disk 80GB


Is SafeZone-edge suitable for outdoor and indoor application?

SafeZone-edge is suitable for both outdoor and indoor deployments.

SafeZone-edge is an i-LIDS® approved primary detection system for operational alert use in outdoor sterile zone monitoring applications. As a consequence, SafeZone-edge is the ideal solution for outdoor sterile zones surveillance.

SafeZone-edge is also suitable for indoor application. In order to get a high level of performance, it is important that the installation meets the SafeZone-edge prerequisites (see the SafeZone-edge requirements sheet).


What makes the SafeZone-edge calibration assistant unique?

In short, SafeZone-edge has redesigned the critical task of calibration around the installer. Instead of multiple screens and manually defined parameters, the Setup Interface incorporates an intelligent calibration assistant and workflow. This uses a brief video clip, recorded from each camera, to extract and process the necessary calibration information. This simply requires an installer to walk through the field of view of the camera, with the system able to detect the presence of the installer to trigger the start/stop of the recording and the subsequent processing step.

What is more – and what really marks SafeZone-edge out as unique – is its ability to automatically queue and calibrate multiple cameras using video captured from a single installer walkabout. Any cameras that for any reason fail auto-calibration are queued automatically for repeat capture, or are else manual calibrated using a simple, intuitive editor. This is the closest thing to ‘one click’ setup that is available on multi-camera VCA deployments – making SafeZone-edge truly class leading...


What is the range of detection?

SafeZone-edge can detect up to 300 metres depending on camera chosen and its deployment, but up to 50m is more typical.

The maximum detection distance depends on two criteria:

  • The Axis camera lens. A higher focal range will allow a longer detection distance.

  • The minimum pixel size a human must cover in the image to be detected by SafeZone-edge. This minimum pixel size is given as a percentage of the image height. The pixel height of a standing person must be at least 10% of the image height (e.g. if the height of the visualized image is 576 pixels, a person to be detected must cover at least 58 pixels in height). This percenatge can be less if the camera is a thermal camera.

Please refer to the SafeZone-edge Axis Compatible Devices in the "Downloads" section for details on detection ranges fro compatible cameras.


What objects and situations SafeZone-edge is able to detect?

SafeZone-edge offers four types of scenario to detect the situations as follow:

  • Intrusion: an alarm is triggered when a person and/or a vehicle enters a zone defined on the ground (from any direction and with any trajectory).

  • Loitering: an alarm is triggered when a person and/or a vehicle remains in a zone defined on the ground greater than a predefined number of seconds.

  • Zone-Crossing: an alarm is triggered when a person and/or a vehicle passes through two or more zones defined on the ground in a given sequence.

  • Conditional: an alarm is triggered when a person and/or a vehicle enters a zone defined on the ground but did not first pass through one or more zones defined on the ground.


How does SafeZone-edge report an alarm?

SafeZone-edge exploits native AXIS alarm outputs and is integrated with Genetec and Milestone.

SafeZone-edge alarm outputs:

  • TCP notification (ACAP framework): send user-configurable text message via TCP to an IP address and port.

  • HTTP request (ACAP framework): send http requests with user-configurable content to a URL.

  • FTP upload (ACAP framework): save individual images on an FTP server using a user-configurable number of frames before and after the alarm event.

  • Email (ACAP framework): send a video clip of the alarm event within an email.

  • Digital Output port (ACAP framework): activate an electrical output port at the rear of the AXIS device.


How does SafeZone-edge behave in bad weather conditions?

SafeZone-edge has excellent performance even in bad weather conditions (e.g. wind and rain) thanks to an advanced reasoning engine which automatically mitigates external effects including: adverse weather, camera movement, shadows and variable lighting, which can often lead to false alarms.


Is the performance different in day and night?

SafeZone-edge offers a high level of performance during the day and night and is able to compensate automatically the passage from one to the other.


For which devices is SafeZone-edge currently available?

The launch device platform for SafeZone-edge is the world-leading AXIS range of network video cameras and encoders. This currently includes devices based on the AXIS Camera Application Platform (ACAP), which spans around 100 devices – including encoders. Specific models are detailed in the "SafeZone-edge Axis Compatible Devices" document that can be found in the downloads section.


How easy is it to incorporate SafeZone-edge into existing security setups?

SafeZone-edge is designed for seamless incorporation into existing security and surveillance setups. It supports full integration into Milestone and Genetec video management systems (VMS) and a range of alarm output formats can be used, depending on the native alarm outputs of the cameras or encoders. For example, SafeZone-edge is able to issue notifications via the AXIS electrical alarm output.

The Outputs Integration Guide will guide you in how to exploit the multiple output possibilities provided by the Axis cameras as well as the open-format metadata outputs that we are providing for you to integrate into more advanced setups.


For which operational applications is SafeZone-edge most suited?

SafeZone-edge features four intrusion detection scenarios to address typical perimeter security and sterile zone monitoring requirements. It is designed for medium range detection and is capable of operating in a range of challenging environments, including harsh weather and other demanding conditions. Typical outdoor sterile zone monitoring applications include: perimeter security at small to medium size commercial properties, industrial facilities or storage sites; the rear or loading bay of a retail unit; high-value asset protection e.g. car dealerships; other vulnerable or secure sites e.g. education campuses and sports facilities. It is also suitable for internal sterile zone monitoring application, such as controlled zones in transportation hubs, hospitals, commercial facilities and other sites.


What are the limitations of SafeZone-edge?

Despite its resilience to a range of external effects, SafeZone-edge may not always be suitable in certain environments or situations. This is typically where false detections or missed detections may result from: vegetation that is moving close to or immediately in front of the camera; a plastic bag or flag that is moving in the detection zone; inadequate scene lighting; or insects in front of the camera.


What updates and upgrades will be made available for SafeZone-edge?

During the support life of each generation of SafeZone-edge product (e.g. v1.x), product fixes are free and simple to download, prior to uploading onto the camera or encoder. Upgrades to the next generation release of the product (e.g. v2.x) require the purchase of a full licence and re-installation of the product.


How can a PTZ camera be configured so that SafeZone-edge only runs while the camera is in the preset selected for analysis

In order to make an Axis camera stop and start according to PTZ presets you need to define 2 Action Rules in the Axis camera. One Action Rule that stops SafeZone-edge when the camera moves and one Action Rule that activates SafeZone-edge when the camera comes to the preset for SafeZone-edge has been configured.

Action Rule to stop SafeZone-edge when PTZ camera moves:

1. You need to give it a name (in this example: PTZ_move_stop_SZe)
2. The 1st trigger is "PTZ"
3. The 2nd trigger is "Moving"
4. "Is Moving" should be left at Yes
5. You can choose a schedule (in this example: always)
6. The Action Type is "Send Notification"
7. Click on New recipient to create a new recipient. This will open a new window:


In the recipent window do the following things:

1. Enter a name (in this example: SafeZone_edge_stop
2. Select HTTP as type
3. Enter the following url:
http://IP_OF_THE_CAMERA/axis-cgi/applications/control.cgi?action=stop&package=SafeZoneEdge&reload_page=yes
replacing IP_OF_THE_CAMERA with the real IP address of the camera.
Example:
http://192.168.2.22/axis-cgi/applications/control.cgi?action=stop&package=SafeZoneEdge&reload_page=yes
4.Enter username and password for the camera.
5. You can test the connection, but be aware that this test will not stop the analysis
6. Click OK

You will come back to the Action rule configuration:

There are some additional fields that have appeared, but you only need to click "OK".


Action Rule to start SafeZone-edge when PTZ camera comes back to the preset chosen for SafeZone-edge:

1. You need to give it a name (in this example: PTZ_preset_start_SZe)
2. The 1st trigger is "PTZ"
3. The 2nd trigger is "Preset Reached"
4. Select the preset that SafeZone_edge is configured for (in this example: preset_groupe_Vehicle (H)
5. "Preset reached" should be left at Yes
6. You can choose a schedule (in this example: always)
7. The Action Type is "Send Notification"
8. Click on New recipient to create a new recipient. This will open a new window:

In the recipent window do the following things:
1. Enter a name (in this example: SafeZone_edge_start
2. Select HTTP as type
3. Enter the following url:
http://IP_OF_THE_CAMERA/axis-cgi/applications/control.cgi?action=start&package=SafeZoneEdge
replacing IP_OF_THE_CAMERA with the real IP address of the camera.
Example:
http://192.168.2.22/axis-cgi/applications/control.cgi?action=start&package=SafeZoneEdge
4.Enter username and password for the camera.
5. You can test the connection, but be aware that this test will not necessarily start the analysis
6. Click OK

You will come back to the Action rule configuration:

There are some additional fields that have appeared, but you only need to click "OK".

Once the 2 Action rules have been defined the analyse should stop when the camera is moving and be stopped while in a different preset, but when the camera comes back to the preset chosen for SafeZone-edge the analysis should start after about 20 seconds.

Be aware that if you are using the Setup interface to verify if the analysis is running or not you should select the camera and click F5 and the eye icon regurarly so that status and the overlay is updated.


Does SafeZone-edge analyze the image before or after the Axis camera has stabilized the image?

SafeZone-edge will analyse the image after it has been stabilized by the Axis camera.


Calibration: How can I make sure that my calibration is as good as possible?

Make sure you read the documentation on calibration and especially that you respect the requirements in terms of camera placement (height, orientation, no roll, lense, etc...) !

Automatic Calibration

  • Make sure the installer walks in the whole image and especially that he covers both the left and right side and both near and far. Avoid walking in an area you already covered.

  • Make sure you verify the calibration using the tools provided

    • Display the grid and make sure it corresponds to the ground in the image

    • Display the person avatar and move it around all over the image. Check that its size and direction correspond to that of the installer everywhere in the scene (even in parts where he did not walk)

    • Check if the horizon corresponds to the real world horizon. Note: This is sometimes hard to see depending on the scene.

    • Press on "Detection results" and look at the replay of the calibration walk being played through the analytics engine. If you notice:

      • Several yellow or blue boxes around the installer

      • Several missing boxes around the installer or not fully covering him

    • Then you should redo the automatic calibration or go to manual calibration.

Manual Calibration

  • If you place person sticks, make sure they have the right size and also the right orientation related to the person/object you are using to calibrate

  • Make sure you have person sticks well distributed across the image: left and right should be covered. Back and front should be covered.

  • If you use horizontal sticks, make sure there are at least 2 and that they are all parallel in the world – NOT in the image.

  • Check using measure tool that scene dimensions are as close as possible to the real ones.


What happens if I upgrade the firmware on an Axis camera?

In most cases you need to reinstall the SafeZone-edge application onto the camera. The configurations are kept.

UPDATE: Starting from firmware 7.10 no actions should be needed on the SafeZone-edge side.
UPDATE 2: There is an issue when upgrading to firware 8.30.1 which means you need to reinstall the SafeZone-edge application onto the camera again. The configurations are kept.


Can SafeZone-edge run on a F41 or A F44?

SafeZone-edge can run on F41 depending on the FoV(Field of View) of the camera as SafeZone-edge doesn not perform very well on FoV larger than 110 degrees.

For F44, there are 4 connections on this unit and we do not know if there are also 4 processors. If there are 4 processors and one IP per processor, you can use SafeZone-edge on all channels. If there is one processor, you can only use SZ on channel 1.

Be careful as F44 and F41 are often used with Fisheye, discreet or pinhole cameras which in most cases do not suit to SafeZone-edge use cases and requirements that is why they are not added them into the official compatible list to avoid the use of cameras not suitable for SZe.


What is the minimum camera installation height?

We strongly recommend not having a camera below 2.5m except if the detection is very short range (15m max) and in that case you can put camera at 2 meters height.

When camera is mounted at 2.5m, official requirements are stating max detection distance is 20m. However, if we limit the detection scenario to pure Intrusion scenario without discrimination between human and vehicle in alarms and the environment is a sterile zones (no big animal like sheep) with a flat ground without the presence of significant vegetation, we can consider extending the max detection distance for cameras mounted at 2.5m height up to approx 100m, provided the focal length of the camera allow this distance to be achieved, see max detection distance using the SZ Camera Placement Tool.

The blind spot under the camera will depend on the camera ref and its focal length, you can estimate it using the SZ Camera Placement Tool (for 100m detection distance, it will be probably between 15m and 25m).
We would recommend however a test to be conducted in the targeting environment and used camera to validate in real condition of use this extended detection distance.

Official table (as of August 2016):


Why reducing pole height affects the object classification

The lower is the camera and the greater is the detection range, the more significant is the image crushing effect (i.e. the 2D position and dimension of a moving person remains unchanged or close while his 3D location changes significantly in the real world). As a consequence, the 3D dimension (width and height in meters) of the moving objects is not estimated by the system with enough precision to calibrate in an effective way (is it a human, is it a vehicle or is it an animal) based on the 3D dimension. That is why it is important that your camera installation meets the SafeZone requirements in term of camera height.


Does the color palette chosen for a thermal camera affect the SafeZone-edge analysis?

The color palette chosen will be the one used by the analysis. SafeZone-edge has been developed for and tested extensively with the White-hot palette. Other palettes have been tried occasionally and seem to work ok, but there may be issues that haven't been discovered because of limited testing. One SafeZone-edge user reported to us that he had performed several tests with Ice-and-fire and claimed that the results were good with this palette. You are obviously free to try other palettes, but we recommend using the White-hot palette.


What is the smallest area required for an intrusion of a human or a vehicle to be detected ?

One important SafeZone requirement is: Humans or vehicles to be detected are fully visible in the image during at least 2 seconds and are not obscured by other fixed or moving objects.

As a consequence, we recommend:

  • for human detection: an area size of at least 3m length in the direction of the human movement.

  • for vehicle detection: an area size of at least 15m length in the direction of the vehicle movement.


What is the minimum size for a human to be detected in an image

The minimum size for human detection in image depends on the camera type and processor:

  • For thermal cameras:

    • 3.5% of image height for Artpec-5 cameras (Q194x)

    • 3.5% as well Artpec-4 ones (Q193x) if long-range is activated

    • 7% for Artpec-4 cameras without long-range

  • For color cameras:

    • 8% for Artpec-4 cameras

    • 5% otherwise


How to reduce false alarms

This article includes some tips and tricks on how to reduce false alarms.

CALIBRATION

Calibration mistakes can be a reason for false alarms as the 3D information used to classify humans and vehicles shapes would then be wrong and this 3D information is crucial for us to discriminate objects and classify them as intruders or as objects which we do not care about.

When performing manual calibration, check every aspect:

  • The horizon should correspond to the real horizon. When it is not seen in the image, it is hard to evaluate.

  • The grid: Should be “laying” on the real ground.

  • The avatar: Slide him around and he should have the size and direction of the installer doing the calibration.

  • Horizontal and Vertical measure: They can help you adjust the person-sticks so that you get good measurements all around, but especially in the zone where you want to detect.

  • Replay the detection video: You should not see any wrong classification (human taken as vehicle, etc…) or yellow boxes.


Any field of view that are near top-view configuration (camera with a lot of tilt) can yield problems. Often, they are out of prerequisites and, when they are not, the part of the scene closest to the camera could lead to false alarms with animals (such as dogs and cats) because they are seen from above and could be a human seen by above.

CAR HEADLIGHTS

Several false alarms are due to car headlights which are a known cause of problems, especially during rainy nights where the ground is reflective. However, SafeZone has some internal mechanisms to help reduce the false alarms due to headlights.

The main thing to know with headlights is that we detect them in order to detect cars in the intrusion zone when “Vehicle Detection” is activated in the scenario. On some cameras, the only thing you really see of a car are the headlights and we are looking for them in the scene. However, any headlights effects coming into the zone to protect will also be classified as a vehicle and thus trigger a (false) alarm. So if vehicle detection is not absolutely necessary (for example, if the site is closed at night), we really recommend to not activate vehicle detection to avoid false alarms due to headlights. When we want to detect vehicles as well, the probability of false alarms rises due to the headlight detection and careful positioning of camera and zones has to be performed.

For any scene having headlights effects coming from outside the scenes, we advise putting loitering scenarios of a few seconds instead of intrusion scenarios. The larger the field of view, the more seconds you can put in the scenario. This will reduce the risk of false alarms.

SCENARIO DEFINITION

When drawing scenario zones, they are situated on the ground. Thus, they should not go “on the walls” but stay limited to the base of the walls. When the scenarios zones includes walls or fences, and vehicle detection is activated, there is a higher probability of false alarms due to headlights as they will reflect them, especially the fence whose posts could have some human aspect, so avoid to include walls and fences in the scenario zones.

MISCELLANEOUS COMMENTS

  • Regarding water droplets, the general advice is to protect the camera from the rain.

  • When evaluating a scene to protect, beware of any secondary lighting (such as a passing road in the back or on the side) which can alter the image quality at time and then degrade the performances.

  • We advise to (re)read the section of the Setup User Guide related to “Installation considerations and prerequisites”.

  • To see what is detected before they an object gets classified as human or vehicle, you can press Ctrl-Alt-K in the camera and yellow boxes will be displayed. They represent objects that we are following in order to decide if they are humans or vehicles. Yellow objects (aka “Unknowns”) never trigger alarms except if they become human/vehicle and they had passed into a scenario zone when they were still yellow.

  • Vehicle detection is not guaranteed when the vehicle takes more than 30% of the field of view.

Did this answer your question?