Emergency guidance system based on voice and video

摘要

The present invention detects the number of moving subjects during the period of the detection area S, and when the object quantity N is equal to or greater than a preset first set value (TH1, Threshold1), the voice recognition is not used, but only an image is used for an emergency situation. It is possible to increase the accuracy and reliability of the judgment by judging whether or not it occurs, and when the number of moving subjects during the period of the detection area S is less than the set value TH, the obtained voice data is compared with the preset comparison data to make voice. When the degree of similarity with specific comparison data is greater than or equal to a threshold, it is possible to accurately detect an emergency by determining that an emergency has occurred, and by analyzing the image acquired by the camera, a motion vector of any one of the subjects is above a critical time If it stops for a while, it is possible to immediately determine the accident by determining that an emergency has occurred, and after detecting the average moving speed of the subjects acquired by the camera, the rate of change of the average moving speed is a preset second set value (TH2 ) In case of abnormality, it is related to an emergency guidance system that can immediately determine an emergency such as a fire by determining that an emergency has occurred.