Object Filter Fine Tunes Your Video Analytics

Object Filter Fine Tunes Your Video Analytics

Video analytics use algorithms to examine video feeds and can be configured to meet your specific security or business needs. In order to be notified of an event or capture useful data, you need to tell the video analytics software what you want the system to analyze and what you want the system to disregard. This is done when you use an object filter. An object filter is the filtering of objects (people, vehicles, objects) that meet certain characteristics. Depending on which video analytics product you work with, filtering can be inclusive or exclusive of objects.

Inclusive Object Filter

In the previous two articles, we looked at license plate recognition and video analytics for retail. These applications are designed to let you filter specific items to yield the results that you want. For instance, when you configure license plate recognition software to filter for vehicles that are red, the software will provide you with video feeds that contain a red vehicle.

In retail, you can configure video analytics to show you how many people stopped to look at a display. In addition to people counting, using a dwell filter with time limits will provide you with this information.

Exclusive Object Filter

In order to minimize false alarms, object filtering is used for objects or conditions that should be excluded, such as rain, snow and wind, trees, flags, shadows…

Making the Best Use of Your Surveillance Video

There are a multitude of video analytics object filters that can be combined to produce the desired results. Whether inclusive or exclusive, many object filters fall into these categories: color, direction, motion, size, speed and time.

Let’s take a look at each and how a combination of two or more factors can create specific detection rules for schools, multi-dwelling / multi-tenant units and a retail environment.

Video Analytics for Schools

Receive real-time alerts when any of these conditions are detected:

People Running: people running could indicate running from a problem area or someone being pursued. (MOTION)
People Entering: unauthorized access to a secured area (MOTION; DIRECTION; SIZE)
People Dwelling: exceeding a people count and time limit could indicate a problem area (MOTION; TIME)
Vehicle Speed: vehicle(s) exceeding school ground speed limits put students, staff and faculty at risk (MOTION; DIRECTION; SPEED)
Vehicle Entering: unauthorized vehicle entrance (MOTION; DIRECTION)
Item Removed: unauthorized removal of an item from a defined area (MOTION)

Video Analytics for Apartment and Office Complexes (MDU / MTU)

Receive real-time alerts or quickly search for pre-recorded events of:

People Entering: quickly view all people entering a building during a set period of time (MOTION; DIRECTION)
People Dwelling: exceeding a people count and time limit could indicate a problem area (MOTION; TIME)
Camera Tampering: detects incidents that disturb the security camera’s performance (MOTION; DIRECTION; TIME)
Vehicle Speed: vehicle(s) exceeding parking area speed limits put tenants and visitors at risk (MOTION; DIRECTION; SPEED)
Vehicle Color: quickly search through video feeds for a specific vehicle color (MOTION; DIRECTION; COLOR)

Video Analytics for Retail

Review automated reports & easily access pre-recorded events of:

People Entering: quantity and time of people entering the store (MOTION; DIRECTION)
People Exiting: quantity and time of people exiting the store (MOTION; DIRECTION)
People Dwelling: exceeding a people count and time limit could indicate interest in a particular store display (MOTION; TIME)

Object filtering allows you to fine tune your video analytics enabling a quicker response time and quicker access to video evidence. These are simple examples of the object filtering capability of video analytics. We can help you to define what object filters would work best in your environment.

Posted in: Video Analytics, Video Management Software

Leave a Comment (0) ↓