HUMAN PRESENCE DETECTION BASED ON IMAGING SENSORS AND IMAGE PROCESSING AND ANALYSIS

EXISTING HUMAN PRESENCE DETECTION SYSTEMS SUFFER FROM HIGH FALSE ALARM RATE AND LIMITED PROGRAMMABILITY BECAUSE THEY LACK INTELLIGENT SIGNAL PROCESSING. A SYSTEM BASED ON IMAGE PROCESSING AND ANALYSIS TECHNIQUE WILL ADDRESS THESE PROBLEMS. THE IMAGE DATA IS ACQUIRED BY AN INFRARED IMAGING SENSOR AND A CHARGE-COUPLED DEVICE (CCD) CAMERA. MOTION DETECTION AND PATTERN RECOGNITION ARE EMPLOYED TO DETECT HUMAN PRESENCE. IN ADDITION, THE SYSTEM INCORPORATES AN INNOVATIVE SIGNS-OF-LIFE DETECTION TECHNIQUE, BASED ON DETECTING MICRO-MOTIONS OF BREATHING, EYE BLINKING, AND HEAD TURNING. DATA FUSION REDUCES FALSE ALARM RATE. THE PHASE I WORK WILL ESTABLISH THE FEASIBILITY OF THE PROPOSED APPROACH. ALGORITHMS FOR MOTION DETECTION AND PATTERN RECOGNITION WILL BE IMPLEMENTED IN A PERSONAL COMPUTER. THE EFFECTIVENESS OF THE SIGNS-OF-LIFE DETECTION WILL BE EVALUATED EXPERIMENTALLY USING HUMAN SUBJECTS, A CCD CAMERA, A CCD-TO-COMPUTER INTERFACE BOARD, AND A COMPUTER FOR SIGNAL PROCESSING. PHASE I WILL INCLUDE AN ESTIMATE OF THE SYSTEM'S DETECTION AND FALSE ALARM RATES. IN PHASE II, AN OPERATIONAL HUMAN PRESENCE DETECTION SYSTEM WILL BE CONSTRUCTED.