Classification-based vehicle detection in high-resolution satellite images

In this study, we have looked into the problem of vehicle detection in high-resolution satellite images. Based on the input from the local road authorities, we have focused not only on highways, but also on inner city roads, where more clutter is expected. The study site is the city of Oslo, Norway. To do vehicle detection in these areas, we propose an automatic approach, consisting of a segmentation step, followed by two stages of object classification. In the process, we utilize multispectral images, panchromatic images and a road network. The approach has been tested on Quickbird images, and the results that are obtained have been compared with manual counts and classifications.     

Detecting shadows in multi-temporal aerial imagery to support near-real-time change detection

Multi-temporal aerial imagery captured via an approach called repeat station imaging (RSI) facilitates post-hazard assessment of damage to infrastructure. Spectral-radiometric (SR) variations caused by differences in shadowing may inhibit successful change detection based on image differencing. This study evaluates a novel approach to shadow classification based on bi-temporal imagery, which exploits SR change signatures associated with transient shadows. Changes in intensity (brightness from red–green–blue images) and intensity-normalized blue waveband values provide a basis for classifying transient shadows across a range of material types with unique reflectance properties, using thresholds that proved versatile for very different scenes. We derive classification thresholds for persistent shadows based on hue to intensity ratio (H/I) images, by exploiting statistics obtained from transient shadow areas. We assess shadow classification accuracy based on this procedure, and compare it to the more conventional approach of thresholding individual H/I images based on frequency distributions. Our efficient and semi-automated shadow classification procedure shows improved mean accuracy (93.3%) and versatility with different image sets over the conventional approach (84.7%). For proof-of-concept, we demonstrate that overlaying bi-temporal imagery also facilitates normalization of intensity values in transient shadow areas, as part of an integrated procedure to support near-real-time change detection.     

Quantifying the building stock from optical high-resolution satellite imagery for assessing disaster risk

This study uses high-resolution (HR) satellite imagery to quantify the stock of buildings, referred herein as building stock. The risk assessment requires information on the natural hazards and on the element at risk, that is the building stock in this article. This study combines (1) texture-based image processing to map built-up areas, (2) statistical sampling that allows locating the building samples and (3) photo-interpretation to encoding building footprints. Statistical inference is then used to quantify the building stock per class of building size. Legaspi in the Philippines is used as a case study. The results show that texture-based computer algorithms provide accurate area estimations of the built-up, that the detail of HR imagery allows the mapping of single buildings using photo-interpretation, and that a systematic sampling approach that uses building encoding and built-up maps can be used to quantify the building stock.     

A cognitive framework for road detection from high-resolution satellite images

Road network extraction from high-resolution satellite (HRS) imagery is a complex task. It is an important field of research and is widely used in various cartographic applications such as updating and generating maps. The objective of this research work is to develop a novel framework, emulating human cognition, for detection of roads from HRS images. Roads network from HRS images are detected using support vector machines within the different stages of cognitive task analysis. In the first stage, basic information about the cognitive parameters which are required for image interpretation is collected. In the second stage, the rule-based method is used for knowledge representation. Lastly, during knowledge elicitation, the developed rules are used to extract roads from HRS images. The proposed method is validated using 16 HRS images of developed suburban, developed urban, emerging suburban and emerging urban region.     

高分辨率卫星影像车辆检测的抗体网络

将车辆检测的过程视为一种“抗体”检测“危险抗原”的过程, 其中车辆是“危险抗原”, 车辆检测模板是“抗体”。利用一些车辆图像作为训练样本, 建立一种抗体网络学习并获取一组优化的“抗体”。这些“抗体”经过与待测影像的匹配, 实现对道路车辆目标的有效提取。采用0.6m分辨率的QuickBird全色数据进行实验, 实验结果验证了该方法的有效性和可行性。     

Evaluating the performance of high-resolution satellite imagery in detecting ephemeral water bodies over West Africa

Pastoralism is a major occupation in Africa’s Sahel region, which spans from Senegal in the west to Ethiopia in the east and has a semi-arid climate. Pastoralists from the region are usually nomadic, moving regularly in search of water and pasture, often for hundreds of kilometers. Traditionally, pastoralists have relied on historical knowledge and transhumance pathways to reach the water bodies (WB). Due to climatic conditions, a vast majority of the WBs in this region are small, sporadic, and ephemeral, therefore there is a need to map and frequently monitor water availability. Remote sensing observations can be used to complement current monitoring efforts. High-resolution near-daily imagery from the recently launched PlanetScope (PS) constellation of small satellites has the potential to complement the existing medium-resolution and infrequent but well established and extensively calibrated sensors like Landsat. The aim of this study is to investigate the performance and applicability of high-resolution PS data to detect and monitor WBs in the Ferlo region of West Africa, in comparison to Landsat-8 imagery. The results indicate that for cloud-free days, PS data was comparable (r >0.88; ubRMSE <0.01) to Landsat, but due to inadequate cloud masking, the results were poor for cloudy days (r < 0.49 and ubRMSE>0.058). The PS images were able to detect nearly 95% of the WBs, whereas Landsat was able to identify only 32% at the water fraction threshold >40%. Initial results from this study show that the high spatial and temporal resolution PS datasets show promise in monitoring WBs in arid regions of West Africa.     

Automatic detection of residential buildings using LIDAR data and multispectral imagery

This paper presents an automatic building detection technique using LIDAR data and multispectral imagery. Two masks are obtained from the LIDAR data: a ‘primary building mask’ and a ‘secondary building mask’. The primary building mask indicates the void areas where the laser does not reach below a certain height threshold. The secondary building mask indicates the filled areas, from where the laser reflects, above the same threshold. Line segments are extracted from around the void areas in the primary building mask. Line segments around trees are removed using the normalized difference vegetation index derived from the orthorectified multispectral images. The initial building positions are obtained based on the remaining line segments. The complete buildings are detected from their initial positions using the two masks and multispectral images in the YIQ colour system. It is experimentally shown that the proposed technique can successfully detect urban residential buildings, when assessed in terms of 15 indices including completeness, correctness and quality.     

国产高分辨率卫星影像云检测方法分析

云检测方法大都针对特定的传感器或依赖多个波段,对参数要求高,而国产高分辨率卫星影像通常包含波段数较少,多数云检测方法不适用。本文采用深度学习的方法,以融合后的高分一号影像为例,应用基于双重视觉注意机制模型进行云检测,并与人工采集、全卷积网络模型的检测结果进行对比。理论分析和研究结果表明:基于双重视觉注意机制的模型云检测结果与人工采集进行对比,正确率为0.986 4;通过增加云样本数量和非云样本数量可有效解决模型对道路、河流、居民地的误检测问题;基于双重视觉注意机制的模型与全卷积网络模型相比,云边界更为准确,模型适用性更强。利用较少的波段信息进行云检测为国产其他高分辨率卫星影像云检测提供了参考。     

面向对象的多特征分级CVA遥感影像变化检测

变化矢量分析CVA方法在中低分辨率遥感影像变化检测中已得到广泛应用,但由于高分辨率遥感影像存在不同地物尺度差异大、不同类别地物光谱相互重叠的问题,因此对于高分影像的变化检测具有局限性。为提高高分影像变化检测精度,提出了一种面向对象的多特征分级CVA变化检测方法,首先,利用基于区域邻接图的影像分割方法分别对两时相遥感影像进行多尺度分割,提取分割图斑的光谱、纹理和形状特征;然后,在各级尺度下,分别运用随机森林方法进行特征选择,计算CVA变化强度图;最后,根据信息熵对多级变化强度图进行自适应融合,利用Otsu阈值法检测变化区域,并与仅考虑光谱特征的分级CVA变化检测方法、像元级多特征CVA变化检测方法以及仅考虑光谱特征的像元级CVA变化检测方法进行比较分析。实验表明:与比较方法相比,本文方法的变化检测精度较高,误检率和漏检率较低。     

资源三号影像中城市高大地物阴影检测方法

针对遥感影像处理中阴影检测和信息补偿不准确的问题,该文在研究已有阴影检测算法的基础上,结合资源三号(ZY-3)影像数据的特性,构建了阴影检测方法:首先对原始图像分别做差值运算和主成分变换,并利用多峰阈值自动提取算法检测出阴影区域;其次将差值运算提取的粗阴影区域与主成分变换提取的阴影区域做并运算生成一个新的阴影区域;然后判断影像中是否含有水体,如果含有水体则利用多峰阈值自动提取算法检测出水体并与新合并的阴影区域影像做布尔运算得到完整的阴影区域,反之则新合并的区域即为完整的阴影区域。实验结果表明该方法针对ZY-3具有较好的普适性、较高的提取精度和提取效率。     

Road detection from high-resolution satellite images using artificial neural networks

This article treats the possibility of using artificial neural networks for road detection from high-resolution satellite images on a part of RGB Ikonos and Quick-Bird images from Kish Island and Bushehr Harbor, respectively. Attempts are also made to verify the impacts of different input parameters on network's ability to find out optimum input vector for the problem. A variety of network structures with different iteration times are used to determine the best network structure and termination condition in training stage.It was found that when the input parameters are made up of spectral information and distances of pixels to road mean vector in a 3 × 3 window, the network's ability in both road and background detection can be improved in comparison with simple networks that simply use spectral information of a single pixel in their input vector.     

Introducing mapping standards in the quality assessment of buildings extracted from very high resolution satellite imagery

Many municipal activities require updated large-scale maps that include both topographic and thematic information. For this purpose, the efficient use of very high spatial resolution (VHR) satellite imagery suggests the development of approaches that enable a timely discrimination, counting and delineation of urban elements according to legal technical specifications and quality standards. Therefore, the nature of this data source and expanding range of applications calls for objective methods and quantitative metrics to assess the quality of the extracted information which go beyond traditional thematic accuracy alone. The present work concerns the development and testing of a new approach for using technical mapping standards in the quality assessment of buildings automatically extracted from VHR satellite imagery. Feature extraction software was employed to map buildings present in a pansharpened QuickBird image of Lisbon. Quality assessment was exhaustive and involved comparisons of extracted features against a reference data set, introducing cartographic constraints from scales 1:1000, 1:5000, and 1:10,000. The spatial data quality elements subject to evaluation were: thematic (attribute) accuracy, completeness, and geometric quality assessed based on planimetric deviation from the reference map. Tests were developed and metrics analyzed considering thresholds and standards for the large mapping scales most frequently used by municipalities. Results show that values for completeness varied with mapping scales and were only slightly superior for scale 1:10,000. Concerning the geometric quality, a large percentage of extracted features met the strict topographic standards of planimetric deviation for scale 1:10,000, while no buildings were compliant with the specification for scale 1:1000.     

Normalizing shadows in multi-temporal aerial frame imagery using relative radiometric adjustments to support near-real-time change detection

This study addresses the problem of shadows in multi-temporal imagery, which is a key issue with change detection approaches based on image comparison. We apply image-to-image radiometric normalizations including histogram matching (HM), mean-variance (MV) equalization, linear regression based on pseudo-invariant features (PIF-LR), and radiometric control sets (RCS) representing high- and low-reflectance extrema, for the novel purpose of normalizing brightness of transient shadows in high spatial resolution, bi-temporal, aerial frame image sets. Efficient shadow normalization is integral to remote sensing procedures that support disaster response efforts in a near-real-time fashion, including repeat station image (RSI) capture, wireless data transfer, shadow detection (as precursor to shadow normalization), and change detection based on image differencing and visual interpretation. We apply the normalization techniques to imagery of suburban scenes containing shadowed materials of varied spectral reflectance characteristics, whereby intensity (average of red, green, and blue spectral band values) under fully illuminated conditions is known from counterpart reference images (time-1 versus time-2). We evaluate the normalization results using stratified random pixel samples within transient shadows, considering central tendency and variance of differences in intensity relative to the unnormalized images. Overall, MV equalization yielded superior results in our tests, reducing the radiometric effects of shadowing by more than 85 percent. The HM and PIF-LR approaches showed slightly lower performance than MV, while the RCS approach proved unreliable among scenes and among stratified intensity levels. We qualitatively evaluate a shadow normalization based on MV equalization, describing its utility and limitations when applied in change detection. Application of image-to-image radiometric normalization for brightening shadowed areas in multi-temporal imagery in this study proved efficient and effective to support change detection.     

遥感影像云检测网络泛化性能研究：以DeepLabv3+为例

近年来,深度学习算法得到了长足的发展,并开始应用于云检测。但是深度神经网络模型参数众多,依赖大量训练样本,因此理解其泛化性能对于深度学习在不同遥感影像的实际应用具有重要的参考价值。本文以深度语义分割算法DeepLabv3+为例,以一组广泛使用的云标记数据集"L8 Biome"为验证数据,探讨该算法用于云检测时在不同的地表景观、空间分辨率和光谱波段组合的遥感影像上的泛化性能。云标记数据集"L8Biome"包含96景具有全球代表性的Landsat 8 OLI影像及相应的人工云掩膜,被广泛用于测试云和云阴影检测算法性能。首先,利用Landsat 8 OLI云标记数据集"L8 Biome",构建不同类型景观、不同空间分辨率、不同波段组合的训练影像集和测试影像集;其次基于不同训练样本集和测试集,评估了DeepLabv3+算法在不同情况下的云检测精度,并与Fmask算法作对比分析。研究结果表明:(1)使用全混合景观类型的训练集训练出来的云检测网络在总体检测精度(92.81%)与稳定度(标准差12.08%)上都优于使用单一景观类型的训练集训练得到的云检测网络,也优于Fmask的总体精度(88.75%)与稳定度(标准差17.34%),说明在构建深度学习算法的训练集时,应该尽可能包含多类型的地表景观;(2)将全混合景观训练集中剔除一类景观的样本(冰/雪景观除外)构建的"混合-1"训练集与全混合景观训练集训练的DeepLabv3+网络的云检测精度也相差不大,说明现有训练样本集已具备较强的景观泛化能力;(3)基于30 m空间分辨率的全混合景观训练样本集训练得到的DeepLabv3+云检测网络在不同分辨率(30 m、60 m、120 m、240 m)的测试集上云检测精度差异不大,都取得较好的效果,说明DeepLabv3+能够泛化应用于不同空间分辨率的遥感影像,相反Fmask直接应用于低分辨率影像时精度明显下降;(4) DeepLabv3+能充分自适应不同波段的信息用于云检测,总体来说更多的光谱波段输入能够提高DeepLabv3+的云检测的精度和稳定度,其中短波红外波段对于DeepLabv3+区分冰/雪与云具有重要价值,而热红外波段对DeepLabv3+云检测网络的性能提升很微小。以上结果说明利用现有数据集"L8 Biome"训练的DeepLabv3+云检测网络能够适用于多种类型的遥感影像,并优于Fmask算法。     

Generic rule-sets for automated detection of urban tree species from very high-resolution satellite data

The sustainable management and monitoring of urban forests is an important activity in the urbanized world, and operational approaches require information about the status of urban trees to determine the best strategy. One limitation in urban forest studies is the detection and discrimination of tree species using limited training data. Thus, this study focuses on developing generic rule sets from high-resolution WorldView-2 imagery in conjunction with spectral, spatial, colour and textural information for automated urban tree species detection. The object-based image analysis and its combination with statistical analysis of object features is utilized for this purpose. Results of attribute selection indicated that from 55 attributes, only 26 were useful to discriminate urban tree species, namely Messua ferrea L., Samanea saman and Casuarina sumatrana. Finally, the high overall accuracy, approximately 86.87% with kappa of 0.75 confirmed the transferability of the generic model.     

Interpretation of satellite imagery for delineation of ravines

Ravines have a vast agricultural potential Survey of ravines has attracted the attention of the Photo-interpreters for quite sometime. LANDSAT, imagery offer a new scope in this field. Study of a LANDSAT imagery was made in order to find out the possibility of delineating broad reclamation units and their areal extent. Pattern analysis with visual interpretation shows that ravines could be classified into at least four reclamation units which could be useful for planning purposes.     

Automated detection of slum area change in Hyderabad,India using multitemporal satellite imagery

This paper presents an approach to automated identification of slum area change patterns in Hyderabad, India, using multi-year and multi-sensor very high resolution satellite imagery. It relies upon a lacunarity-based slum detection algorithm, combined with Canny- and LSD-based imagery pre-processing routines. This method outputs plausible and spatially explicit slum locations for the whole urban agglomeration of Hyderabad in years 2003 and 2010. The results indicate a considerable growth of area occupied by slums between these years and allow identification of trends in slum development in this urban agglomeration.     

高分辨率光学卫星影像全球重点区域三维建模方法——以迪拜市为例

针对境外典型区域精细化三维建模数据获取与产品制作技术难点,本文设计归纳出一种高分辨率光学卫星影像精细化三维建模方法。选取迪拜市中心城区为试验区域,以WorldView-3、GeoEye-1等主流高分卫星影像为主要数据源,以互联网众源影像为辅助数据源,利用卫星影像精准定向精化的RPC参数及生成的精准几何信息（DSM/DOM）,完成区域内主要建筑物的三维模型生产制作,并形成高分辨率卫星影像三维建模原型系统,满足全球地理信息资源建设与维护更新工程,快速制作符合精度和精细度要求的三维数字模型的需要。     

像斑直方图相似性测度的高分辨率遥感影像变化检测

基于像斑的变化向量分析法CVA(Change Vector Analysis)过分依赖像斑的灰度均值信息,而未能有效利用其灰度分布信息,这在高分辨率遥感影像变化检测中存在不足。本文提出了一种基于像斑直方图相似性测度的变化检测方法。利用G统计量构建不同时期像斑之间的相似性测度。假设所有像斑的相似性测度值符合混合高斯分布模型,通过期望最大化算法EM(Expectation Maximization)求解相关参数,最后采用基于最小错误率的贝叶斯判别规则获取最终的变化结果。实验表明,本文提出的上述方法能够有效提高变化检测的精度。