基于图像识别的边坡表面位移检测方法研究

Study on the Slope Surface Displacement Detecting Method Based on Image Recognition

针对边坡表面状态发生的变化检测,提出了一种新的智能检测方法。人工设置两类标识体,一类为定位标识,一类为观测标识。用图像识别方法对标识体的质心进行提取和计算,通过一定的判别准则来判断边坡的表面位移状况。具体处理方法为先将图像转换到HIS空间,根据设定的颜色提取感兴趣区域。进行图像边缘二值化处理,计算出感兴趣区域的中心坐标值。实验证明,该方法具有鲁棒性好、检测准确率高等特点。

In this paper an intelligent method based on image recognition for the detection of slope surface state was presented. Firstly two objects were set by artificial body, with one as an anchor point, and another as a monitoring point. Secondly the centroid of object was found and calculated by using the method of image recognition. Then the change of the state of the slope surface was characterized through some criteria. This method was mainly carried out in follow steps： convert RGB color space to HIS color space, and extract pixels with given color, and determine the region of interest, followed by converting the image to binary image and measuring center coordinates of the ROI. The experimental results revealed the good robustness, accurate detection rate of the presented method.