Video analytics and camera calibration

Following years of slow market penetration due to low performance against high expectations, video analytics adoption is growing. Faster processing, more effectively designed algorithms and a more accessible price levels have seen an increased use of these technologies, not only for gathering security-related infomation and loss prevention information but also to provide valuable data to improve efficiency and business performance.

Supporting both analogue and IP cameras, including thermal cameras, Mirasys Video Content Analytics (VCA) can add greater detail and intelligence to your video management system. Increase understanding of motion paths and events.

Based on advanced hermeneutic detection algorithms to turn video into actionable intelligence, Mirasys VCA offers a range of options depending on your requirements such as:

  • An object entering a scene
  • An object exiting a scene
  • Objects appearing in the scene
  • Objects disappearing from the scene
  • Objects stopping in the scene
  • Directionality constraints – one way systems
  • Object counting – mixed with directionality can count objects going the wrong way in a one way system
  • Loitering
  • Object type – dogs, people, cars, vans, lorries
  • Object speed – with alarms if a speed limit is exceeded.

So how do you ensure a great VCA system?

The camera has to be setup correctly and the software correctly calibrated to reflect the cameras position and field of view as best as possible. We are heavily dependent on the camera calibration to allow Pythagorean theorem formulas to be used to work out the size of objects which we can then classify, we do this on each frame sent from the camera, which allows us to track the classified objects. If they then hit a zone or line we check to see what active rule to trigger. If the camera isn’t calibrated correctly or has been moved the engine is not able to classify the object properties such as height, area, speed and classification.

To calibrate the camera and set the Mirasys VCA, you need to know the height, tilt angle and the field of view of the camera.

camera-calibration

 

 

 

 

 

 

 

 

If you don’t know the parameters, below are few steps you can follow, when setting Mirasys VCA:

Step 1: Find some people in the scene

Find some people, or some people-sized objects in the scene. Try to find a person near the camera, and a person further away from the camera. It is useful to use the Video Control button to pause the video so that the mimics can be accurately placed. Place the mimics on top of or near the people. Enter in the known height or estimated height.

camera-calibration_1

Step 2: Adjust tilt angle and camera height

Adjust the camera tilt angle and vertical field of view until both mimics are approximately the same size as a real person at that position in the scene. If an estimated height is entered then adjust this along with the other two parameters.
Click and drag the ground plane to change the tilt angle, use the mouse wheel to adjust the camera height and drag the slider to change the vertical field of view. The calibration parameters will be reflected in the edit boxes in the Camera Setup section of the web page. Then apply the changes.
Once the changes have been applied, the Calibration Status reflects whether the operation was successful, and if so, in which mode the camera is calibrated.

Step 3: Verify the setup

Once you are happy that the scene is correctly calibrated, verify the settings by dragging the mimics around and comparing them to other people or people-sized objects in the scene. The more locations in the scene where the calibration is verified, the more accurate the calibration will be.
If you find that the mimics get lost or disappear due to an odd calibration configuration, the process can be restarted by clicking the “Restore Defaults” button.

camera-calibration_-settings

Vertical Field of View

The vertical field of view is required when calibrating a camera. If you know the vertical field of view of the camera it can be entered directly. If however, you don’t know the vertical field of view, it can be determined from the following table, given the vertical image dimension and its focal length:

focal-lenght

focal-lenght_2

If you are unable to determine the vertical field of view for your camera from the above tables, leave the field of view at its default setting, 40°.

Use of PTZ cameras or cameras with automatic zoom lenses are discouraged from being used for server-based VCA, as the camera calibration, rules and counters will be invalidated by camera/lens movement, resulting in false alarms and incorrect calculations.

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s