Intelligent image smog and flame detector and flame detection method

Abstract

The invention relates to an intelligent image smog and flame detector, a compound type intelligent image smog and flame detector and an intelligent image smog and flame detecting system. The detectorcomprises a camera or an imaging CCD component, a DSP core processing circuit, an input and output circuit, an infrared lamp and filter component, a shell packaged by metal or other materials, a window lens, and front and back covers, and is characterized by reliably monitoring fires happening in the early stage such as smog and flare, thereby realizing functions of alarming early and reducing theoccurrence of disaster.

Description

technical field [0001] The invention designs fire detection or alarm technology, in particular relates to an intelligent image smoke and fire detection digital information system. Background technique [0002] With the continuous development of global pattern recognition technology, and the visual demand in the development of fire protection and security systems in the past century, especially the increased demand for security monitoring after 911 in the United States, intelligent detection systems based on video images emerge in endlessly. Due to the different monitoring objectives, it is divided into different research and application directions. The invention relates to the field of fire detection and alarm. [0003] Over the years, large spaces (convention centers, stadiums, hangars, large warehouses, large factories, etc.), special buildings (traffic tunnels, petroleum and petrochemical production areas, explosive storage areas, museums, subways, train stations, hospit...

AI Technical Summary

Problems solved by technology

[0004] The characteristics of large-scale spaces are: large space, generally as high as 20m or more, and large floor area of ​​a single building; it is difficult to detect fires in the early stage. , smoke concentration, temperature, etc. are difficult to respond quickly. For example, it takes a long time for the smoke to reach a certain height and generate the smoke concentration that makes the through-ray smoke detector act as an alarm; location where the fire extinguishing system cannot be controlled to effectively extinguish the fire

[0005] The characteristics of special buildings will be more complicated, and their common features are: there are some defects in conventional detection technology, it is difficult to achieve the purpose of rapid detection; the environmental conditions are complex, and in some cases the environment is very poor, facing special environments such as explosive Problem; requires rapid detection, early forecast, and integration of fire safety monitoring

[0006] The characteristics of large outdoor venues are: outdoor, large area, environment changes with natural weather, and complex conditions; fire location is uncertain and random, and there is currently no suitable detection method; external environment interference is large, light, fog, etc. , Moisture, rain, snow, lightning, etc. are greatly affected; it is necessary to form a large-area and wide-area monitoring

[0007] Because of these particularities, the current conventional detection technology is difficult to achieve the expected purpose, especially the purpose of early detection of fire, early control and treatment

[0008] Chinese patent 95100519.7 proposes a method for detecting and locating fires using the difference of three primary colors of color images. The problems of this method are: first, the characteristics of three primary colors of many artificial or natural light sources are similar to those of flames. The primary color difference method has a very high false alarm rate; second, this algorithm is only for flame detection and does not have the ability to detect smoke, so it does not have the ability to detect and alarm early in large spaces and special buildings

Chinese Patent 00805204.2 (International Publication No. WO00 / 63863, Publication Date 2000.10.26) proposes a smoke detection method with an optical cross-sectional image. The existing problem is that the principle is basically similar to that of a linear beam smoke detector and is not suitable for In places with high heights, it is difficult to detect the rising smoke generated at the initial stage of the fire, and the installation of infrared light-emitting arrays is very inconvenient

[0009] Foreign patent WO01 / 57819, US7002478 "Smoke and Flame Detection" proposed some fire detection algorithms, the main problems of which are: first, this method uses complex algorithms such as correlation analysis and Bayesian statistical methods, and uses a large number of Operations such as nth power and square root etc., plus the number of image points in one frame is more than 104 points, so the amount of calculation is very large, it is impossible to transplant and use it in the DSP processor, only high-performance computer analysis can be used; second, the system does not have The control mechanism of the camera and supporting light source is difficult to use at night, which is the biggest obstacle to its use; third, because it needs to use an industrial computer including a keyboard, it is difficult to pass the international standard on fire detectors This is the reason why it has not been able to obtain international certification so far

Images