一种中文地址知识库支撑的中文地址分词算法

针对中文地址非结构化、不规范的特点,在构建中文地址模版和中文地址词典等重要知识库基础上,基于中文地址模版引入了中文地址分词的预处理;并在中文地址词典的支撑下采用逆向最大匹配算法实现了中文地址分词。新中文地址分词算法不仅在正确率和召回率等指标上优于传统算法,更提出了一种用于解决未登录地址名词识别问题的新方法。     

基于局部约束的建筑物屋顶点云平面分割方法

建筑物屋顶面点云分割结果的好坏对建筑物三维模型重建起着重要的作用。针对传统RANSAC算法建筑物屋顶面点云的分割问题,提出了一种基于局部约束的建筑物点云平面分割方法。利用点云局部曲面法向约束构建法向准则,利用半径约束的点云空间聚类的方法对共面屋顶面点云进行分解,从而抑制"伪屋顶面"的产生;利用局部抽样策略降低算法的迭代次数,减少运算量。实验表明该方法能够获得稳定可靠的建筑物屋顶点云分割结果,将有利于后续的建筑物三维模型重建。     

高分辨率遥感影像的随机森林变化检测方法

基于面向对象分析(OBIA)的遥感影像变化检测研究已取得显著的进展,代表了遥感影像变化检测的发展范式,未来是发展更加智能的解译分析方法。随机森林作为一种新的机器学习算法,其预测效果和性能稳定性要优于许多单预测器和集成预测方法。本文充分利用OBIA及随机森林机器学习算法的优势,提出了利用随机森林进行面向对象的遥感影像变化检测。首先基于熵率对影像进行超像素分割,通过最优超像素个数评价指数来获取最佳的影像分割结果,并提取每个超像素在前、后时相影像上的光谱特征和Gabor特征作为随机森林的特征输入数据,用于模型的训练。在初始像素级检测结果之上,自动进行分类样本选择并构建分类器模型,用训练好的模型来提取最终的变化区域。利用Quickbird、IKONOS、SPOT-5等3组多光谱影像进行试验,结果表明,本文方法在变化检测精度上要优于对比方法。     

基于无人机影像的城市规划指标计算

在采用传统测绘方式形成测绘成果后,基于该成果进行建筑密度、绿地率、建筑占地面积、绿地面积等现状规划指标计算时,存在时效性差、便捷度低等问题。本文提出了一种利用无人机影像进行现状规划指标计算的思路和方法,开展研究并初步设计了一套基于无人机影像的规划指标计算系统。对广州市番禺区洛溪岛区域进行试验航飞,并利用图像分割方法对城市要素进行分析提取后,进行各类规划指标计算。     

基于GF-1遥感影像面向对象的海岸线自动提取——以秦皇岛为例

近年来,人类活动的影响加上自然因素的作用引起了一系列海岸带生态环境问题,因此,如何快速准确地提取海岸线已成为一个热门的研究课题。本文基于高分一号遥感数据,运用面向对象分类方法,结合光谱、几何等特征在图像上提取海岸线,利用已有"908基线"进行精度验证,得到沙质岸线、淤泥质岸线、生物岸线以及基岩岸线的均方根误差分别为1.7 m、34.79 m、15.27 m和4.69 m。     

Formation and preservation of an overstepped segmented lagoon complex on a high‐energy continental shelf

The duration of shoreline occupation at a given sea‐level, coastal response to sea‐level change and the controls on preservation of various shoreline elements can be recognized by detailed examination of submerged shorelines on the continental shelf. Using bathymetric and seismic observations, this article documents the evolution and preservation of an incised valley and lithified barrier complex between ?65 m and ?50 m mean sea‐level on a wave‐dominated continental shelf. The barrier complex is preserved as a series of aeolianite or beachrock ridges backed by laterally extensive back‐barrier sediments. The ridges include prograded cuspate lagoonal shoreline features similar to those found in contemporary lagoons. The incised valley trends shore‐parallel behind the barrier complex and records an early phase of valley filling, followed by a phase of extensive lagoonal sedimentation beyond the margins of the incised bedrock valley. Sea‐level stability at the outer barrier position (ca ?65 m) enabled accumulation of a substantial coastal barrier that remained intact during a phase of subsequent slow sea‐level rise to ?58 m when the lagoon formed. These lagoonal sediments are stripped seawards by bay ravinement processes which caused the formation of several prograded marginal cuspate features. An abrupt rise in sea‐level to ?40 m, correlated with melt‐water pulse 1B, enabled the preservation of thick lagoonal sediments at the top of the incised valley fill and preservation on the sea bed of the cemented core of the barriers. This situation is unique to subtropical coastlines where early diagenesis is possible. The overlying sandy sediment from the uncemented upper portion of the barriers is dispersed by ravinement, partly burying the ridges and protecting the underlying sediments. The high degree of barrier or shoreline preservation is attributed to rapid overstepping of the shoreline, early cementation in favourable climatic conditions and the protection of the barrier cores by sand sheet draping.     

遥感图像分割的迭代上下文融合小波域HMT模型

由于已有小波域HMT(hidden Markov tree)图像分割算法在上下文融合阶段直接对数据块大小不等的相邻两尺度进行信息融合,导致细节信息分割不充分。为此,提出一种基于迭代上下文融合的小波域HMT模型图像分割算法。该算法在上下文融合阶段采用迭代融合方法,将每一尺度的融合结果作为该尺度的上下文信息再次融合,并设置变化阈值作为迭代终止条件。利用Brodatz纹理组合图像和Formosat-2遥感图像进行分割试验。定性和定量分析表明本文算法能改善图像分割的细节效果,进一步提高图像分割精度。     

面向对象的遥感变化检测研究

通过分析传统的遥感变化检测方法存在的问题,提出了面向对象的遥感变化检测方法。本文利用某地ETM+两个时相的遥感影像,将面向对象和传统变化检测方法进行定性定量的比较,从而得出面向对象的遥感变化检测方法的优势。该方法采用了基于相邻影像区域合并异质性最小的面向对象的多尺度分割方法和模糊分类的方法对变化检测图像进行处理,从而提高了变化检测结果的精度。最终得到较理想的实验分析结果。     

一种利用像素邻域信息的模糊聚类图像分割算法

本文提出一种结合空间信息的模糊C均值聚类图像分割算法。该方法是利用每个像素的邻域像素的隶属度来修正FCM算法的隶属度函数,从而引入图像的空间信息,对隶属函数做了改进;依据平方误差和最小准则,从而确定模糊分类矩阵及聚类中心;并依据最大隶属度原则,划分图像像素的类别归属。实验结果表明,该方法能快速有效地分割图像,并且具有较强的抗噪能力。     

Individual tree crown segmentation in tropical peat swamp forest using airborne hyperspectral data

Abstract

Individual tree crown segmentation is important step for deriving various information for fine-scale analysis of ecological process. However, only several studies have applied tree crown segmentation in tropical forest ecosystems, especially in mixed peat swamp forests. In this study, hyperspectral data were used to detect changes in the biochemical and biophysical characteristics, which are important factors for tree crown segmentation. Principal Component Analysis method was performed to investigate its influence on crown segmentation. Visually Selected PCs, 160 PCs and 160 Spectral Bands image were used and two segmentation techniques; Watershed Transformation and Region Growing segmentation were applied on those images. The highest accuracy was achieved for the crown segmentation is using Region Growing segmentation, based on 1:1 measurement, D value and RMSE value. The results obtained from 160 PCs image using region growing algorithm shows better accuracy with D value of 0.2 (80% accuracy, 20% error) and RMSE of 9.9 m².