Video-based flame detecting device and method

Abstract

The invention discloses a video-based flame detecting device and a video-based flame detecting method. The method comprises the following steps of: firstly, acquiring an edge shape of a foreground area by using a photographing unit, a foreground extraction unit and a foreground edge extraction unit; and secondly, acquiring the number of sharp angle shapes by the treatment of a sharp angle detection unit and a sharp angle counting unit of the current frame, calculating by using a continuous multi-frame sharp angle number storage unit and then making a further judgment on the calculating result by using a flame judgment unit; or directly making an judgment on the number of the sharp angle shapes by the flame judgment unit according to the counting of the sharp angle shape counting unit, and thus judging whether a flame exists or not. The device and the method identify and detect the sharp angle features of the flame by video image identification algorithm respectively, judge the identification and detection results, and if flame standards are met, can give an alarm or further perform treatment. The detecting device and the detecting method have the advantages of effectively eliminating interference of external factors such as light illumination and shallow, along with high identification accuracy, low false alarm rate and low omitted alarm rate.

Description

technical field [0001] The invention relates to a video-based flame detection device and method. Background technique [0002] In modern society, more and more electrical appliances are used in various occasions, which increases safety hazards. People's awareness of prevention has gradually increased, and they have begun to pay attention to various security products. At present, there are related monitoring and security products sold in the market, which can be basically divided into two categories in terms of monitoring means: sensor monitoring and video monitoring. Sensor monitoring is to use smoke and fire sensors to detect whether there is smoke or fire and give an alarm. This monitoring method is easily disturbed by external factors (such as: temperature, distance, vibration, sound and other external factors), and the false alarm rate is high. The sensor-type smoke detection system reacts slowly and requires a certain concentration and temperature. When it is found, th...

AI Technical Summary

Problems solved by technology

Sensor monitoring is to use smoke and fire sensors to detect whether there is smoke or fire and give an alarm. This monitoring method is easily disturbed by external factors (such as: temperature, distance, vibration, sound and other external factors), and the false alarm rate is high. The smoke sensor system of the sensor type reacts slowly and requires a certain concentration and temperature. When it is found, the loss has already been caused.

Video surveillance is to use the camera to collect live video in real time, and the on-duty personnel will guard it at all times. When there is smoke or fire on the scene, it will be dealt with in a timely manner. Although the false alarm rate is low, due to the need for manual guarding, it is easy to produce visual acuity after a long time. fatigue, lead to omission

Images