Free Newsletter
Register for our Free Newsletters
Access Control
Deutsche Zone (German Zone)
Education, Training and Professional Services
Government Programmes
Guarding, Equipment and Enforcement
Industrial Computing Security
IT Security
Physical Security
View All
Other Carouselweb publications
Carousel Web
Defense File
New Materials
Pro Health Zone
Pro Manufacturing Zone
Pro Security Zone
Web Lec
ProSecurityZone Sponsor
ProSecurityZone Sponsor
ProSecurityZone Sponsor
ProSecurityZone Sponsor
ProSecurityZone Sponsor
ProSecurityZone Sponsor
Editor's Blog and Industry Comments

Smart Outdoor Surveillance Tracks Targets

25 August, 2010
SightLogix provides algorithmic error elimination for smarter edge analytics for intruder detection in challenging outdoor surveillance applications
The President and CEO of SightLogix, John Romanowich took the time recently to talk about the company's automated outdoor surveillance system and the benefits that it brings to pipeline and railway protection.

SightLogix' innovative automated outdoor surveillance system uses a combination of a range of cameras, on-board analytics, geometric positioning and peer-to-peer camera control to detect and track intruders in harsh outdoor environments without suffering from the effects of wind, rain and camera shake.

During the rise of video analytics as an emerging technology a few years ago and the market flooded with unreliable products generating high false alarm levels, the challenge for SightLogix to be convincing in the market was not an easy brief but one which the company rose too by not being over-complicated in what it was trying to achieve. The detection of a real intruder on the scene without the system being duped by blowing debris, swaying trees or camera instability was the goal.

Simple, clear and reliable analytics is essential to providing comprehensive surveillance of such remote installations as pipelines and railways since constant monitoring of 100's or 1000's of cameras is not feasible. This analysis has to take place on the camera to avoid constant streams over the IP network and it has to eliminate errors through algorithms rather than sensitivity adjustments so that the system operates autonomously.

There are several components to the SightLogix system:

The Cameras are available as a wide area day/night unit or long range day/night or thermal imaging variants. All the cameras are available with rugged nitrogen-purged housings to suit harsh outdoor environments.

Edge analytics are incorporated into each camera model including electronic stabilisation, image processing, intrusion detection and algorithmic correction for weather and vibration effects as well as anomalous events such as flying paper and debris.

Target mapping provides accurate positioning information of the intruder. This is achieved through geometric calibration of fixed objects within the camera's field of view at the time of installation. Once this has been done, spatial positioning including distance from the camera can be calculated.

Edge-to-edge camera control provides instructions to PTZ units to zoom in on suspect intruders and track them through the scene. Once an intruder has been detected, all images and meta-data from the cameras is streamed to the video management system in the control centre.

I asked John why the company had chosen to go for algorithm based error correction. "Sensitivity adjustment is a recipe for disaster", he told me. "There are too many parameters to consider for sensitivity adjustments to be accurate on each camera. By building it into the image processing algorithm, we can account for such variable factors as the camera swaying on the pole, vibration from train movements or weather effects on the field of view". With long range detection capabilities affecting only a few pixels of the scene, the accuracy of such error correction is essential in eliminating high levels of false alarms.

What does the system actually detect? "We try not to be too complex in what is identified. After eliminating spurious activity on the scene, we can detect suspicious activity. The range of the camera enables this detection at considerable distance and at that range, we don't pretend to be able to detect the difference between a bear and a man in a bear suit. What we can do is to identify the area of activity and through geometric positioning focus on that activity using a PTZ camera and present this to the operator for further analysis while the potential intruder is tracked. Some systems try to do too much and this is where the errors start increasing".

I was interested in the cost implications of deploying such technology and John explained the benefits the system has. By covering long ranges and wide areas, the infrastructure costs are reduced. Control room staffing and server storage costs are also kept to a minimum through the use of on-board analytics and autonomous camera control. In this way, only events which are of interest are streamed for storage and analysis.

For the remote distances involved in railway and pipeline surveillance applications, the SightLogix Automated Outdoor Surveillance System fulfils the need for low camera density, local analytics and reliable event detection with low false alarm rates.
Bookmark and Share