证据推理的遥感图像分类方法与应用

遥感图像分类一直是遥感研究领域的核心问题。然而,传统遥感分类方法在地物复杂地区不能取得满意效果。不但分类精度不够理想,分类灵活性也存在不足。本文尝试引入证据推理软分类方法,选择小兴安岭山区一景TM遥感图像,基于用户知识和经验,通过人机交互处理,以累计信任度(CBV)最大为划分像元归类的原则得到证据推理方法的分类图像。结果表明,整体分类精度从最大似然法的78.74%提高到82.28%,kappa系数从0.67提高到0.71。但该方法对于裸地分类精度不高,通过人为设定各类别CBV阈值的方法,获得了人为干预的证据推理方法分类图,其整体分类精度达到了87.80%,kappa系数也达到了0.81,所有类别的生产者和用户精度相比于最大似然法都有提高。研究表明,证据推理方法在遥感分类精度和分类灵活性方面都具有优越性。     

基于本体的地理信息查询检索方法研究

传统的GIS查询方式和信息检索方式已经难以满足人们对地理信息查询和检索的越来越广泛的需要,为此需要发展新的地理信息查询检索方法。本文将地理本体的思想引入到信息查询应用中,从查询式优化、查询式扩展、可视化与检索结果知识化等方面展开讨论,是对传统查询检索方法的丰富和发展。在此基础上,设计了基于本体的信息检索模型框架,详细并讨论了实现该检索模型的两个关键技术-自然语言处理技术、本体推理技术。     

基于深度学习的GF-1卫星WFV影像赤潮探测方法

赤潮是我国主要的海洋生态灾害,有效监测赤潮的发生和空间分布对于赤潮的防治具有重要意义。传统的赤潮监测以低空间分辨率的水色卫星为主,但是其对于频发的小规模赤潮存在监控盲区。GF-1卫星WFV影像具有空间分辨率高、成像幅宽大和重访周期短等优点,在小规模赤潮监测中表现出较大的潜力。然而,GF-1卫星WFV影像的光谱分辨率较低,波段少,传统面向水色卫星的赤潮探测方法无法应用于GF-1卫星WFV数据。而且赤潮具有形态多变、尺度不一的特点,难以精确提取。基于此,本文提出了一种面向GF-1卫星WFV影像的尺度自适应赤潮探测网络（SARTNet）。该网络采用双层主干结构以融合赤潮水体的形状特征与细节特征,并引入注意力机制挖掘不同尺度赤潮特征之间的相关性,提高网络对复杂分布赤潮的探测能力。实验结果表明,SARTNet赤潮探测精度优于现有方法,F1分数达到0.89以上,对不同尺度的赤潮漏提和误提较少,且受环境因素的影响较小。     

MARC: a robust method for multiple-aspect trajectory classification via space,time, and semantic embeddings

ABSTRACT

The increasing popularity of Location-Based Social Networks (LBSNs) and the semantic enrichment of mobility data in several contexts in the last years has led to the generation of large volumes of trajectory data. In contrast to GPS-based trajectories, LBSN and context-aware trajectories are more complex data, having several semantic textual dimensions besides space and time, which may reveal interesting mobility patterns. For instance, people may visit different places or perform different activities depending on the weather conditions. These new semantically rich data, known as multiple-aspect trajectories, pose new challenges in trajectory classification, which is the problem that we address in this paper. Existing methods for trajectory classification cannot deal with the complexity of heterogeneous data dimensions or the sequential aspect that characterizes movement. In this paper we propose MARC, an approach based on attribute embedding and Recurrent Neural Networks (RNNs) for classifying multiple-aspect trajectories, that tackles all trajectory properties: space, time, semantics, and sequence. We highlight that MARC exhibits good performance especially when trajectories are described by several textual/categorical attributes. Experiments performed over four publicly available datasets considering the Trajectory-User Linking (TUL) problem show that MARC outperformed all competitors, with respect to accuracy, precision, recall, and F1-score.     

Context inference and prediction modeling in ubiquitous health GIS

The ever‐increasing population in cities intensifies environmental pollution that increases the number of asthmatic patients. Other factors that may influence the prevalence of asthma are atmospheric parameters, physiographic elements and personal characteristics. These parameters can be incorporated into a model to monitor and predict the health conditions of asthmatic patients in various contexts. Such a model is the base for any asthma early warning system. This article introduces a novel ubiquitous health system to monitor asthmatic patients. Ubiquitous systems can be effective in monitoring asthmatic patients through the use of intelligent frameworks. They can provide powerful reasoning and prediction engines for analyzing various situations. Our proposed model encapsulates several tools for preprocessing, reasoning and prediction of asthma conditions. In the preprocessing phase, outliers in the atmospheric datasets were detected and missing sensor data were estimated using a Kalman filter, while in the reasoning phase, the required information was inferred from the raw data using some rule‐based inference techniques. The asthmatic conditions of patients were predicted accurately by a Graph‐Based Support Vector Machine in a Context Space (GBSVMCS) which functions anywhere, anytime and with any status. GBSVMCS is an improved version of the common Support Vector Machine algorithm with the addition of unlabeled data and graph‐based rules in a context space. Based on the stored value for a patient's condition and his/her location/time, asthmatic patients can be monitored and appropriate alerts will be given. Our proposed model was assessed in Region 3 of Tehran, Iran for monitoring three different types of asthma: allergic, occupational and seasonal asthma. The input data to our system included air pollution data, the patients’ personal information, patients’ locations, weather data and geographical information for 270 different situations. Our results showed that 90% of the system's predictions were correct. The proposed model also improved the estimation accuracy by 15% in comparison to conventional methods.     

案例推理的地学应用背景和方法

案例推理 (CBR)研究 ,是人类认知的一种合理性模型 ,接近于人类认识、解决问题最原始的思维方式 ,具有在无法获取机理模型、确定规则或统计模型时 ,采用简单的历史相似性推理实现问题的定量求解和预测 ,成为人工智能领域近年来的研究热点。     

基于本体的地理信息服务查询组件设计

介绍了一种基于本体的服务发现组件设计方法,将基于本体的元数据和基于本体的查询结合起来,提高了基于OGC标准的网络地理服务发现的查询效率和准确性.     

利用PLSA技术进行图像分割

引用概率潜语义分析(PLSA)技术作了航空影像图像分割的尝试,介绍了PLSA图像分割的原理方法和分割的具体算法。通过三幅航空影像分割实验,并与生物地理学图像分割(BBOIS)方法和基于吸引子算法的结果作了对比,结果表明,本文方法具有一定的优越性,是一个很有潜力的图像分割方法。     

面向对象的语义数据模型及其在空间数据库中的应用

传统的数据库模型适合于表示数字字母数据,难以处理复杂的空间数据。面向对象的语义数据模型克服了传统数据模型的许多不足,越来越多地应用于计算机科学各个领域。本文先描述了面向对象的语义数据模型,然后介绍了建立于这种模型基础上的空间数据库实验系统 spobase。