QQ群聊

• CN 62-1112/TF
• ISSN 1005-2518
• 创刊于1988年

## Research on Automatic Extraction Method of Top and Bottom Line of Open-Pit Slope Based on DEM

SHI Tiandong,, BI Lin,

School of Resources and Safety Engineering，Center South University，Changsha 410083，Hunan，China

 基金资助: 国家自然科学基金项目“基于深度学习和距离场的复杂金属矿体三维建模技术研究”.  41572317

Received: 2020-12-26   Revised: 2021-03-08   Online: 2021-10-08

Abstract

The slope top and bottom line plays an important role in open-pit mine mining design， production planning，calculation of stripping volume，road network planning，etc.Traditional extraction methods require the use of total stations and other equipment to rely on manual extraction，which is inefficient.In order to solve the problem of low efficiency of the traditional method of manually extracting the slope top and bottom line from the open-pit mine 3D model，a method of automatically extracting the slope top and bottom line from the open-pit mine DEM （Digital Elevation Model） 3D model was studied.The first step of the method is to pre-process the elevation data with up-sampling and bilateral filtering.The second step is to set a threshold，then according to the different elevation values of different regions in the DEM 3D model，different regions are converted into grayscale images with different pixel values.The last step is to perform gradient operation on the gray image，extract the edge pixels in the image，combine with an improved Zhang-Suen algorithm to extract the skeleton of the edge lines，and design a search algorithm based on eight neighborhoods to order the edge pixels and connect them into a complete contour line，and use the Douglas-Peucker algorithm for thinning to get the final step line.The experiment used the measured three-dimensional model data of an open-pit mine for testing.With the aid of a three-dimensional visualization software platform，the extracted step line is superimposed with the three-dimensional model of the pit，visually analyzed and compared，and the accuracy of the extracted results is judged.The visualization results show that compared with the traditional manual extraction method，the automation method greatly improves the work efficiency under the condition of ensuring the extraction accuracy，and most importantly realizes the automation of the extraction process.This method is also applicable to the terrain features of the DEM 3D model line extraction.

Keywords： digital mine ; extraction of top and bottom line of open pit slope ; digital elevation model ; edge detection ; terrain feature extraction ; eight neighborhood edge tracking

SHI Tiandong, BI Lin. Research on Automatic Extraction Method of Top and Bottom Line of Open-Pit Slope Based on DEM[J]. Gold Science and Technology, 2021, 29(4): 612-619 doi:10.11872/j.issn.1005-2518.2021.04.006

### 图1

Fig.1   Flow chart of automatic extraction process of slope top and bottom line based on DEM

### 1.1 原始数据预处理

（1）上采样加密数据。在提取处理之前，需要对原始数据进行预处理。由于通过图片处理库读取的DEM三维模型尺寸较小，会影响到提取精度，因此需要用插值法对其进行上采样处理，扩大模型尺度（Fuentes et al.，2009钟宝江等，2016）。

（2）模型数据深拷贝备份。对原数据进行上采样处理之后，由于最后需要输出的是数据的空间坐标，因此在将DEM模型转化为灰度图像之前，需要对模型数据进行深拷贝备份，最终在灰度图像中确定坡面顶底线所处的行列号，根据行列号来计算对应的备份模型数据的空间坐标，实现将灰度图像中的边缘像素行列坐标转换为模型数据的空间坐标。

（3）滤波处理。原始的DEM三维模型存在干扰噪点，会影响到后续根据高程标准差阈值对DEM三维模型进行灰度划分的过程，因此需要对其进行滤波处理。在滤波过程中，为尽可能减少图像边缘信息的损失，以便不影响后续的边缘检测过程，使用双边滤波对原始模型进行滤波处理（马先明等，2017李广金等，2019）。

### 图2

Fig.2   Image segmentation processing

### 图3

Fig.3   Edge detection graph

### 1.3 后续处理

（1）针对第一个问题，采用一种Zhang-Suen细化算法解决（Zhang et al.，2009Chen et al.，2012）。原始的Zhang-Suen算法在本问题中存在断线和线宽不正确的问题，本文采用了一种修改的算法，将原算法的八邻域点二进制编码修改为十进制编码，作为扫描点的限制条件。算法流程如下：

①2≤NP1）≤6；

SP1）=1||BP1）∈｛65，5，20，80，13，22，52，133，141，54｝；

P2×P4×P6=0；

P4×P6×P8=0。

（2）针对第2个问题，设计了一种基于中心点八邻域范围的搜索算法，避免了逐行逐列搜索，减少用时。该算法将待处理的点划分为三类，即中间点、端点和交叉点，如图4所示。创建TypeMatrix类型矩阵来表示点的类型，该矩阵存储4个整形数值，即0、1、2、3。其中，0表示该点类型为背景点，不需要处理；1表示该点类型为中间点，2表示该点类型为端点，3表示该点类型为交叉点。

### 图4

Fig.4   Schematic diagram of three types of points

### 图5

Fig.5   Flow chart of connection algorithm

（3）针对第3个问题，采取Douglas-Peuker算法（Yu et al.，2013）解决。该算法的思想是设置一个距离阈值，当点到线的距离小于该阈值时，该点视为可以删除的点，通过不断地迭代，最后剩下的点即为最终结果。算法流程如图6所示。

### 图6

Fig.6   Flow chart of Douglas-Peucker algorithm

### 1.4 输出坐标

$x=Xcord+j×Dx$
$y=Ycord+i×Dy$

### 图7

Fig.7   Processing and result of slope top and bottom line

### 图8

Fig.8   Comparison of slope top and bottom line extracted by automatic method and super position of 3D model

## 3 结论

（1）针对传统人工勾勒提取坡顶底线的低效率问题，提出了一种自动化提取方法。通过将三维DEM模型数据转化为二维图像数据，在转化后的图像数据基础上，采用边缘检测方法提取出边缘像素，并进行必要的后续处理，从而实现自动化提取，提高了提取效率。

（2）本文方法基于Windows操作系统的C++程序实现，未来可将程序嵌入至数字采矿软件中，作为其中的某一功能模块。该方法可推广应用至野外环境中地形特征线的提取，通过将三维模型根据高程值标准差的差异转化为二维图像，结合图像处理的方法，提取所需要的地形特征。

（3）本文方法在设定标准差阈值方面尚未实现自适应确定最佳阈值，仍需人为设置，后续将继续完善研究，尝试利用最大类间方差法并结合概率统计方法实现自适应确定最佳阈值，进一步提高坡面顶底线自动提取的精细化和自动化。

http://www.goldsci.ac.cn/article/2021/1005-2518/1005-2518-2021-29-4-612.shtml

## 参考文献 原文顺序 文献年度倒序 文中引用次数倒序 被引期刊影响因子

Chen HuaweiYuan XiaocuiWu Lushenal et2019.

Automatic point cloud feature-line extraction algorithm based on curvature-mutation analysis

［J］.Optics and Precision Engineering，275）：1218-1228.

Chen WSui L CXu Z Cal et2012.

Improved Zhang-Suen thinning algorithm in binary line drawing applications

［C］//International Conference on Systems and Informatics（ICSAI2012）.YantaiIEEE1947-1950.

Cheng X LCheng X JLi Qal et2018.

Automatic registration of terrestrial and airborne point clouds using building outline features

［J］.IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing，112）：628-638.

Fuentes I O HBravo-Zanoguera M EYanez G G2009.

FPGA implementation of the bilinear interpolation algorithm for image demosaicking

［C］//International Conference on Ele-ctrical，Communications，and Computers.CholulaPueblaIEEE25-28.

Kang XLi JFan X2018.

Line feature extraction from RGB laser point cloud

［C］//International Congress on Image and Signal Processing，BioMedical Engineering and Informatics （CISP-BMEI）.BeijingIEEE1-5.

Kang Z ZZhong R FWu Aal et2016.

An Efficient planar feature fitting method using point cloud simplification and threshold-independent baysac

［J］.IEEE Geoscience and Remote Sensing Letters，1312）：1842-1846.

Kim HLiu BMyung H2017.

Road-feature extraction using point cloud and 3D LiDAR sensor for vehicle localization

［C］//International Conference on Ubiquitous Robots and Ambient Intelligence （URAI）.JejuIEEE891-892.

Li GuangjinYang Tao2019.

A guided point cloud filtering algorithm combined with Gaussian statistics

［J］.Manufacturing Automation，414）：80-84.

Liu LijunLiu JipingZhou Yuanyingal et2003.

A study of automatic vecror system about topographic map based on mathematical morphology

［J］.Application Research of Co-mputers，（2）：83-85.

Ma XianmingLi YongshuXie Jiali2017.

Experiment and analysis of point cloud denoising using bilateral filtering method

［J］.Bulletin of Surveying and Mapping，（2）：87-89.

Sun ChaoTan Zhenghua2019.

Optimal selection of the final boundary of open-pit mine under multi-factor conditions

［J］.Nonferrous Metals Engineering，95）：72-80.

Wang QingGu XiaoweiXu Xiaochuanal et2014.

Holistic optimization of production planning elements and its application for open-pit mine

［J］.Journal of Northeastern University（Nature Science），3512）：1796-1800.

Wang S MChen Y MLiu Y Fal et2015.

The extraction of feature points from DEM geographic data in cloud computing environment

［C］//International Conference on Geoinforma-tics.WuhanIEEE1-5.

Wu QunyongWu ZufeiZhang Liangpan2019.

GPS trajectory agglomeration and refined road network extraction

［J］.Acta Geodaetica et Cartographica Sinica，484）：502-511.

Yu JChen GZhang Xal et2013.

An improved Douglas-Peucker algorithm aimed at simplifying natural shoreline into direction-line

［C］//21st International Conference on Geoinformatics.KaifengIEEE1-5.

Yuan RuidongYang XinfengWang Leial et2018.

Application of automatic generation method of open-pit characteristic line based on point cloud data

［J］.Modern Mining，3410）：168-170.

Zhang Y YJing X JWang X Y2009.

An approach to improve the ZS thinning algorithm

［C］// International Conference on Network Infrastructure and Digital Content.BeijingIEEE750-753.

Zhong BaojiangLu ZhifangJi Jiahuan2016.

Review on image interpolation techniques

［J］.Journal of Data Acquisition and Processing，316）：1083-1096.

Zhu HaibinWang YanLi Yamei2018.

Surveying and mapping of open-pit mining area using unmanned aerial vehicle

［J］.Coal Engineering，5010）：162-166.

Zhu QingZhao JieZhong Zhengal et2004.

The extraction of topographic patterns based on regular grid DEMs

［J］.Acta Geodaetica et Cartographica Sinica，（1）：77-82.

［J］.光学精密工程，275）：1218-1228.

［J］.制造业自动化，414）：80-84.

［J］.计算机应用研究，（2）：83-85.

［J］.测绘通报，（2）：87-89.

［J］.有色金属工程，95）：72-80.

［J］.东北大学学报（自然科学版），3512）：1796-1800.

［J］.测绘学报，484）：502-511.

［J］.现代矿业，3410）：168-170.

［J］.数据采集与处理，316）：1083-1096.

［J］.煤炭工程，5010）：162-166.

［J］.测绘学报，（1）：77-82.

/

 〈 〉