Accurate and automatic detection of human presence in order toprevent unauthorized use of high value assets is an importantissue for both DOE and society in general. To date, securitysystems have been very expensive in terms of capital and verylabor intensive to operate. In addition, the systems have beenprone to false alarms, reducing confidence in the alarmsgenerated. In this project a security monitoring system based ona unique combination of video acquisition, wireless spreadspectrum communications, and morphological image processing todetect the presence of human intruders is planned. The ultimategoal is to reduce by a factor of 10 the capital costs andinstallation labor (compared to conventional security systems),reduce false alarms to near zero, produce the positive detectionof humans with near 100% accuracy, and provide fully automatedoperation. Phase I investigates morphological image processingfor human detection, implements the test bed detection system,and validates the system concept based on actual video. Thesoftware for image processing runs on a workstation or personalcomputer for the test bed. The software will be analyzed todetermine the requirements for porting to low cost dedicateddigital signal processing. Cost analysis will determine if thesignal processing should be done in hardware, in software, atremote nodes, or at central hubs. Design alternatives will besuggested for changes in the image processing software or imagefiles transferred to optimize cost and performance. The finalreport will provide test results, suggest design alternativesbased on test results, and summarize the analysis of software andhardware requirements to develop a prototype for field testing.Anticipated Results /Potential Commercial Applications as described by the awardee: The expected result of this project isthe proof of concept that human presence can be automaticallydetected and confirmed with a very low false alarm rate usingwireless data communication systems and advanced signalprocessing and can be manufactured for a l-ow cost (under $600 anode). Controlling crime is a major issue in the United States.Currently around 9.95 billion dollars per year is spent onsecurity systems. These monitored systems have excessive falsealarm rates. With the successful completion of this system,government, business, and consumers can use a new generation ofsecurity system with near zero false alarm rate and very lowinstallation and operating costs.
