城市道路的施工进度关系到城市道路建设的建设周期、成本控制以及质量目标的实现。传统施工进度监测主要依靠监测员进行实地踏勘，但若在工程体量大、施工周期长、施工人员众多等复杂情况下，传统监测方法可能存在施工时效性低、项目经济成本提高等缺陷。因此，探究一种高时效性、高可靠性的城市道路施工进度监测方法，对于城市道路建设的精细化管理具有重要意义。 本文以西安市浐灞区广运潭大道路段的施工道路为研究对象，尝试使用面向对象的遥感分类方法提取施工不同阶段无人机获取的影像中城市施工道路，通过对三种不同分类方法进行实例验证，选择一种最优的城市施工道路提取方法，分别提取三期影像中的施工道路，并根据提取的结果对三个时期城市道路施工进度进行分析研究，以达到施工监测的目的。主要研究内容和结论如下： （1）对城市施工道路最佳分割尺度展开研究。通过探究不同同质性组合参数下多尺度分割结果的差异性，并分析影像中施工道路的分割效果，确定出无人机影像城市施工道路的最佳同质性组合参数；在确定城市施工道路最优分割尺度时，对三期影像进行迭代分割，确定出三期影像的最优分割尺度参数。结果表明：第一期影像最佳分割参数组合为：形状因子为0.3，紧致度因子为0.3，最佳分割尺度为74；第二期影像最佳分割参数组合为：形状因子为0.3，紧致度因子为0.5，分割尺度为98；第三期影像最佳分割参数组合为：形状因子为0.3，紧致度因子为0.5，最佳分割尺度为72。 （2）城市施工道路特征空间构建及优化研究中，通过对三期无人机影像中城市施工道路特征进行分析，结合光谱、纹理、指数、几何等特征，构建出35类特征集合。分别计算三期无人机影像中各类样本在此特征空间的分离度，比较不同特征空间产生的分离度，最终确定出三期影像的最优特征组合。结果表明：通过对构建特征空间的优化，三期影像中，用于城市施工道路提取的最优特征空间维数分别为20、22、29维。 （3）对城市道路施工进度监测方法进行研究。采用决策树、随机森林、贝叶斯三种分类方法对研究区城市施工道路进行提取，在三期影像中生成随机点并目视解译其类型，利用混淆矩阵对分类结果进行精度验证，结果表明：四种精度评价指标对比下随机森林法的分类精度最高，三期影像中采用随机森林法的城市施工道路提取的生产者精度和用户精度分别为：第一期91.10%，76.40%；第二期87.75%，89.26%；第三期92.70%，85.04%。故面向对象的随机森林法对城市施工道路提取的精度最佳，能够满足高分辨率遥感影像城市施工道路的提取，采用该方法能够快速实现对城市道路的施工进度监测。

The construction progress of urban roads is related to the construction period, cost control and the realization of quality objective of urban road construction. The traditional construction progress monitoring mainly relies on monitors to carry out field reconnaissance. However, under complex situation such as huge project volume, long construction period, and numerous construction personnel, the traditional monitoring method may have the defects of low construction time efficiency and high project economic cost. Therefore, it is of great significance to explore a monitoring method of urban road construction progress with high efficiency and high reliability for the fine management of urban road construction projects. In this paper, the construction roads of Guangyuntan Avenue in Chanba District of Xi'an City are taken as the research object. The object-oriented remote sensing classification method is used to extract the urban construction roads in the images obtained by UAV at different construction stages. Through the verification of three different classification methods, an optimal extraction method for urban construction roads is selected to extract the construction roads in the three phase images respectively. According to the extraction results, the construction progress of urban roads in the three periods is analyzed and studied to achieve the purpose of construction monitoring. The main research contents and conclusions are as follows : (1)The optimal segmentation scale of urban construction roads is studied. By exploring the difference of multi-scale segmentation results under different homogeneity combination parameters and analyzing the segmentation effect of construction roads in images, the optimal homogeneity combination parameters of urban construction roads in UAV images are determined. When determining the optimal segmentation scale of urban construction roads, the iterative segmentation of the three phase images is carried out to determine the the optimal segmentation scale parameters of the three phase images. The results show that the optimal segmentation parameter combination for the first-phase image is: 0.3 for shape factor, 0.3 for compactness factor and 74 for the optimal segmentation scale; the optimal segmentation parameter combination for the second-phase image is: 0.3 for shape factor, 0.5 for compactness factor and 98 for segmentation scale; the optimal segmentation parameter combination for the third-phase image is: 0.3 for shape factor, 0.5 for compactness factor and 72 for the optimal segmentation scale. (2)The construction and optimization of the feature space of urban construction roads are studied. the feature of urban construction roads in the three phase of UAV images is analyzed, 35 types of feature sets are constructed by combining spectrum, texture, index, geometry and other features. The degree of separation of various types of samples in this feature space in the three phase of UAV images is calculated separately, and the the degree of separation generated by different feature spaces are compared. Finally, the optimal feature combination of the three phase images is determined. The results show that through the optimization of the constructed feature space, the optimal feature space dimensions for the extraction of urban construction roads in the three images are 20, 22 and 29 dimensions respectively. (3)The monitoring method for the construction progress of urban roads is studied. The classification methods of Decision Tree, Random Forest and Bayes are used to extract the urban construction roads in the study area. Random points are generated in the three phase images, and their types can be visually interpreted. The accuracy of classification results is verified by using confusion matrix. The results show that the classification accuracy of the random forest method is the highest under the comparison of the four accuracy evaluation indexes. The producer accuracy and user accuracy of the urban construction roads extracted by the random forest method in the three phase images are respectively as follows: 91.10%, 76.40% in the first-phase image; 87.75%, 89.26% in the second-phase image; 92.70%, 85.04% in the third-phase image. Therefore, the object-oriented random forest method has the best accuracy for the extraction of urban construction roads, which can meet the needs of the extraction of urban construction roads from high-resolution remote sensing images. Using this method, the construction progress monitoring of urban roads can be quickly realized.

[1]文裕. 基于关键链的城市道路改造工程施工进度管理研究[D]. 西安: 西安理工大学, 2016.

[2]包洁玮. 基于BIM技术的道路工程施工进度管理应用研究[D]. 兰州: 兰州交通大学, 2021.

[3]廖小罕, 肖青, 张颢. 无人机遥感: 大众化与拓展应用发展趋势[J]. 遥感学报, 2019, 23(06): 1046-1052.

[4]蔡红玥. 高分辨率遥感图像城市道路自动提取方法研究[D]. 北京: 中国地质大学, 2015.

[5]戴激光, 王杨, 杜阳, 等. 光学遥感影像道路提取的方法综述[J]. 遥感学报, 2020, 24(7): 804-823.

[6]Raziq A, Xu A, Li Y. Automatic extraction of urban road centerlines from high-resolution satellite imagery using automatic thresholding and morphological operation method[J]. Journal of Geographic Information System, 2016, 8(4): 517-525.

[7]Kass M, Witkin A, Terzopoulos D. Snakes: Active contour models[J]. International journal of computer vision, 1988, 1(4): 321-331.

[8]朱长青, 王耀革, 马秋禾, 等. 基于形态分割的高分辨率遥感影像道路提取[J]. 测绘学报, 2004(4): 347-351.

[9]孙可, 张钧萍. 局部自适应方向模板匹配的高分辨率遥感影像道路提取[C]. //2015年光学精密工程论坛论文集. 2015: 509-515.

[10]徐胜军, 欧阳朴衍, 郭学源, 等. 多尺度特征融合空洞卷积 ResNet遥感图像建筑物分割[J]. 光学精密工程, 2020, 28(7): 1588-1599.

[11]符喜优, 张风丽, 王国军, 等. 结合张量投票和Snakes模型的SAR图像道路提取[J]. 中国图象图形学报, 2015, 20(10): 1403-1411.

[12]曹帆之, 徐杨斌, 朱宝山, 等. 利用动态规划半自动提取高分辨率遥感影像道路中心线[J]. 测绘科学技术学报, 2015, 32(6): 615-618+625.

[13]Pudaruth S . Extraction of Roads from Remotely Sensed Images using a Multi-Angled Template Matching Technique[C]// International Symposium on Computer Vision & the Internet. ACM, 2016: 21-29.

[14]李建飞, 文志强, 胡永祥, 等. 基于改进区域生长的遥感影像道路提取[J]. 计算机工程与应用, 2016, 52(2): 209-213+238.

[15]Trinder J C, Wang Y, Sowmya A, et al. Artificial intelligence in 3-D feature extraction[M]//Automatic Extraction of Man-Made Objects from Aerial and Space Images (II). Birkhäuser, Basel, 1997: 257-266.

[16]Ton J, Jain A K, Enslin W R, et al. Automatic road identification and labeling in Landsat 4 TM images[J]. Photogrammetria, 1989, 43(5): 257-276.

[17]Baumgartner A, Steger C, Mayer H, et al. Automatic road extraction in rural areas[J]. International Archives of Photogrammetry and Remote Sensing, 1999, 32(3; SECT 2W5): 107-112.

[18]杨先武, 韦春桃, 吕建刚, 等. 基于形态重建的高分辨率遥感影像城市道路提取[J]. 测绘通报, 2012(1): 45-47.

[19]Dell"Acqua F , Gamba P . Detection of urban structures in SAR images by robust fuzzy clustering algorithms: the example of street tracking[J]. IEEE Transactions on Geoscience & Remote Sensing, 2001, 39(10): 2287-2297.

[20]王文锋, 丁伟利, 李勇, 等. 一种高效的基于平行边缘的道路识别算法[J]. 光学学报, 2015, 35(7): 232-238.

[21]杨孝翠, 孟万利. 基于对比度增强和形态学的遥感影像道路提取[J]. 测绘通报, 2017(8): 45-49.

[22]侯学勇, 刘士荣, 张波涛, 等. 基于聚类和区域生长的道路可行区域提取[J]. 华中科技大学学报(自然科学版), 2013, 41(S1): 276-279.

[23]赵文智, 雒立群, 郭舟, 等. 光谱特征分析的城市道路提取[J]. 光谱学与光谱分析, 2015, 35(10): 2814-2819.

[24]李磊. 一种基于无人机倾斜摄影的三维路网提取方法[J]. 中国公路学报, 2019, 32(11): 219-226+254.

[25]范雯, 何鄂龙, 李天琪, 等. 融合空谱特征的车载LiDAR点云道路标识线提取[J]. 测绘通报, 2018(8): 97-101.

[26]贾建鑫, 孙海彬, 蒋长辉, 等. 多源遥感数据的道路提取技术研究现状及展望[J]. 光学精密工程, 2021, 29(2): 430-442.

[27]李磊. 利用遥感影像和点云数据的道路提取方法研究[D]. 郑州: 战略支援部队信息工程大学, 2018.

[28]彭检贵. 融合点云与高分辨率影像的城区道路提取与表面重建研究[D]. 武汉: 武汉大学, 2012.

[29]Shi W , Miao Z , Debayle J . An Integrated Method for Urban Main-Road Centerline Extraction From Optical Remotely Sensed Imagery[J]. IEEE Transactions on Geoscience & Remote Sensing, 2014, 52(6): 3359-3372.

[30]阙昊懿, 黄辉先, 徐建闽. 基于双阈值SSDA模板匹配的遥感图像道路边缘检测研究[J]. 国土资源遥感, 2014, 26(4): 29-33.

[31]李建, 张其栋. 基于霍夫变换的遥感图像城市道路的提取识别[J]. 电脑知识与技术, 2017, 13(3): 172-174.

[32]谭媛, 黄辉先, 徐建闽, 等. 基于改进Sobel算子的遥感图像道路边缘检测方法[J]. 国土资源遥感, 2016, 28(3): 7-11.

[33]Wang J, Qin Q, Yang X, et al. Automated road extraction from multi-resolution images using spectral information and texture[C]//2014 IEEE Geoscience and Remote Sensing Symposium. IEEE, 2014: 533-536.

[34]Ghaziani M, Mohamadi Y, Koku A B, et al. Extraction of unstructured roads from satellite images using binary image segmentation[C]//2013 21st Signal Processing and Communications Applications Conference (SIU). IEEE, 2013: 1-4.

[35]曾发明, 杨波, 吴德文, 等. 基于Canny边缘检测算子的矿区道路提取[J]. 国土资源遥感, 2013, 25(4): 72-78.

[36]张永宏, 何静, 阚希, 等. 遥感图像道路提取方法综述[J]. 计算机工程与应用, 2018, 54(13): 1-10+51.

[37]张曦. 基于时频特征和支持向量机的高分辨率遥感影像道路提取[D]. 合肥: 安徽大学, 2016.

[38]陈杰, 邵权斌, 梅小明, 等. 结合尺度空间的面向对象高分辨率影像城市道路提取[J]. 测绘工程, 2016, 25(12): 5-11+15.

[39]Huang X , Zhang L . Road centreline extraction from high-resolution imagery based on multiscale structural features and support vector machines[J]. International Journal of Remote Sensing, 2009, 30(8): 1977-1987.

[40]Maboudi M , Amini J , Malihi S , et al. Integrating fuzzy object based image analysis and ant colony optimization for road extraction from remotely sensed images[J]. Isprs Journal of Photogrammetry & Remote Sensing, 2018, 138(APR.): 151-163.

[41]MM, Stein, Bijker, et al. Region-based urban road extraction from VHR satellite images using Binary Partition Tree[J]. INT J APPL EARTH OBS, 2016, 2016, 44(-): 217-225.

[42]李洪川, 楚恒, 霍英海. 城市道路的多特征多核SVM提取方法[J]. 测绘通报, 2018(2): 72-77.

[43]刘如意, 宋建锋, 权义宁, 等. 一种自动的高分辨率遥感影像道路提取方法[J]. 西安电子科技大学学报, 2017, 44(1): 100-105.

[44]Mohan R. Deep deconvolutional networks for scene parsing[J]. arXiv preprint arXiv:1411.4101, 2014.

[45]Caltagirone L, Scheidegger S, Svensson L, et al. Fast LIDAR-based road detection using fully convolutional neural networks[C]//2017 ieee intelligent vehicles symposium (iv). IEEE, 2017: 1019-1024.

[46]朱宏, 张艳, 王欢. 基于卷积神经网络的遥感影像地表覆盖分类[J]. 地理信息世界, 2020, 27(5): 58-64.

[47]Rezaee M, Zhang Y. Road detection using deep neural network in high spatial resolution images[C]//2017 Joint Urban Remote Sensing Event (JURSE). IEEE, 2017: 1-4.

[48]张永宏, 夏广浩, 阚希, 等. 基于全卷积神经网络的多源高分辨率遥感道路提取[J]. 计算机应用, 2018, 38(7): 2070-2075.

[49]Wei Y, Wang Z, Xu M. Road structure refined CNN for road extraction in aerial image[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(5): 709-713.

[50]戴激光, 王杨, 杜阳, 等. 光学遥感影像道路提取的方法综述[J]. 遥感学报, 2020, 24(7): 804-823.

[51]Uuemaa E , Roosaare J , ülo Mander. Scale dependence of landscape metrics and their indicatory value for nutrient and organic matter losses from catchments[J]. Ecological Indicators, 2005, 5(4): 350-369.

[52]Zhang X , Feng X , Xiao P , et al. Segmentation quality evaluation using region-based precision and recall measures for remote sensing images[J]. Isprs Journal of Photogrammetry & Remote Sensing, 2015, 102: 73-84.

[53]Xiao P, Zhang X, Zhang H, et al. Multiscale optimized segmentation of urban green cover in high resolution remote sensing image[J]. Remote Sensing, 2018, 10(11): 1813.

[54]洪亮, 楚森森, 彭双云, 等. 顾及全局和局部最优的高分辨率遥感影像多尺度分割优化算法[J]. 遥感学报, 2020, 24(12): 1464-1475.

[55]Dragut L , Csillik O , Eisank C , et al. Automated parameterisation for multi-scale image segmentation on multiple layers[J]. Isprs J Photogramm Remote Sens, 2014, 88: 119-127.

[56]张俊, 朱国龙, 李妍. 面向对象高分辨率影像信息提取中的尺度效应及最优尺度研究[J]. 测绘科学,2011, 36(2): 107-109+58.

[57]王芳, 杨武年, 王建, 等. 遥感影像多尺度分割中最优尺度的选取及评价[J]. 遥感技术与应用, 2020, 35(3): 623-633.

[58]张锐, 尤淑撑, 杜磊, 等. 基于改进超像素和标记分水岭的高分辨率遥感影像分割方法[J]. 国土资源遥感, 2021, 33(1): 86-95.

[59]王露, 刘庆元. 高分辨率遥感影像多尺度分割中最优尺度选取方法综述[J]. 测绘与空间地理信息, 2015, 38(3): 166-169.

[60]张露洋, 雷国平, 郭一洋, 等. 基于Landsat影像的面向对象土地利用分类——以下辽河平原为例[J]. 应用基础与工程科学学报, 2021, 29(2): 261-271.

[61]毛宁, 刘慧平, 刘湘平, 等. 基于RMNE方法的多尺度分割最优分割尺度选取[J]. 国土资源遥感, 2019, 31(2): 10-16.

[62]张华, 张改改. 面向对象的GF-1遥感影像多尺度分割研究[J]. 甘肃农业大学学报, 2018, 53(4): 116-123.

[63]游炯, 裴志远, 王飞, 等. 基于改进多元纹理信息模型和GF-1影像的县域冬小麦面积提取[J]. 农业工程学报, 2016, 32(13): 131-139.

[64]周春艳, 王萍, 张振勇, 等. 基于面向对象信息提取技术的城市用地分类[J]. 遥感技术与应用, 2008(1): 31-35+123.

[65]吴健生, 刘建政, 黄秀兰, 等. 基于面向对象分类的土地整理区农田灌排系统自动化识别[J]. 农业工程学报, 2012, 28(8): 25-31.

[66]田有文, 程怡, 王小奇, 等. 基于高光谱成像的苹果虫害检测特征向量的选取[J]. 农业工程学报, 2014, 30(12): 132-139.

[67]林雪, 彭道黎, 黄国胜, 等. 结合多尺度纹理特征的遥感影像面向对象分类[J]. 测绘工程, 2016, 25(7): 22-27.

[68]裴欢, 孙天娇, 王晓妍. 基于Landsat 8 OLI影像纹理特征的面向对象土地利用/覆盖分类[J]. 农业工程学报, 2018, 34(2): 248-255.

[69]黄邵东, 徐伟恒, 熊源, 等. 结合纹理和空间特征的多光谱影像面向对象茶园提取[J]. 光谱学与光谱分析, 2021, 41(8): 2565-2571.

[70]温小乐, 钟奥, 胡秀娟. 基于随机森林特征选择的城市绿化乔木树种分类[J]. 地球信息科学学报, 2018, 20(12): 1777-1786.

[71]耿仁方, 付波霖, 金双根, 等. 面向对象的无人机遥感影像岩溶湿地植被遥感识别[J]. 测绘通报, 2020(11): 13-18.

[72]沈盛彧, 张彤, 程冬兵, 等. 融合无人机高分辨率DOM和DSM数据语义的崩岗识别[J]. 农业工程学报, 2020, 36(12): 69-79.

[73]魏青, 张宝忠, 魏征. 基于无人机多光谱影像的地物识别[J]. 新疆农业科学, 2020, 57(5): 932-939.

[74]邬宁珊, 王佳希, 张岩, 等. 基于无人机可见光影像的树种和树冠信息提取——以晋西黄土区蔡家川流域为例[J]. 浙江农业学报, 2021, 33(8): 1505-1518.

[75]王舒. 基于面向对象分析的无人机影像侵蚀沟提取研究[D]. 西安: 西安科技大学, 2019.

[76]杨樟平, 曹碧凤, 谢巧雅, 等. 应用无人机影像提取毛竹林立竹度[J]. 东北林业大学学报, 2021, 49(7): 66-71.

[77]耿仁方, 付波霖, 金双根, 等. 面向对象的无人机遥感影像岩溶湿地植被遥感识别[J]. 测绘通报, 2020(11): 13-18.

[78]刘敏. 面向地理国情普查的地表覆盖解译技术与方法[D]. 阜新: 辽宁工程技术大学, 2013.

[79]Louhaichi M, Borman M M, Johnson D E. Spatially located platform and aerial photography for documentation of grazing impacts on wheat[J]. Geocarto International, 2001, 16(1): 65-70.

[80]Zhengxing W, Chuang L, Alfredo H. From AVHRR-NDVI to MODIS-EVI: Advances in vegetation index research[J]. Acta ecologica sinica, 2003, 23(5): 979-987.

[81]Bendig J, Yu K, Aasen H, et al. Combining UAV-based plant height from crop surface models, visible, and near infrared vegetation indices for biomass monitoring in barley[J]. International Journal of Applied Earth Observation and Geoinformation, 2015, 39: 79-87.

[82]Haralick R M, Shanmugam K, Dinstein I H. Textural features for image classification[J]. IEEE Transactions on systems, man, and cybernetics, 1973 (6): 610-621.

[83]Puissant A, Hirsch J, Weber C. The utility of texture analysis to improve per‐pixel classification for high to very high spatial resolution imagery[J]. International Journal of Remote Sensing, 2005, 26(4): 733-745.

[84]Regniers O, Bombrun L, Lafon V, et al. Supervised classification of very high resolution optical images using wavelet-based textural features[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(6): 3722-3735.

[85]马长辉. 基于高分影像的土地覆盖分类方法研究[D]. 湘潭市: 湘潭大学, 2018.

[86]刘小波. 基于面向对象的多特征综合高分辨率遥感影像变化检测研究[D]. 兰州: 兰州交通大学, 2017.

[87]谢静. 基于多季相遥感信息的三江平原湿地信息提取[D]. 长春: 中国科学院研究生院(东北地理与农业生态研究所), 2013.

[88]胡金梅. 面向对象的高分辨率遥感影像分类算法研究[D]. 合肥: 合肥工业大学, 2020.

[89]Breiman L. Random forests[J]. Machine learning, 2001, 45(1): 5-32.

[90]Breiman L, Friedman J H, Olshen R A, et al. Classification and regression trees[M]. Routledge, 2017.

[91]齐雁冰, 王茵茵, 陈洋, 等. 基于遥感与随机森林算法的陕西省土壤有机质空间预测[J]. 自然资源学报, 2017, 32(6): 1074-1086.

[92]汪小钦, 邱鹏勋, 李娅丽, 等. 基于时序Landsat遥感数据的新疆开孔河流域农作物类型识别[J]. 农业工程学报, 2019, 35(16): 180-188.

[93]陈蕊, 张继超. 基于eCognition的遥感图像面向对象分类方法研究[J]. 测绘与空间地理信息, 2020, 43(2): 91-95.