Tailings Dam Dry Beach Length Measurement System Research Based On Image Segmentation

Tailings dam,as an important element in mine production,is an important source of environmental pollution as well as hazards in mines.For this reason,tailings dam safety has become an important research subject in disaster prevention and alleviation.Dry beach length is one of the essential parameters reflecting the safety of a tailings dam.This paper proposes a dry beach length survey technique that is based on image segmentation,which focuses on: The pixel coordinates of individual dry beach water lines are found using image segmentation and the dry beach length is estimated using conversion formula.Major work includes:Firstly,a modified watershed algorithm based on HSV gradient and self-adaptive marker is proposed to address over-segmentation in the watershed algorithm.On the basis of HSV gradient graph,self-adaptive HSV gradients are computed so as to decrease the H-channel dissimilarity gradient value,The markers are extracted self-adaptively in H-minima transformation having regard to the statistical characteristics,and the watershed segmentation is made using minima marker technique.This approach solves satisfactorily the over-segmentation problem with the watershed algorithm.Secondly,an algorithm of individual dry beach water line segmentation is developed from the modified watershed algorithm.Non-water line outlines are identified and filtered appropriately,Hsafe values are set and used to redraw 2 markers,representing the dry beach and the water surface respectively,and then individual dry beach water lines are segmented using watershed transformation.Pixel coordinates are identified by means of pixel detection.This approach makes up for the lack of a method in dry beach survey sector that is capable of identifying the pixel coordinates of an entire water line on a complicated dry beach water line image.Further,2 innovative survey methods have been designed which convert water line pixel coordinates into the length of the dry beach.A discussion on their errors and their source of errors is given.The costly and risky conversion formula process of traditional video image processing methods in surveying dry beach length is thusrectified.The 2 algorithms have been tested by measuring the tin & iron tailings reservoir in Zhejiang Chun'an county,and the relative error of both algorithms is below 1%,satisfying the requirements for dry beach survey.In the end,an image segmentation and survey system terminal has been developed using Emgu CV visual processing library and the C# platform.Use instructions of this terminal are described.