Some challenges in global land cover mapping: An assessment of agreement and accuracy in existing 1 km datasets

Many investigators need and use global land cover maps for a wide variety of purposes. Ironically, after many years of very limited availability, there are now multiple global land cover maps and it is not readily apparent (1) which is most useful for particular applications or (2) how to combine the different maps to provide an improved dataset. The existing global land cover maps at 1 km spatial resolution have arisen from different initiatives and are based on different remote sensing data and employed different methodologies. Perhaps more significantly, they have different legends. As a result, comparison of the different land cover maps is difficult and information about their relative utility is limited. In an attempt to compare the datasets and assess their strengths and weaknesses we harmonized the thematic legends of four available coarse-resolution global land cover maps (IGBP DISCover, UMD, MODIS 1-km, and GLC2000) using the LCCS-based land cover legend translation protocols. Analysis of the agreement among the global land cover maps and existing validation information highlights general patterns of agreement, inconsistencies and uncertainties. The thematic classes of Evergreen broadleaf trees, Snow and Ice, and Barren show high producer and user accuracy and good agreement among the datasets, while classes of mixed tree types show high commission errors. Overall, the results show a limited ability of the four global products to discriminate mixed classes characterized by a mosaic of trees, shrubs, and herbaceous vegetation. There is a strong relationship between class accuracy, spatial agreement among the datasets, and the heterogeneity of landscapes. Suggestions for future mapping projects include careful definition of mixed unit classes, and improvement in mapping heterogeneous landscapes.     

An application of MODIS data to snow cover monitoring in a pastoral area: A case study in Northern Xinjiang, China

Snow is an important land cover on the earth's surface. It is characterized by its changing nature. Monitoring snow cover extent plays a significant role in dynamic studies and prevention of snow-caused disasters in pastoral areas. Using NASA EOS Terra/MODIS snow cover products and in situ observation data during the four snow seasons from November 1 to March 31 of year 2001 to 2005 in northern Xinjiang area, the accuracy of MODIS snow cover mapping algorithm under varied snow depth and land cover types was analyzed. The overall accuracy of MODIS daily snow cover mapping algorithm in clear sky condition is high at 98.5%; snow agreement reaches 98.2%, and ranges from 77.8% to 100% over the 4-year period for individual sites. Snow depth (SD) is one of the major factors affecting the accuracy of MODIS snow cover maps. MODIS does not identify any snow for SD less than 0.5 cm. The overall accuracy increases with snow depth if SD is equal to or greater than 3 cm, and decreases for SD below 3 cm. Land cover has an important influence in the accuracy of MODIS snow cover maps. The use of MOD10A1 snow cover products is severely affected by cloud cover. The 8-day composite products of MOD10A2 can effectively minimize the effect of cloud cover in most cases. Cloud cover in excess of 10% occurs on 99% of the MOD10A1 products and 14.7% of the MOD10A2 products analyzed during the four snow seasons. User-defined multiple day composite images based on MOD10A1, with flexibilities of selecting composite period, starting and ending date and composite sequence of MOD10A1 products, have an advantage in effectively monitoring snow cover extent for regional snow-caused disasters in pastoral areas.     

Use of a dark object concept and support vector machines to automate forest cover change analysis

An automated method was developed for mapping forest cover change using satellite remote sensing data sets. This multi-temporal classification method consists of a training data automation (TDA) procedure and uses the advanced support vector machines (SVM) algorithm. The TDA procedure automatically generates training data using input satellite images and existing land cover products. The derived high quality training data allow the SVM to produce reliable forest cover change products. This approach was tested in 19 study areas selected from major forest biomes across the globe. In each area a forest cover change map was produced using a pair of Landsat images acquired around 1990 and 2000. High resolution IKONOS images and independently developed reference data sets were available for evaluating the derived change products in 7 of those areas. The overall accuracy values were over 90% for 5 areas, and were 89.4% and 89.6% for the remaining two areas. The user's and producer's accuracies of the forest loss class were over 80% for all 7 study areas, demonstrating that this method is especially effective for mapping major disturbances with low commission errors. IKONOS images were also available in the remaining 12 study areas but they were either located in non-forest areas or in forest areas that did not experience forest cover change between 1990 and 2000. For those areas the IKONOS images were used to assist visual interpretation of the Landsat images in assessing the derived change products. This visual assessment revealed that for most of those areas the derived change products likely were as reliable as those in the 7 areas where accuracy assessment was conducted. The results also suggest that images acquired during leaf-off seasons should not be used in forest cover change analysis in areas where deciduous forests exist. Being highly automatic and with demonstrated capability to produce reliable change products, the TDA-SVM method should be especially useful for quantifying forest cover change over large areas.     

Path Hitting in Acyclic Graphs

An instance of the path hitting problem consists of two families of paths, and ℋ, in a common undirected graph, where each path in ℋ is associated with a non-negative cost. We refer to and ℋ as the sets of demand and hitting paths, respectively. When p∈ℋ and share at least one mutual edge, we say that p hits q. The objective is to find a minimum cost subset of ℋ whose members collectively hit those of . In this paper we provide constant factor approximation algorithms for path hitting, confined to instances in which the underlying graph is a tree, a spider, or a star. Although such restricted settings may appear to be very simple, we demonstrate that they still capture some of the most basic covering problems in graphs. Our approach combines several novel ideas: We extend the algorithm of Garg, Vazirani and Yannakakis (Algorithmica, 18:3–20, 1997) for approximate multicuts and multicommodity flows in trees to prove new integrality properties; we present a reduction that involves multiple calls to this extended algorithm; and we introduce a polynomial-time solvable variant of the edge cover problem, which may be of independent interest. An extended abstract of this paper appeared in Proceedings of the 14th Annual European Symposium on Algorithms, 2006. This work is part of D. Segev’s Ph.D. thesis prepared at Tel-Aviv University under the supervision of Prof. Refael Hassin.     

基于集成的年龄估计方法

近十年来, 由于广泛的应用前景, 关于人脸识别的研究得到了广泛的关注. 但目前有一种影响人脸识别技术的因素尚未被研究者所重视, 那就是年龄变化. 而在适用于年龄变化的人脸识别技术中有一个重要的问题, 即年龄估计. 本文基于典型相关分析和代价敏感学习提出了两种年龄估计算法, 并在此基础上利用集成技术来提高年龄估计的准确性. 最终实验结果验证了本文方法的有效性.     

Approximation algorithms for partially covering with edges

The edge dominating set (EDS) and edge-cover (EC) problems are classical graph covering problems in which one seeks a minimum cost collection of edges which covers the edges or vertices, respectively, of a graph. We consider the generalized partial cover version of these problems, in which failing to cover an edge, in the EDS case, or vertex, in the EC case, induces a penalty. Given a bound on the total amount of penalties that we are permitted to pay, the objective is to find a minimum cost cover with respect to this bound. We give an 8/3-approximation for generalized partial EDS. This result matches the best-known guarantee for the {0,1}-EDS problem, a specialization in which only a specified set of edges need to be covered. Moreover, 8/3 corresponds to the integrality gap of the natural formulation of the {0,1}-EDS problem. Our techniques can also be used to derive an approximation scheme for the generalized partial edge-cover problem, which is -complete even though the uniform penalty version of the partial edge-cover problem is in .     

随机图点覆盖1度顶点核化算法分析

将随机图引入参数计算领域,利用随机图统计和概率分布等特性,从全局和整体上研究参数化点覆盖问题1度点核化过程中问题的核及度分布演变的内在机制和变化规律,并得出关于随机图1度点核化强度与顶点平均度关系及随机图点覆盖问题的决策与度分布关系的两个重要推论.最后分别从MIPS和BIND提取数据进行1度核化实验和分析.初步结果表明,对随机图点覆盖问题的分析方法不仅具有理论上的意义,而且随着问题随机度的大小而对问题有不同程度的把握能力.     

基于鲁棒规范变量分析的故障诊断方法

针对工业过程的建模数据中含有离群点的情况,提出一种基于鲁棒规范变量分析(CVA)的故障诊断方法.该方法使用相关系数的鲁棒估计代替传统的相关系数,通过基于粒子群算法的投影寻踪技术计算最大化鲁棒相关系数的规范变量,从而建立统计模型并监控统计量检测过程的变化.连续搅拌反应器(CSTR)系统的仿真结果说明,鲁棒规范变量分析方法能在含离群点数据的基础上建立准确的统计模型,比规范变量分析更有效地监控过程变化.     

基于C5.0决策树算法的西北干旱区土地覆盖分类研究——以甘肃省武威市为例

西北干旱区面积广阔,由于土地利用类型多样,成因复杂,对环境变化敏感、变化过程快、幅度大、景观差异明显等特点,在影像上表现出的“同物异谱”现象明显 |利用常规目视解译、监督非监督分类、人工参与的决策树分类等方法在效率或精度等方面各有其缺陷。采用机器学习C5.0决策树算法,综合利用地物波谱、NDVI、TC、纹理等信息,根据样本数据自动挖掘分类规则并对整个研究区进行地物分类。机器学习的决策树可以挖掘出更多的分类规则,C5.0算法对采样数据的分布没有要求,可以处理离散和连续数据,生成的规则易于理解,分类精度高,可以满足西北干旱区大面积的土地利用/覆被变化制图的需要。     

基于覆盖粗糙集模型的层次聚类算法

目前大部分聚类算法只适用于处理属性取值为单值的数值型数据,介绍了一种新的基于粗糙集理论的聚类算法,该算法不仅可用于取值为单值的数值型数据聚类,而且能够用于取值为多值的非数值型数据聚类.该算法利用基于相容关系的属性最小覆盖来求解对象各属性的对象属性信息粒.在此基础上,通过对象属性信息粒和对象粗糙相似度的运算构建各对象的相容粒.最后,把具有相同相容粒的对象视为同一等价类,从而实现对论域的聚类,进而对数据对象进行层次聚类.实验结果表明,该算法是可行的.