光学遥感影像道路提取的方法综述

道路信息在多个应用领域中发挥着基础性的作用。光学遥感影像能够以较高的空间分辨率对目标地物进行精细化解译，可大幅增强地物目标的提取能力。充分利用光学遥感影像丰富的几何纹理信息，进行道路的精确提取，已成为当前遥感学界研究的热点与前沿问题。鉴于此，本文依据近年来大量相关文献，对现有的理论与方法进行了归类与总结，通过分析不同方法采用的道路特征组合，将道路提取方法划分为模板匹配、知识驱动、面向对象和深度学习4类方法，简要介绍了道路提取普适性的评价指标并对部分方法进行了分析与评价；最后对现有光学遥感影像道路提取的发展提出了建议和展望。     

城市道路的多特征多核SVM提取方法

针对高分辨率遥感影像中城市道路提取的复杂性及SVM的分类性能,提出了一种城市道路的多特征多核SVM提取方法。首先利用FCM算法将原始影像粗分为建成区和非建成区两类,剔除非建成区;然后根据分水岭分割算法分割建成区并提取分割对象的光谱特征与空间特征,以全局核函数和局部核函数加权组合的方式构建多核SVM对建成区进行二次分类,去除建成区中的建筑物等非道路信息;最后利用数学形态学处理,获得最终的道路提取结果。试验结果表明：文中所提方法能够较精确地提取城市道路信息,分类精度高于单核SVM提取及其他对比方法。     

Coregistration of multiangle fine spatial resolution SAR images

Provides a first assessment of a coregistration technique suitable for multiangle fine spatial resolution synthetic aperture radar (SAR) images. The technique is based on crossroad and road junction extraction and matching and exploits recently introduced road extraction routines for SAR data. These features are matched using relational and geometrical analysis. Results are encouraging and show the possibility to exploit multiangle SAR available from future airborne and satellite missions.     

Automatic main road extraction from high resolution satellite imageries by means of particle swarm optimization applied to a fuzzy-based mean calculation approach

Manual extraction of road network by human operator is an expensive and time-consuming procedure. Alternatively, automation of the extraction process would be a great advancement. For this purpose, an automatic method is proposed to extract roads from high resolution satellite images. In this study, using few samples from road surface, a particle swarm optimization is applied to a fuzzy-based mean calculation system to obtain road mean values in each band of high resolution satellite colour images. Then, the images are segmented using the calculated mean values from the fuzzy system. Optimizing the fuzzy cost function by particle swarm optimization enables the fuzzy approach to be the best mean value of road with sub-grey level precision. Initially, this method was applied to simulated images where the calculated mean values are consistent with the hypothetic mean values. Application of the method to IKONOS satellite images has shown a prospective outcome for automatic road extraction. Mathematical morphology is subsequently used to extract an initial main road centreline from the segmented image. Then, small redundant segments are automatically removed. The quality of the extracted road centreline indicates the effectiveness of the proposed approach.     

Automatic urban road extraction using airborne laser scanning/altimetry and high resolution satellite data

Automatic road extraction from remotely sensed images has been an active research in urban area during last few decades. But such study becomes difficult in urban environment due to mix of natural and man-made features. This research explores methodology for semiautomatic extraction of urban roads. An integrated approach of airborne laser scanning (ALS) altimetry and high-resolution data has been used to extract road and differentiate them from flyovers. Object oriented fuzzy rule based approach classifies roads from high resolution satellite images. Complete road network is extracted with the combination of ALS and high-resolution data. The results show that an integration of LiDAR data and IKONOS data gives better accuracy for automatic road extraction. The method was applied on urban area of Amsterdam, The Netherlands.     

航拍图像跨数据域特征迁移道路提取方法

针对传统道路提取方法应用于新数据泛化能力不足的问题,研究了通过特征迁移和编解码网络实现跨数据域的道路提取方法。首先,构建了基于编解码网络的道路提取基本模型,用于实现单一数据来源的道路提取任务。然后,基于道路提取网络结构和循环一致性原则,提出了用于跨数据域图像特征迁移的循环生成对抗网络,使目标域图像映射入源域特征空间。使用预训练的道路提取模型处理特征迁移后的目标域图像,即可实现跨数据域道路提取任务。试验结果表明,本文所提方法能够拓展道路提取网络的泛化能力,准确有效地提取跨数据域图像中的道路目标。相较于未特征迁移的结果,本文所提方法大幅改善了道路提取指标,使得F1提升了50%以上。本文方法不需要目标域的标注信息,也不需要对道路提取网络进行微调训练,而只需训练由目标域向源域的特征迁移模型,所耗时间和人力成本较低,因而具有良好的应用价值。     

An Integrated Multistage Framework for Automatic Road Extraction from High Resolution Satellite Imagery

Automated procedures to rapidly identify road networks from high-resolution satellite imagery are necessary for modern applications in GIS. In this paper, we propose an approach for automatic road extraction by integrating a set of appropriate modules in a unified framework, to solve this complex problem. The two main properties of roads used are: (1) spectral contrast with respect to background and (2) locally linear path. Support Vector Machine is used to discriminate between road and non-road segments. We propose a Dominant singular Measure (DSM) for the task of detecting linear (locally) road boundaries. This pair of information of road segments, obtained using Probabilistic SVM (PSVM) and DSM, is integrated using a modified Constraint Satisfaction Neural Network. Results of this integration are not satisfactory due to occlusion of roads, variation of road material, and curvilinear pattern. Suitable post-processing modules (segment linking and region part segmentation) have been designed to address these issues. The proposed non-model based approach is verified with extensive experimentations and performance compared with two state-of-the-art techniques and a GIS based tool, using multi-spectral satellite images. The proposed methodology is robust and shows superior performance (completeness and correctness are used as measures) in automating the process of road network extraction.     

空间信息感知语义分割模型的高分辨率遥感影像道路提取

道路信息自动化提取已经成为遥感领域热门的研究方向,而基于深度学习的遥感影像道路信息提取方法已经取得了许多成果。但由于受到网络中卷积和池化等操作的影响,基于深度学习的道路提取方法存在着空间特征和地物细节信息丢失等问题,造成许多误提现象。针对此问题,本文设计了一种改进的道路提取语义分割网络模型,该网络以改进的ResNet网络为主体,并引入坐标卷积和全局信息增强模块,用于增强空间信息和全局上下文信息的感知能力,突出道路边缘特征进而确保道路分类的精确性。本文方法在公开道路数据集和高分数据集上获得了显著的提取效果,与其它方法相比取得了明显提高;并且,在一定程度上减少了树木、建筑阴影等自然场景因素遮挡的影响,可以完整准确地提取出道路;此外,模型对多尺度道路也可以实现有效地提取。     

基于 GVF 和 Snake 模型的高分辨率遥感图像四元数空间道路提取

基于Snake模型的优势及缺陷,提出一种高分辨率多光谱图像道路提取方法。通过建立多光谱图像信息的四元数表达式,解决Snake模型应用于多光谱图像的局限性;通过多光谱图像四元数空间标记分水岭分割,获取道路初始轮廓,解决道路初始边界设置问题;通过梯度矢量流模型迭代求解,解决道路凹陷位置逼近问题,最终实现道路提取。用QuickBird 高分辨率多光谱遥感数据验证方法的可行性,与直接采用Snake 模型进行道路提取结果相比,本文所提出的方法对于高分辨率影像上的高等级道路提取具有较好的精度。     

Road condition assessment by OBIA and feature selection techniques using very high-resolution WorldView-2 imagery

Accurate information on the conditions of road asphalt is necessary for economic development and transportation management. In this study, object-based image analysis (OBIA) rule-sets are proposed based on feature selection technique to extract road asphalt conditions (good and poor) using WorldView-2 (WV-2) satellite data. Different feature selection techniques, including support vector machine (SVM), random forest (RF) and chi-square (CHI) are evaluated to indicate the most effective algorithm to identify the best set of OBIA attributes (spatial, spectral, textural and colour). The chi-square algorithm outperformed SVM and RF techniques. The classification result based on CHI algorithm achieved an overall accuracy of 83.19% for the training image (first site). Furthermore, the proposed model was used to examine its performance in different areas; and it achieved accuracy levels of 83.44, 87.80 and 80.26% for the different selected areas. Therefore, the selected method can be potentially useful for detecting road conditions based on WV-2 images.     

Extraction of Alteration Mineral Information from Moderate Remote Sensing Images Using MPS Method

Although alteration minerals related to metallogenesis is very important in mineral exploration, information of alteration mineral is weakly expressed in remote sensing imagery, which is often subject to interfering noise and sometimes limited in spectral and spatial resolutions. Because of easy access, moderate images are the main sources of alteration mineral information. Therefore, it is very important to develop alteration mineral information extraction methods from remote sensing images. In this paper, a combined method based on Mask, principal component analysis (PCA) and support vector machine method (SVM) was used to extract alteration mineral information from Enhanced thematic mapper plus remote sensing data with limited spectral and spatial resolutions. First, a mask image of the remote sensing imagery was created to remove interference information such as vegetation, shadow and water. Then, PCA was employed to collect sample data relating to iron, argillic, and carbonatization alteration. Finally, SVM was used to deal with alteration anomaly and build a feature extraction model of high accuracy. The Mask-PCA-SVM model is used to extract alteration mineral information from remote sensing images of Hatu area, Xinjiang Uygur Autonomous Regions, China. The results show that the new methods proposed in this paper can coincide well with known deposits occurrences, rate reached 86.51%. While, the consistent rate with known deposits of the ratio model, PCA model and Spectral angle mapper model were only 3.37, 65.08 and 69.05% respectively. This suggests that the proposed model can find the actual distribution of mineral deposits more effectively by reducing interference to a greater degree.     

支撑向量机及其遥感影像空间特征提取和分类的应用研究

提出了基于支撑向量机（SVM）的遥感影像空间特征提取的新方法，并以SPOT全色波段影像上城市特征信息的提取为应用实例，并与人工神经网络（ANN）等特征提取方法进行综合比较，认为SVM方法不但能够获得比较高的分类精度，而且在学习速度、自适应能力、特征空间高维不限制、可表达性等方面具有优势。     

Methods and prospects of road and linear structure automatic extraction from remote sensing images

Automatic extraction of road and linear structure from remote sensing images is a very important problem.This paper analyses several existing methods of the automatic road and linear structure extraction by using some multi-spectral remote sensing images acquired from different spatial resolutions,districts and road characteristics. Their advantages and disadvantages have been generalized.     

Methods and prospects of road and linear structure automatic extraction from remote sensing images

Automatic extraction of road and linear structure from remote sensing images is a very important problem. This paper analyses several existing methods of the automatic road and linear structure extraction by using some multi-spectral remote sensing images acquired from different spatial resolutions, districts and road characteristics. Their advantages and disadvantages have been, generalized.     

基于GF2号卫星影像的农业信息提取方法对比分析

以GF2卫星0.8 m全色/3.2 m多光谱分辨率遥感影像为基础数据源,对基于GF2号卫星影像的农业信息提取流程和方法进行了研究与对比分析。首先对GF2号卫星影像进行波谱分析;其次对GF2号影像进行融合,并对多种融合方法进行质量评价;最后选择阈值法、波谱间关系法、非监督分类法和面向对象法分别对GF2号影像数据进行农业信息提取试验,并对信息提取结果进行精度验证和结果分析。试验表明,面向农业信息提取的GF2号卫星影像融合方法中,Pansharp融合算法融合影像色彩正常,无虚影,清晰度高,地类对比度正常,纹理清晰,熵值及与原始多光谱影像的相关系数高。阈值法和谱间关系法适用于提取单要素农业信息,非监督分类法能够初步获取研究区土地利用情况,面向对象法提取研究区全要素信息精度高。总体来说,不同信息提取方法具有各自的优势,在具体实际应用中,可以根据目标地类的波谱特性,选择适宜的遥感影像处理和信息提取方法。     

面向对象和GURLS结合的高空间分辨率遥感数据云检测

遥感信息获取过程中云是重要的干扰因素，随着国产高空间分辨率卫星数据的应用，实现数据的准确云检测对有效获取地面信息具有重要意义。本文以高分一号、高分二号多光谱影像为数据源，利用图像分割获取了同质对象，基于对象光谱、纹理和几何8种属性特征建立了规则集，以规则集为输入，利用阈值法和GURLS分类器结合进行了云检测。针对不同时相和场景的高分数据，将该方法与基于像素的最大似然法和SVM法进行了对比，结果表明该方法云提取精度均在95%以上，Kappa系数在0.9以上。     

兼顾连通性的U-Net网络高分辨率遥感影像道路提取

遥感影像道路提取结果中的断线一方面降低了提取精度，另一方面影响了道路形态完整性，使得提取结果不能直接应用于空间决策与分析。本文基于U-Net网络在高分辨率遥感影像道路提取时全局特征表达的优势，提出一种兼顾连通性的道路断线修复方法完善U-Net网络局部特征表达的劣势。首先，利用数据增强和扩充数据量后的样本数据作为U-Net网络的输入以此训练模型并进行最优模型的道路提取；然后，对提取结果中出现的道路断线以三次多项式曲线拟合的形式进行优化处理。实验表明，与相近网络比较，本文道路提取的精度和形态完整性有了明显的提高，查准率为86.25%，查全率为85.50%，F1-score达到了85.87%。其成果数据能直接地应用于地理决策分析，特别有利于灾后的路径规划，本文提出的方法对道路、电网、轨道、河流等线性地物分类结果中出现类似断线问题具有一定的参考意义。     

改进全卷积网络方法的高分二号影像农村道路提取

针对目前利用高分遥感数据提取农村道路的研究与应用少,提取结果精准度不够的问题,提出了结合空洞卷积和ASPP（Atrous Spatial Pyramid Pooling）结构的改进全卷积农村道路提取网络模型DC-Net（Dilated Convolution Network）。该模型基于全卷积的编解码结构来提取道路深度特征信息,同时针对农村道路细长的特点,在解编码层之间加入了以空洞卷积为基础的ASPP（Atrous Spatial Pyramid Pooling）结构来提取道路的多尺度特征信息,在不牺牲特征空间分辨率的同时扩大了特征感受野FOV（Field-of-View）,从而提高细窄农村道路的识别率。以长株潭城市群郊区部分区域为试验对象,以高分二号国产卫星遥感影像为实验数据,将本文提出的方法与经典的几种全卷积网络方法进行实验结果对比分析。实验结果表明：（1）本文所提出的道路提取模型DC-Net在农村道路的提取上具有可行性,整体提取平均精度达到98.72%,具有较高的提取精度;（2）对比几种经典的全卷积网络模型在农村道路提取上的效果,DC-Net在农村道路提取的精度和连结性、以及树木和阴影的遮挡方面,均表现出了较好的提取结果;（3）本文提出的改进全卷积网络道路提取模型能够有效地提取高分辨率遥感影像中农村道路的特征信息,总体提取效果较好,为提高基于国产高分影像的农村道路提取精度提供了一种新的思路和方法。     

Wavelet-Based Technique to Extract Convective Clouds From Infrared Satellite Images

Extraction of all cumuliform clouds from infrared satellite images is important for cloud studies. Existing methods have focused on extracting only the cumulonimbus clouds. Over monsoon Asia, warm cumulus and cumulus congestus clouds are a large fraction of total cumuliform clouds and are covered by cirrus. An extraction method based only on brightness temperatures (BT) is not sufficient for the detection of these cumuliform clouds. In this letter, a new cloud extraction technique based on spatial characteristics of the convective clouds is presented. The À trous wavelet transform (ATWT) is shown here to successfully extract both shallow and deep convective clouds. The depression in BT caused by the cold cloud tops corresponds to negative wavelet components found by the ATWT.      

一种改进光谱角匹配的水体信息提取方法

快速、准确地从卫星遥感影像上获取水体信息已成为水资源调查及监测、湿地保护、洪涝灾害评估等领域的重要技术手段。本文以GF-4卫星的PMS传感器影像为数据源,提出了一种改进光谱角匹配（MSAM）的水体信息提取方法,以内蒙古中东部和长江中下游两个试验区为研究对象,对比分析了MSAM与单波段阈值法、NDWI阈值法、支持向量机和光谱角匹配等传统方法的水体信息提取结果,并进行了精度评价。试验结果表明,本文提出的MSAM方法不仅能准确地提取水体信息,而且能很好地区分水体与云阴影,对细小水体的提取也具有很好的效果。在内蒙古中部和长江中下游两个试验区的水体提取精度分别达到99.86%和98.37%,在5种水体提取方法中的精度最高,可以有效地提取水体信息。