基于YCbCr和帧间差分法的火焰分割算法

Flame segmentation algorithm based on YCbCr and inter frame difference methodtong

为了提高森林火灾的火焰检测效率和准确率,提出一种融合YCbCr和三帧差分法森林火焰前景提取算法。首先通过三帧差分算法对视频进行分帧处理,利用形态学进行去噪处理,再计算两帧参考帧和当前帧之间像素点的位置差异,得到像素点在参考帧中的位置和在当前帧中的位置之间的偏移量,从而分割出不断变化的火焰,结合YCbCr算法来进一步分割火焰疑似区域,达到精细化分割的目的。实验结果表明:所提方法克服了单个算法对火焰分割精度不高的缺点,可以较好地完成火焰分割,达到预期要求。

In order to improve the efficiency and accuracy of flame detection in forest fires,a fusion algorithm of YCbCr and three frame difference method for forest flame foreground extraction is proposed.Firstly,the video is segmented using a three frame difference algorithm,followed by morphological denoising.Then,the position difference between the two reference frames and the current frame is calculated to obtain the offset between the pixel positions in the reference frame and the current frame.The constantly changing flames are segmented,and the YCbCr algorithm is combined to further segment the suspected flame areas,achieving the goal of fine segmentation.The experimental results show that the proposed method overcomes the disadvantage of low accuracy in flame segmentation by a single algorithm,and can effectively complete flame segmentation,meeting the expected requirements.