Smoke identification method and device based on rapid bilateral filtering

Abstract

The invention relates to a smoke identification method and device based on rapid bilateral filtering. One embodiment of the method comprises the steps of obtaining an original image; determining the spatial distance between any adjacent pixel point in the neighborhood of the target pixel point in the original image and the target pixel point, and the deviation degree between the pixel value of theadjacent pixel point and the pixel value of the target pixel point; obtaining the spatial weight of the adjacent pixel point according to the spatial distance, obtaining the similarity weight of theadjacent pixel point according to the deviation degree, and obtaining the comprehensive weight of the adjacent pixel point according to the spatial weight and the similarity weight; determining a weighted sum of the pixel values of each adjacent pixel point in the neighborhood of the target pixel point based on the comprehensive weight, and taking the weighted sum as a pixel value of the target pixel point in the filtering image; carrying out the smoke recognition on the filtered image. According to the embodiment, a rapid bilateral filtering method can be provided for image filtering, so thatthe smoke detection is facilitated.

Description

technical field [0001] The present invention relates to the technical field of pattern recognition, in particular to a smoke recognition method and device based on fast bilateral filtering. Background technique [0002] Fire is one of the common serious disasters of human beings. In order to reduce the damage caused by fire, the most important thing is to give early warning to fire. In the early stage of a fire, the fuel is usually not fully burned, often accompanied by smoke, so the smoke can be used as a precursor to the fire. With the development of industry, the video monitoring system based on visible light has become an important means to protect people's safe life, and its advantages of low cost, high resolution, wide range and non-contact make it widely used in fire early warning system. Currently, video-based smoke detection is an important research and application direction in fire warning. [0003] However, there is a lot of noise in the original image sequence,...

AI Technical Summary

Problems solved by technology

[0003] However, there is a lot of noise in the original image sequence, especially in the smooth area, so the image needs to be filtered, and the image filtering process not only needs to smooth the image, but also needs to maintain the edge information of the target in the image

[0004] Bilateral filtering has a large amount of calculation in the process of image processing, and the real-time performance is poor. Therefore, fast bilateral filtering is required to improve the real-time performance of filtering.

Images