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GIS的TM数据分类及辅助制图研究

本文初步研究了遥感、GIS和制图一体化实用技术方法。对黄土丘陵区和沙漠地区TM数据进行了特征信息分析;给出了分层分类和GIS辅助分类结果;经模糊推理和人机交互修改,将提“纯”的遥感专题数据作为GIS的动态信息源,对GIS进行扩充与更新;最后在GIS支持下分层提取专题图并进行辅助制图。     

基于规则知识的空间推理研究

在一般空间推理的基础上 ,提出了空间推理扩展 ,即基于规则知识的空间推理。该方法结合了人工智能的基本原理 ,继承并发展了空间推理的方法学 ,其推理过程建立在空间知识与模型集成的基础上 ,以规则知识为推理控制 ,因而更符合实际 ,具有更大的灵活性。实例验证了本方法的正确性。     

RS与GIS集成环境下的遥感图像分析技术

通过多源数据融合，将目前GIS中包含的地理信息作为辅助数据，引入到RS图像分析过程中。GIS数据在空间信息、目标属性和目标范围等方面对RS图像分析具有重要的引导作用。     

利用案例推理（CBR）方法对雷达图像进行土地利用分类

提出了基于案例推理 (CBR)的遥感分类的新方法。基于规则的专家系统被用来提高遥感分类的效率。但所涉及的规则可能多达上百或上千条 ,有的问题根本无法用规则来表达。CBR只是根据以往的案例进行推理 ,克服了基于规则的推理方法的不足 ,可以用来解决一些复杂的资源环境问题。利用CBR和模糊数学结合 ,并通过分层随机采样来控制案例在空间上的分布 ,以反映复杂环境所造成的光谱变化 ,由此解决遥感分类中的“同物异谱”现象。实验表明 ,所建立的案例库可以被重复多次使用 ,分类的效果比监督分类和非监督分类要好。     

Geography Education: The Glass Is Half Full and It's Getting Fuller

This article examines the recent progress of geography education as an academic subfield of geography and as an area of activity in K–12 education. Challenges to both endeavors are described. A restructuring of geography education's research program that focuses on spatial thinking and reasoning and merges applied and theoretical research is proposed.     

A variable precision rough set approach to knowledge discovery in land cover classification

This paper presents a granular computing approach to spatial classification and prediction of land cover classes using rough set variable precision methods. In particular, it presents an approach to characterizing large spatially clustered data sets to discover knowledge in multi-source supervised classification. The evidential structure of spatial classification is founded on the notions of equivalence relations of rough set theory. It allows expressing spatial concepts in terms of approximation space wherein a decision class can be approximated through the partition of boundary regions. The paper also identifies how approximate reasoning can be introduced by using variable precision rough sets in the context of land cover characterization. The rough set theory is applied to demonstrate an empirical application and the predictive performance is compared with popular baseline machine learning algorithms. A comparison shows that the predictive performance of the rough set rule induction is slightly higher than the decision tree and significantly outperforms the baseline models such as neural network, naïve Bayesian and support vector machine methods.     

Evaluation of a rule-based compositing technique for Landsat-5 TM and Landsat-7 ETM+ images

Image compositing is a multi-objective optimization process. Its goal is to produce a seamless cloud and artefact-free artificial image. This is achieved by aggregating image observations and by replacing poor and cloudy data with good observations from imagery acquired within the timeframe of interest. This compositing process aims to minimise the visual artefacts which could result from different radiometric properties, caused by atmospheric conditions, phenologic patterns and land cover changes. It has the following requirements: (1) image compositing must be cloud free, which requires the detection of clouds and shadows, and (2) the image composite must be seamless, minimizing artefacts and visible across inter image seams. This study proposes a new rule-based compositing technique (RBC) that combines the strengths of several existing methods. A quantitative and qualitative evaluation is made of the RBC technique by comparing it to the maximum NDVI (MaxNDVI), minimum red (MinRed) and maximum ratio (MaxRatio) compositing techniques. A total of 174 Landsat TM and ETM+ images, covering three study sites and three different timeframes for each site, are used in the evaluation. A new set of quantitative/qualitative evaluation techniques for compositing quality measurement was developed and showed that the RBC technique outperformed all other techniques, with MaxRatio, MaxNDVI, and MinRed techniques in order of performance from best to worst.