基于完整局部二值和阈值优化的火焰边缘检测  

Flame Image Edge Detection:Based on Local Binary Patterns and Thresholding Techniques

作　　者：万瑞军[1] 李小康[1,2] 沈继忱[1] 

机构地区：[1]东北电力大学自动化工程学院,吉林132012 [2]广东粤电华清煤气化联合循环发电有限公司,江门529149

出　　处：《计算机与数字工程》2015年第11期2040-2043,共4页Computer & Digital Engineering

摘　　要：针对火焰图像的边缘提取问题,论文提出了基于完整局部二值和阈值调节的边缘检测方法,该方法具有稳定,简约的优点。首先通过完整局部二值技术对火焰图像的主要边缘特征进行提取,然后利用二次阈值优化使提取的边缘更清晰化。实验结果表明在图像被噪声污染情况下,该方法仍然有效,经合成图像和现场拍摄图像分析验证,证明该方法比当前主流图像边缘提取方法具有更好的效果,为火焰图像后续分析奠定了基础。A simple and robust approach for flame image edge extraction is presented in this paper ,which is based on completed local binary patterns and thresholding techniques .Firstly ,completed local binary patterns is utilized to obtain The main features of an image ,and then the two level thresholding is used to make the edges more clear .The results show that in the case of the image is polluted by noise ,the approach is still valid ,Various experimentations are implemented on syn-thetic as well as real images .The results show the proposed approach gives good localization and effective edge detection as compared to the existing methods .

关 键 词：边缘检测 阈值技术 火焰图像 完整局部二值 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]