Research And Implementation Of Farmland Fire Smoke Detection Technology Based On Video Image

With the continuous development of modern agriculture and industrial technology,large areas of farmland and farmland demonstration area,farmland and vegetable area,such as more and more,all kinds of fields widely used flammable materials and easy to cause fire,also often appear straw burning phenomenon,therefore,based on video image of the fire smoke detection and prevention methods of technology rapid development is imperative.This paper adopts the technology of the video image on analysis,studied the application of higher efficiency and shorter detection time farmland video images of fire smoke detection algorithm,is presented based on video image of farmland fire smoke detection system realization method.The research content mainly includes the following several aspects:1,Introduced the video image of motion detection target or area of the classic method,for the acquisition of the smoke motion video,from the color image into grayscale image,and on the basis of describing the current use of a wide range of frame differential method,optical flow method and background subtraction between division(BDM),and other classic goal or regional motion detection algorithm.2,In view of the fire smoke visual characteristics,and through the fire smoke image color,motion,analyzes regional outline and frequency characteristics,can draw a series of efficient representation of fire smoke class characteristic parameters,finally an accurate fire smoke image construct the key characteristic parameters.3 More features,this paper discusses the irrigation and fire smoke detection technology,in view of the above analysis of smoke image color,motion,area,and the frequency characteristic information,the design of fire smoke detection system based on video image process,overall architecture and algorithm,fire smoke detection method are given.4,In order to explore the application of research results to specific fields of fire prevention work,through the related algorithm can form embedded products tool software is optimized,which according to farmland for real-time monitoring of video image to realize detection of fire smoke image.