Steve Gorski is general manager for the Americas at Mobotix, New York City.
The security industry is plagued by false dispatches, which wastes law enforcement and internal resources and cost taxpayers more than $2 billion every year. With the inception of video analytics, the false alarm issue, reflected by the false positive rate, became a bigger problem for those using advanced intelligence to monitor perimeters, stores and businesses.
Massive false alarms make the real-world implementation of motion detection a challenge, especially in outdoor and crowded environments with constant activity. Finding the real events in traditional motion detection software is like finding a needle in a haystack — every movement triggers an alarm.
But the industry has taken notice and as a result continues to work diligently to improve analytic sensors and algorithms. Today, there are solutions that look closely at video footage to recognize the traveling movement of people and objects while dismissing extraneous and repetitive changes and movement in the scene. This drastically reduces the number of false alarms in video motion detection applications.
Selective activity sensors
New technology advancements enable activity-controlled, software-based image analysis to detect movement in a monitored area. Unlike existing video motion sensors, which detect all events in the pre-defined area, these activity sensors deliver reliable motion detection, even in applications with large amounts of interference.
Using the activity sensor, a camera can distinguish between movements of vehicles, people or objects that should trigger an alarm and movements that are not relevant, such as shadows, changes in light, swaying trees, animals or rain. The sensor is capable of detecting objects or people that cross the pre-defined monitored area regardless of background movement.
These sensors also reduce storage requirements, another plus. A side-by-side comparison with traditional motion-detection cameras and cameras equipped with a sensor to improve analytic reporting demonstrated amazing results. During a 24-hour test at a busy, hillside environment, the traditional motion-detection camera detected 8,181 events that took up 9,341 megabytes of storage. In the same scene using the camera with the sensor, 138 detected events took up only 157 megabytes of storage.
These new developments in software and motion detection can reduce false alarms by more than 90 percent. Augmenting the benefits of the sensor, a decentralized approach to surveillance enables the camera to analyze an event before alarm action occurs. With the decentralized concept, a high-speed computer and if necessary, digital long-term memory (MicroSD Card) is built into every camera. The PC and the video control center serve only for viewing and controlling the cameras (PTZ), not for analysis or recording. This makes it unnecessary to purchase expensive video management software, as computer-intensive functions are integrated in the cameras.
Mobotix recently introduced a technology called MxActivitySensor. After reviewing and testing the software, we believe that in 90 percent of cases MxActivitySensor will replace Video Motion Detection (VMD), which delivers greater accuracy than traditional motion based alarm triggering and is also a lot easier to setup.
Better analytics and fewer false alarms can increase the adoption of video motion detection in standard commercial applications and critical infrastructure at remote sites. Systems integrators benefit from offering technology that limits false alarms from both an industry, customer and revenue perspective, while end users can reap potential cost savings with less live monitoring.
Steve Gorski is general manager for the Americas at Mobotix, New York City. Request more info on the company at www.securityinfowatch.com/10239577.