NCJ Number
233345
Date Published
January 2011
Length
38 pages
Annotation
This project's objective was to develop a prototype intelligent in-car video (ICV) system with higher image quality and machine-based video analysis of the incoming video stream in order to detect critical events during a routine police traffic stop.
Abstract
The detection of the critical events would result in certain automated decisionmaking, the simplest of which is to alert the police headquarters and call for back-up assistance. Other automated decisions upon detection of critical events would include recording the scene with higher resolution video camera. The video analyses could also be linked to data from GPS (global positioning system) and ALPR (automated license plate recording) devices in order to supply supporting information, such as location of the incident and the identity of the car stopped. The project succeeded in building a computer system that analyzes video and identifies certain critical events. The system was developed as a software prototype on a laptop computer with a digital video recording device connected to the laptop. The laptop could be mounted in the passenger front seat of a patrol car for data collection and testing purposes. The project also deployed the test system in police cars and in simulations, so as to capture realistic video recordings of the intended subject matter. Data were collected from actual police car videos and from simulated situations, which were then used to test the developed algorithms. In addition, the system's ability was tested to determine critical situations. The situations included an open door of the stopped car, a person running out of the stopped car, and officer falling down. Algorithms were successfully developed and tested for detecting these events. These achievements have been presented in scientific venues, and the commercialization of the prototype system is open for interested parties. 24 figures, 3 tables, and 18 references
Date Published: January 1, 2011