自适应Retinex与Canny的炉膛火焰边缘检测算法

摘要

对炉膛火焰图像的增强与检测是获得炉膛内部燃烧情况的重要手段。首先，提出了一种自适应多尺度 Retinex算法，提高了火焰清晰度和炉膛火焰与炉膛的对比度。其次，提出了识别火焰边缘改进的 Canny算法，提高了火焰边缘清晰度。仿真结果表明，炉膛火焰与环境对比度得到提高，改善了图像的视觉效果，火焰边界连续且轮廓清晰。%Analysis of furnace flame is an important method to obtain the combustion condition inside the furnace.Accordingly.an adaptive multi-scale Retinex (MSR)algorithm is proposed.which effectively im-proves the clarity and detail of the fire flame and enhances the contrast of fire flame with the back-ground.In order to get a better monitoring result,improved Canny algorithm is adopted to identify the flame edge,and improves the ability of extracting the flame edge details and suppressing the false edge. After image enhancement and edge detection,it can more accurately grasp the internal combustion situa-tion of furnace.The simulation experiment shows that the brightness of the flame is improved effective-ly.the contrast of flame with furnace is obvious.and the edge of the flame is continuous and clear.