METHOD FOR OBJECT INFERENCE USING MULTIPLE IMAGES AND A SYSTEM THEREOF, CAPABLE OF RECOGNIZING AND MONITORING CRIMES AND EVENTS/ACCIDENTS IN OBJECT INFERENCE DATA FOR PERFORMING PROTECTION OF VICTIMS AND EARLY APPREHENSION OF CRIMINALS

摘要

PURPOSE: A method for object inference using multiple images and a system thereof are provided to extract object inference data from image data at a smart node and to execute an algorithm on the extracted data for detecting crimes and events/accidents, thereby recognizing and observing crimes and events/accidents.;CONSTITUTION: A plurality of image data processors(201-203) processes image data, which is collected by a plurality of image collection cameras(101-103). A correlation extractor(204) calculates correlativity between images by each location. An object inference unit(205) performs object inference using the extracted correlativity data. An algorithm execution unit(206) applies an algorithm to crimes and events/accidents using object inference data. An event/accidents recognition unit(207) recognizes the crimes and events/accidents.;COPYRIGHT KIPO 2013;[Reference numerals] (101,102,103) Image collection cameras; (201,202,203) Image data processors; (204) Correlation extractor; (205) Object inference unit; (206) Algorithm execution unit; (207) Event/accidents recognition unit; (208) Communication processing unit; (300) Communication network; (410) Communication processing unit; (420) Situation management server; (430) Situation recognition processing unit; (440) Situation monitor; (450) Alarming light and bell