GunTrack

Use the GunTrack module to execute the "Detect weapon" scenario in public spaces such as schools, transport hubs, banks, and critical infrastructure, which is to detect in real-time different types of weapons pulled out or carried by bad actors. Like other real-time video analytics modules of the IREX Platform, GunTrack helps operators reduce the response time down to 1 sec and below in control rooms, real-time crime control, and fusion centers, thus increasing public safety and human rights protection.

GunTrack in general

How it works

The convolutional neural network trained to detect "Gun" objects analyses every frame in the incoming video stream to find guns in it; if found, it builds an object box around the gun and calculates a similarity score to pass those to the event setting check module regarding the object box location, similarity score value and the number of frames in a row holding that gun.

Once the event conditions are met, the module generates a "Gun detected" event with a "Critical" priority by default.

Computational complexity

The computational complexity of GunTrack is estimated as "High", which means that approximately 100% of the processing core is in use while the module runs. The IREX developers continuously investigate the module performance on various CPU and GPU platforms to offer the best model-hardware solutions to the customers.

Camera requirements

The best performance of GunTrack  is guaranteed under the following conditions:

• Camera position: fixed camera
• Resolution: from 1280x720 (HD Ready) to QHD 2592x1944.
• Tilt: +15° to 75°
• Pan: -75° to +75°
• Roll -5° to +5°
• Image quality within the region of interest: over 250 px per meter.
• Height: 3 - 6 meters above ground level.
• Minimum detected object size: 4096 (64x64) px.

Export module settings

Export the module settings to a *.xml file, e.g., to pass it to qualified IREX personnel to find a solution to a problem with the module performance, should it happen. To do so:

1. Find the camera running the module in the User menu — Settings — Cameras, and click Video  analytics settings as shown:    

2. On the VIDEO ANALYTICS tab on the camera settings pane on the left, click the three-dot menu 



Export analytics configurations of the module — to export the module settings on the MAIN sub-tab.

Export analytics configurations of the model — to export the entire module configuration, including the rules, if any.

Enable GunTrack

1. Find the camera to enable the module in the User menu — Settings — Cameras section and click Video analytics settings as shown.    

2. On the <module_name> list, click the name of the module.
3. Slide right the On/Off slider 
4. Click Apply at the bottom of the pane.

How to get alerted on guns

1. Think over the gun monitoring scheme as for the camera locations, cross-monitoring, view directions and zooms  and other factors.
2. Enable GunTrack on those cameras and set it up for gun detection with "Critical" priority assigned to the Gun detected event.
3. Build an alarm monitor having Analytics modules - GunTrack - Gun detection checkmarked.
   

Set up gun detection

On the module settings pane that opens, do:

1. Edit the Region of interest,which defaults to the full frame.

To move a vertex, point the cursor onto the vertex and drag and drop it.
To delete a vertex, double-click it.
To add a vertex onto an edge, double-click where the vertex should be.
To move the region as a whole, place the cursor inside it and drag-and-drop.

2. Move the Threshold slider to define a gun similarity threshold to trigger the Gun detected event.

Select a value with account for the module performance stats, which depends on the typical, expected gun image size as viewed by the camera; values too high may lead to too many misses.

3. Move the Frames slider to define the number of frames in a row to trigger the Gun detected event.

The similarity score must be exceeded in all these frames.

4. Depending on the system hardware configuration at the user's end, an extra option can be offered to run the calculations on GPU. Switching it on significantly speeds up the module performance.
5. Click the Apply to start streaming the “Gun detected” events onto the Events screen.

Tips to use GunTrack

• GunTrack performance is critical to the expected gun image size, so pay close attention to finding a good location for the camera, which is the one that ensures the right balance between the surveyed area and gun image sizes.
• Consider cross-monitoring of critical sub-areas by several cameras at various zooms and other event verification measures; thoroughly think over the response protocols to a "Gun detected" alarm.