Simulation of Enhanced Identification of Fire Points in Indoor Fire Environment with High Concentration Smoke
Fire points have uncertainty in their morphology,especially in high concentration smoke fire environ-ments where they have random transparency and are easily mixed with the background of the image,making it difficult to identify the location of fire points.As a results,this paper puts forward an enhanced method for identifying fire points in high concentration smoke environment of indoor fire.At first,the high concentration smoke image was de-noised,sharpened and segmented,and then the contour of target region was extracted.On this basis,the target contour was extracted as well and enhanced,so that the feature vector of the fire point in target area could be obtained.Moreo-ver,TWSVM was used to classify the features of fire point,thus achieving accurate identification for fire point in the environment of high concentration smoke of indoor fire.Experimental results show that the accuracy of identifying fire point is higher than 98%,and the time consumption is less than 210ms.In addition,this method can extract target features effectively,indicating that the application effect is better and the reliability is higher.
Indoor fireHigh-concentration smoke environmentTarget enhancementFire point identificationTwin support vector machine TWSVM
NETL
NSTL
万方数据
