基于时空域密度异常的土地利用/土地覆盖短期变化检测

论文分析了时间序列遥感影像中土地利用/土地覆盖短期变化的特点及其时空异常特征, 认为和环境、物候等因素造成的影像变化相比, 由人为活动引起的土地利用/土地覆盖变化具有典型的时间和空间异常特征, 并提出了基于密度异常的土地利用短期变化检测方法。研究工作选取珠江口地区1—5月作物生长期间的3个时间序列Radarsat雷达影像进行试验, 在影像分割的基础上, 构建了基于对象的特征变化矢量, 并将密度异常检测算法(DBAD)扩展到变化矢量的N维特征空间上, 运用随机搜索策略确定检测参数, 对Radarsat时间序列变化矢量中的“小模式”事件进行了检测。检测结果认为, 密度异常检测算法检测的是变化矢量在特征空间的密度分布, 与变化矢量的强度和方向无关, 因此能在时间序列影像中分离出由典型的、正常的作物生长或农事活动引起的影像光谱或回波变化, 进而识别出由人为活动或突发事件导致的土地利用/土地覆盖变化, 这是通常的图像差值等方法难以做到的。进一步的抽样检测说明, 密度异常检测方法对新增建设用地的检测准确率最高(>88%);林地地表覆盖相对稳定, 检测误差也很低(8%);农用地和养殖水面的异常变化检测误差在11%—22%之间;较大的检测误差主要集中在建设用地、农用地和未利用地之间的转换(16%—25%);此外, 养殖水面的检测误差主要集中在河流沿岸及水面变化较大的养殖区域。影像分割结果特别是一些线状分割图斑以及混合地类图斑对误差也有一定的影响。     

四种基于像元的地表覆盖变化检测方法比较

目前遥感影像变化检测方法很多,但各种方法的适用性各不相同。鉴于灰度差值、NDVI差值、灰度比值、主成分分析法在地表覆盖变化检测中应用广泛,文章从数据更新的角度对这4种方法进行了比较;在分析比较这4种方法的单一变化检测精度、检测结果的相同性、相异性的基础上探索了适合于30m分辨率TM地表覆盖变化检测的组合方法。实验结果表明,在地表覆盖变化检测中,有效组合方法能够取得比单一变化检测方法更好的效果;比值法并NDVI差值法并PCA差异法的检测结果中包含了4种单一检测方法所检测出的全部变化像元,达到了最高的生产者精度,比较适合于地表覆盖数据更新制图应用。     

DMC卫星影像在海啸灾情土地覆盖类型变化分析中的应用

利用DMC卫星影像,对在2004年12月26日发生在印度洋海啸中,受灾严重的印度尼西亚亚齐特别自治区进行土地覆盖类型变化比较分析,科学客观的对灾情进行评估,从而阐述了DMC卫星影像在灾情分析及土地覆盖类型监测等方面的应用前景。     

基于改进的最大类间方差法的遥感影像变化检测

变化检测技术被广泛应用于灾情估计、环境监测、农作物生长、森林覆盖等多个领域.为了自动获取变化检测的阈值,本文仔细分析了最大类间方差的方法和原理,改进了二维最大类间方差方法在变化检测中的应用,并利用多套遥感影像数据进行了实验,实验表明了该方法的可行性和有效性.     

基于Sentinel-2影像的雄安新区2016—2019年土地利用分析

雄安新区是国家层面打造的又一个具有重要战略意义的新区,及时准确掌握该地区的土地利用详情具有重要意义。本文利用10 m分辨率的Sentinel-2影像对雄安新区2016—2019年的土地利用进行分类,进而分析该地区的土地利用时空演变。共测试了决策树、随机森林和支持向量机3种分类器,进而获得最高精度的土地分类结果图;同时,利用随机森林的特征排序功能分析了不同特征的重要性。结果表明,雄安新区的耕地、林地、水生植物面积总体均呈显著减少趋势,建设用地面积变化最为显著,表明雄安新区正在进行中、快速的城市化发展。本研究得到的10 m分辨率土地利用专题图和分析结果对于雄安新区的及时监测与规划有着重要参考意义。     

多特征融合的高分辨率遥感影像飞机目标变化检测

为利用高分辨率遥感影像实现高精度的飞机目标变化检测,提出了一种自适应的多特征融合变化检测与深度学习相结合的方法。首先,通过加权迭代的多元变化检测法获取变化强度图,并结合自适应的直方图统计法自动获取显著的变化与不变化样本;然后,提取多时相影像的光谱、边缘和纹理特征,完成多特征融合的变化检测,并通过形态学处理得到变化图斑;最后,利用训练的NIN(Network in Network)结构的卷积神经网络飞机识别模型,完成变化图斑的类型判别,实现变化飞机的检测。实验结果表明,本文方法在两组数据的正确率分别达到100%和91.89%,均优于对比方法,能实现准确可靠的飞机目标变化检测。     

地理国情地表覆盖的城市建设用地扩张分析——以兰州新区为例

针对传统的土地利用变化研究方法无法实现大规模在建中的建设用地动态监测的问题,该文提出了一种新的动态监测方法。结合多时相遥感影像、规划数据和外业调查数据,能够快速、低成本地提取城市建设用地。选取合适的扩展指数构建综合扩展程度指数模型,便于对建设用地变化时序特征做出评价。定量分析建设用地变化时序特征与规划用地的对比情况,明确了城市建设用地的建设进度和符合度。结果表明,城市建设用地处于高速扩展阶段,建设现状和规划情况基本相符;该方法是对地理国情监测工作的一种新的探索。     

利用最小交叉熵的多时相影像变化检测

针对目前多时相遥感影像变化检测中缺少有效的确定变化阈值方法的问题,文章提出一种基于最小交叉熵的方法:采用中值滤波方法分别消除两个时相遥感影像中的噪声;将比值法和差值法融合构造两个时相遥感影像的差异影像;通过最小交叉熵确定差异影像的最佳变化阈值,并对差异影像进行分割得到变化区域;最后对变化检测精度进行评估。实验结果表明,提出的方法具有较高的检测精度,是一种有效的多时相遥感影像变化检测方法。     

基于孪生神经网络的土地利用现状年度变化检测

近年来,利用多期高分遥感影像,构建土地利用现状年度变化检测模型,并以此为基础支持土地利用现状年度变化检测的智能化发展,正是当前研究的难点所在.本文探讨了构建土地利用现状年度变化检测模型存在的问题.通过对土地资数据进行预处理,构建基于孪生神经网络的变化检测模型,以及对模型输出结果进行GIS优化获取目标区域的变化图斑,实现...     

Polarimetric synthetic aperture radar change detection for specific land cover types

This paper presents a supervised polarimetric synthetic aperture radar (PolSAR) change detection method applied to specific land cover types. For each pixel of a PolSAR image, its target scattering vector can be modeled as having a complex multivariate normal distribution. Based on this assumption, the joint distribution of two corresponding vectors in a pair of PolSAR images is derived. Then, a generalized likelihood ratio test statistic for the equality of two likelihood functions of such joint distribution is considered and a maximum likelihood distance measure for specific land cover types is presented. Subsequently, the Kittler and Illingworth minimum error threshold segmentation method is applied to extract the specific changed areas. Experiments on two repeat-pass Radarsat-2 fully polarimetric images of Suzhou, China, demonstrate that the proposed change detection method gives a good performance in determining the specific changed areas in PolSAR images, especially the areas that have changed to water.     

一种适用于粗匹配的遥感影像快速变化发现方法

针对遥感影像变化检测过程中影像间相对配准精度对检测结果影响大,易产生细碎图斑和错检冗余等问题,本文提出了一种基于感知哈希算法适用于粗匹配的遥感影像变化发现方法,以待检测影像的光谱信息和纹理特征信息共同作为检测特征集,以分割后的格网作为检测单元,采用基于感知哈希算法的影像相似度比对方法,计算格网内图像相似度,提取变化位置...     

Modelling the impacts of civil war on land use and land cover change within Kono District,Sierra Leone: a socio-geospatial approach

Capturing the scope and trajectory of changes in land use and land cover (LULC) is critical to urban and regional planning, natural resource sustainability and the overall information needs of policy makers. Studies on LULC change are generally conducted within peaceful environments and seldom incorporate areas that are politically volatile. Consequently, the role of civil conflict on LULC change remains elusive. Using a dense time stack of Landsat Thematic Mapper images and a hybrid classification approach, this study analysed LULC changes in Kono District between 1986–1991, 1991–2002 and 2002–2007 with the overarching goal of elucidating deviations from typical changes in LULC caused by Sierra Leone's civil war (1991–2002). Informed by social survey and secondary data, this study engaged the drivers that facilitated LULC changes during war and non-war periods in a series of spatial regression models in exploring the interface between civil conflict and LULC change.     

Integration of remote sensing and GIS in monitoring the dynamics of land cover in Oroba-Kibos catchment area, Kenya

Land cover in Kenya is in a state of fl ux at different spatial and temporal scales. This compromises environmental integrity and socioeconomic stability of the population hence increasing their vulnerability to the externalities of environmental change. The Oroba-Kibos catchment area in western Kenya is one locality where rapid land use changes have taken place over the last 30 years. The shrubs, swamps, natural forests and other critical ecosystems have been converted on the altar of agriculture, human settlement, fuel wood and timber. This paper presents the results of a study that aimed at providing spatially-explicit information for effective remedial response through (a) Mapping the land cover; (b) Identifying the spatial distribution of land cover changes; (c) Determining the nature, rates and magnitude of the land cover changes, and; (d) Establishing the drivers of land use leading to land cover changes in Oroba-Kibos catchment area. Bi-temporal Landsat TM imagery, fi eld observation, household survey and ancillary data were obtained. Per-fi eld classifi cation of the Landsat TM imagery was performed in a GIS and the resultant land cover maps assessed using the fi eld observation data. Post-classifi cation comparison of the maps was then done to detect changes in land cover that had occurred between 1994 and 2008. SPSS was used to analyze the household survey data and attribute the detected land cover changes to their causes. The fi ndings showed that 9 broad classes characterize the catchment area including the natural forests, swamps, natural water bodies, woodlands, shrublands, built-up lands, grasslands, bare lands and croplands. Croplands are dominant and accounted for about 65% (57122 ha) of the total land in 1994, which increased at the rate of 0.89% to 73% (64772 ha) in 2008, while natural water bodies has the least spatial coverage accounting for about 0.6% (561 ha) of the total land in 1994, which diminished at the rate of 3.57% to 0.3% (260 ha) in 2008. Climate, altitude, access and rights to land, demographic changes, poverty, political governance, market availability and economic returns are the interacting mix of proximate and underlying factors that drive the land cover changes in Oroba-Kibos catchment area.     

遥感影像变化检测方法探讨

利用同一区域不同时相的遥感影像进行变化检测的应用一直是遥感应用研究的热点之一.本文对遥感变化检测技术的研究现状和新进展进行了一定的归纳和总结,对以往分类方法进行了介绍,分别对基于像素、基于对象以及结合特定理论的变化检测方法进行了重点讨论,并对每种方法涉及的理论和最新研究进展进行了评述.     

地表覆盖遥感产品更新完善的研究动向

近年来,多尺度地表覆盖遥感产品的不断涌现,为环境变化研究、地球系统模拟、地理国(世)情监测和可持续发展规划等提供了重要科学数据。为更好地满足广大用户日益增长的应用需求,应对地表覆盖遥感产品进行持续更新完善,保持其时效性、增强时序性、丰富多样性。针对大面积地表覆盖遥感产品更新完善所面临的主要问题,介绍和评述了国内外有关研究动向,包括影像与众源信息相结合的更新、数据类型细化与完善、地表覆盖真实性验证,并作了简要展望。     

Land change attribution based on Landsat time series and integration of ancillary disturbance data in the Athabasca oil sands region of Canada

The Alberta Oil Sands (AOS) is a unique area in Canada undergoing significant disturbance and recovery due to a variety of anthropogenic and natural factors. Accurately quantifying these changes in space and time is important for assessing ecosystem status and trends. In this research, we implemented an approach to combine Landsat time series for the period 1984–2012 with ancillary change datasets to derive detailed change attribution in the AOS. Detected changes were attributed to causes including fire, forest harvest, surface mining, insect damage, flooding, regeneration, and several generic change classes (abrupt/gradual, with/without regeneration) with accuracies ranging from 74% to 100% for classes that occurred frequently. Lower accuracies were found for the generic gradual change classes which accounted for less than 3% of the affected area. Timing of abrupt change events were generally well captured to within ±1 year. For gradual changes timing was less accurate and variable by change type. A land-cover time series was also created to provide information on “from-to” change. A basic accuracy assessment of the land cover showed it to be of moderate accuracy, approximately 69%. Results show that fire was the major cause of change in the region. As expected, surface mine development and related activities have increased since 2000. Insect damage has become a more significant agent of change in the region. Further investigation is required to determine if insect damage is greater than past historical events and to determine if industrial development is linked to the increasing trend observed.     

Soil characterization based on land cover heterogeneity over a tropical landscape: an integrated approach using earth observation data-sets

Soil is a vital part of the natural environment and is always responding to changes in environmental factors, along with the influences of anthropogenic factors and land use changes. The long-term change in soil properties will result in change in soil health and fertility, and hence the soil productivity. Hence, the main aim of this paper focuses on the analysis of land use/land cover (LULC) change pattern in spatial and temporal perspective and to present its impact on soil properties in the Merawu catchment over the period of 18?years. Post classification change detection was performed to quantify the decadal changes in historical LULC over the periods of 1991, 2001 and 2009. The pixel to pixel comparison method was used to detect the LULC of the area. The key LULC types were selected for investigation of soil properties. Soil samples were analysed in situ to measure the physicochemical soil properties. The results of this study show remarkable changes in LULC in the period of 18?years. The effect of land cover change on soil properties, soil compaction and soil strength was found to be significant at a level of <0.05.     

决策级融合的多分辨率遥感影像变化检测

针对多时相、多分辨率遥感影像数据的特点,充分考虑不同分辨率数据和不同变化检测应用的需求,将由粗到精数据集分层检测和决策级融合的思想引入到变化检测,以多时相多分辨率ALOS遥感影像为例,构建并试验了由粗到精变化检测的技术流程.该方法将ALOS多光谱数据视为粗数据集,将全色数据和融合数据视为精数据集,通过对3种数据集变化检...     

面向类型特征的自适应阈值遥感影像变化检测

为了实现两个不同年份单时相遥感影像之间的土地覆盖变化检测，提出了一种基于土地覆盖类型特征自适应确定阈值的遥感影像变化检测方法。以2015年土地覆盖数据为基础，综合2013年和2015年Landsat 8-OLI影像数据，首先，采用时相不变点群法TIC（Temporally Invariant Cluster）保证了两期影像辐射水平的一致性。其次，对两期影像进行多尺度分割，并在各级尺度下构建分割对象的变化向量。然后，采用最大类间方差的方法分别进行单一变化阈值变化检测以及基于土地覆盖类型的多阈值变化检测分析，并利用目视解译样点进行精度验证与评价。结果表明：（1）单一阈值变化检测结果的总体精度为79.6%，Kappa系数为0.601，多阈值变化检测结果的总体精度为87.2%，Kappa系数为0.741，多阈值变化检测具有更高的精度。（2）进一步逐土地覆盖类型精度评价可知，多阈值变化检测能在一定程度上减弱物候期的影响，具有更高的稳定性。该研究以土地覆盖数据为底图，逐类别的选取变化检测阈值，提高了变化区域检测的精度，在大范围高效更新土地覆盖数据的应用中具有一定的参考价值。     

多层特征与上下文信息相结合的光学遥感影像目标检测

目标检测是遥感影像分析的基础和关键。针对光学遥感影像中目标尺度多样、小目标居多、相似性及背景复杂等问题,本文提出一种将卷积神经网络(CNN)和混合波尔兹曼机(HRBM)相结合的遥感影像目标检测方法。首先设计细节—语义特征融合网络(D-SFN)提取卷积神经网络低层和高层融合特征,提升目标特征的判别力,特别是小目标;其次考虑上下文信息对目标检测的影响,结合上下文信息进一步加强目标表征的准确性,提升检测精度。在NWPU数据集上试验表明,本文方法能够显著提升目标检测精度且具有一定程度的稳健性。