Forensic Search smart scene analysis in video archives

Why Forensic Search?

The main goal of using an intelligent search engine is to find events of interest to the user in the video archive quickly if the exact time when the event occurred is unknown. Forensic Search makes it possible to raise work with the video archive to a new level of quality, from the user viewing the entire time interval which is potentially of interest to situational analysis of the archive according to set parameters.

What is Forensic Search?

With Forensic Search, camera video is processed in real time, and metadata describing all objects in the scene and their characteristics are recorded to a special database that is synchronized with the video footage. No preliminary configuration of video detection tools is needed. To search, the user enters a query (this is similar to configuration of a video detection tool), for example, crossing a line or motion in a zone. The Forensic Search system processes the recorded metadata and in seconds finds all video fragments which correspond to the query.

Search and find with Forensic Search

You can specify the following search criteria via the Forensic Search user interface:

  • crossing a virtual straight line in a user-defined direction
  • motion in the zone
  • zone entry
  • zone exit
  • motion from one area to another.
  • item appeared in the zone
  • item disappeared in the zone
  • stopping by in the zone
  • presence in the zone for over 10 seconds
  • abandoned object

In addition, you can configure Forensic Search to find objects of the selected color or filter objects that cross a line/polyline with the selected speed.

Benefits

  • There is no need for preliminary configuration of video detection tools; data about all moving objects in the scene are automatically saved. This advantage is especially valuable for large systems, as it makes it possible to quickly install and start using Forensic Search.
  • Instant search. The first query results appear on the screen almost immediately.
  • As the system works with video recorded into the archive, a search can be performed several times while you refine your query.

Axxon Intellects Video Detection Capabilities

Basic VMD

Video Motion Detection (VMD). A group of detection tools designed to detect moving objects in the frame. The basic motion detection tool detects motion in the scene as it looks for anything that is different from one video frame to another. An object tracker detects motion and its direction, and can track an object even if the camera is shaking or while the PTZ camera is rotating. The infrared detection tool (for its operation an infrared imager is needed) is triggered by motion in the infrared range. The motion direction detection tool captures motion in set directions.

The abandoned object/object disappearance detection tool notifies of the appearance of an object in the scene or the disappearance of an object from the scene. This tool makes it possible to detect, for example, the disappearance of a laptop from a desk, a briefcase left in a building lobby, or an illegally parked automobile.

Facial recognition detects the appearance of a human face in the scene, distinguishing it from any other object. Besides providing security, the face detection tool can be used for other tasks, for example, for counting the number of visitors to a shop in one day or determining peak hours and the number of visitors at various times of day.

Scene analysis

This group of detection tools makes it possible to filter set types of movement of an object in the scene. The user configures lines, polygonal zones and time intervals, and the system detects events corresponding to the set criteria. Scene video detection tools capture:

  • when an object crosses a straight line in a selected direction;
  • when an object crosses a polyline in a selected direction;
  • motion in a zone;
  • entry of an object into a zone;
  • exit of an object from a zone;
  • appearance of an object in a zone;
  • disappearance of an object from a zone;
  • stopping of an object in a zone;
  • an object remaining in a zone more than 10 seconds.
  • an object abandoned in a zone.

Any of these tools can be fine-tuned to track a specific object: a person, an automobile or all objects.

Camera Tampering

Intellects camera health detection tools trigger in case of malfunctions in a video camera’s operation. They not only make it possible to identify attempts to disable a camera, but also detect various kinds of interference which reduce the quality of event recording.

The covered lens detection tool is triggered by all instances of accidental or deliberate covering / blocking / spray-painting of the lens. This is especially important in situations when a camera is located within reach.

The camera blinding detection tool generates an alert if a beam of bright light, such as a flashlight, searchlight or automobile headlight, is directed into the lens.

The camera shift detection tool notifies of reorientation of a camera in space. This detection tool is especially needed in situations when a camera is located within reach and can easily be turned.

The background change detection tool reacts to changes in the background in front of a camera. This detection tool is very similar to the previous one, but serves somewhat different purposes. While the camera shift detection tool reacts to manipulation of the camera itself, the background change detection tool reacts to manipulations around it, for instance, an attempt to place an artificial background in front of the camera.

The defocusing detection tool notifies of loss of image sharpness as a result of defocusing or dirtying of the camera lens. This is possible, for example, if someone deliberately or accidentally disturbed the focus.

Besides its own video analytics functions, Intellect supports the embedded video detection tools of IP cameras, and also can work with specialized video analytics devices.