Post on 05-Jan-2016
description
IAIInfrared Security System and Method
US Patent 7,738,008 June 15 2010
How Does It Work?
June 2010
IAI = Infrared Applications Inc.
Test Set-up: Visual Orientation
• Two cameras with a common surveillance field of view.
• Camera B can be seen in Camera A’s FOV.
• Camera A is positioned in Camera B’s FOV
• Angles between cameras & targets are as shown
• The Cameras are 105 feet apart.
Camera A
Target
Camera B Location
Camera A Angle
Camera B
Camera A Location
Target
Camera B angle
ISS Geometry
Location Camera A
Location Camera B
Target Planter Location
Distance Between Cameras (105 feet)
R2R1
Set up value, distance between CamerasCameras & target Actual PositionsRange computations, R1 & R2
Real time calculation of Target Size
• Cameras are IR calibrated and balancedGain & Level using common objects
• Each IR camera employs convention 2 dimensional processing.
Target segmentationThreshold
• Two dimensional information is processed in real time into 3 dimensional information
Precise object location (x,y,z, coordinates)Precise physical size (sq. ft)
Threat Determination
• Targets are defined by size – (eg: truck, car, large animal, human, small
animal/child, very small animal)
• A threat is defined as a specific target in a defined location
• The Location– All or part of surveillance field – Or a specific threat area: No fly zone
Target Upgrade & Tracking
• Target was defined by actual size. • Once classified as a threat
– Actual the actual target size is stored.– Actual Inherent Thermal Contrast is stored.
• Threat is continuously tracked by:– Both cameras or one camera if either camera
becomes obscured. (2 D tracking using Actual target size for ranging and Inherent contrast for improved discrimination)
• Continuously tracking: – Allows higher order threat determinations
Intruder Example
• Series of snap shots of an intruder• The initial detection is by Camera A.• Then, Intruder enters Camera B FOV• The intruder enters the yard.• The intruder is continuously tracked
through partial and total camera obscurations.
• Snap shots are 1/40 the actual number of independent samples at video frame rates.
Camera B, initial detection
Camera A ------ Camera B
Camera A ------ Camera B Almost fully obscured
Camera A ------ Camera B
Camera A ------ Camera B
Camera A ------ Camera B
Camera A ------ Camera B Partially Obscured Mostly Obscured
Intruder
• The target alarm sounded approximately 0.5 seconds after Camera B detection.
• The highly cluttered scene caused each camera to lose the target because of complete or partially obscurations.
• The “arc” path of the intruder causes an aspect change with small changes in computed size.
• With more than 300 independent sampled image pairs, the confidence level is extremely high.
• The Intruder was observed to be carrying a tool or a weapon.
Object in Hand
Advance Discrimination Techniques
• Target Refinement: – ITC & Size of each target allows discrimination between targets in a
multi-target environment.• Target Image Dropout
– “Inherent thermal contrast” and actual size are used to re-acquire and separate new targets from old target.
• Multiple targets: – Can merge together and then separate, where ITC and physical size
assigned to each target are used to maintain the identity of each target. • Behavioral traits
– Movement over time against a preset criteria are associated with a certain kinds of threat.
• Redundant information– ITC and physical size provides redundant data that support the
application of best estimate theory.
Advance Discrimination Techniques
• Designed for multiple targets, each target having a separate threat definition, and threat response. (examples one or more)– “People-size” targets in specific areas at defined times– People congregating (crowd recognition)– Loitering (excessive time)– Stalking, (time history relationship between two targets)– Lying in wait, (serious home evasion threat)– A unattended child entering a swimming pool– Animals entering controlled areas– People exhibiting threatening behavioral – “Man down” recognition– Cars, time and location criteria– Trucks, time and location criteria– People count, matching entering with exits, tagging size– Verification, matching size with independent data, e.g. RFID data
Summary
• Field tests have demonstrated the attributes of the Infrared Security System.
• ISS provides reliable target detection and threat classification.
• High level of confidence that all false alarms have been rejected or minimized.
• ISS has the capability to be the first fully automatic physical security system
• ISS minimizes or eliminates the costly dedicated control rooms of TV monitors and security analysts.
• ISS provides the real time information needed by a first responder, or Information needed by the occupants of the home to avoid the threat.
ISS Applications• Major Business Sectors
– Home Security (adjunct to existing home security)– Factories (upgrade from forensic to threat negation)– 24/7 High Value (integrated threat assessment)
• Power Plants• Refineries• Farms (man and environmental threats, seasonal threat) • Military Installations & portable field operations• Shopping Malls, parking lot security (host of threats)• Airports: intrusion, unattended luggage, & threat tracking
• Green Applications– Automobile Sales lots– Correctional Institutions– Transportation Depots/shipyards/docks
• Other forms employing the core patented principle of 3D processing