基于图像分割的尾矿坝干滩长度监测

Dry Beach of Tailings Dam Length Monitoring Based on Image Segmentation

干滩长度是反映影响尾矿库安全稳定的重要参数之一;为了测量尾矿坝干滩的长度,提出一种基于图像分割的干滩长度监测系统;根据尾矿库地形特点通过获取尾矿库水域边界图像,在Emgu CV环境下,使用OTSU阈值分割法、轮廓识别、分水岭法等算法过滤图片中尾矿库周围的植物、水面的波纹和倒影,自动清晰地识别干滩和水面的分界水线,并对水线像素坐标点进行分析与处理,实时得出最准确的干滩水线像素坐标;创新性地只通过一次标志物标定就能得出水线坐标与干滩长度的函数,从而得到尾矿坝干滩长度;经过浙江建德铜矿尾矿坝实地测量,该方法长度误差小于2.6%。

The length of dry beach is an important monitoring content that affects the safety of tailing pond. In order to measure the length of the tailings dam, proposed a dry beach length monitoring system based on image segmentation. According to the terrain features of the railings dam, the water boundary image of the railings dam is obtained by the acquisition of the water. In the Emgu CV environment, the OTSU algorithm threshold segmentation, contour recognition, and watershed algorithm can be used to retrieve the boundary between dry beaches and water surface. Analyze and deal with the image pixel coordinates, the most accurate pixel coordinates of waterline can be real-time measured. Only through one calibration can he obtained waterline coordinates and dry beach length function innovatively, so it＇s easy to measure the dry beach of tailings dam length. After field survey of tailings dam in Zhejiang Jiande copper mine, the error of this method is less than 2.6 %.