Abstract: A detection method of foamed slag height based on computer vision was put forth. The image of self-made foamed slag in high temperature stove was gathered several times in laboratory，and the digital image was processed by computer. The contour line of foamed dreg was displayed in real-time on CRT monitor in the form of black and white image. According to the calibration tail of contour background，the actual height of foamed slag in stove can be observed directly. Compared with the experimental results of several kinds of traditional edge detection algorithms, it approves that this edge detection method has better immunity and adaptability.

Key words: Computer vision, Foamed dregs, Image processing, Edge detection

摘要： 提出一种基于计算机视觉的高炉泡沫渣高度检测方法。在实验室对自制的高温炉内的泡沫渣图像进行了多次采集，并由计算机对该数字图像进行处理，由CRT监示器以黑白图像的形式实时显示泡沫渣的轮廓线，并由轮廓线背景的刻度线，即可以直接看出高炉泡沫渣各处的实际高度。通过与几种传统边缘检测方法的实验结果比较，充分论证了该检测方法具有更好的抗干扰性和适应性。

关键词: 计算机视觉, 泡沫渣, 图像处理, 边缘检测