【作者】 鲍文东；

【导师】 靳奉祥；

【作者基本信息】 山东科技大学 ， 大地测量学与测量工程， 2007， 博士

【摘要】 土地利用／覆被变化(简称LUCC)是当前全球环境变化研究领域的前沿课题，而开展典型地区的案例研究是认识全球变化的重要途径。我国正经历着一场前所未有的城市化浪潮，快速城市化背景下的大规模土地利用／覆被变化已经成为一种典型的人地系统相互作用过程。本文选择东部沿海发达城市-济南市作为研究区，构建了基于光谱特征和地学辅助知识的决策树信息提取模型，提高了Landsat TM遥感数据分类的精确度和可靠性，并建立了三期土地利用时空数据库：综合地理学、景观生态学、统计学、分形理论、数据挖掘、土地评价学等多种方法，集成—系列模型，采用定性与定量相结合的手段，从中、小两个尺度、不同侧面系统研究和分析了土地利用动态变化的特征、机制、过程、趋势与规律，不仅可为该区域的土地资源管理和利用提供决策依据，也可为国内其它经济发达地区的同类研究提供借鉴。中尺度宏观过程研究中，以“时间过程”与“空间格局”为主线，采用动态度、综合程度指数，空间洛伦茨曲线、转移矩阵、景观指数、半变异函数、分形维数等模型，对土地利用时空变化进行了深入剖析；小尺度微观过程研究，选取经十路一线为特定研究区，基于地学信息图谱和地类邻接度指数模型，揭示和凸显了城市扩张占用耕地的特征和规律。将定性分析和定量计算相结合，采用缓冲区分析、主成分分析、岭回归分析、BP神经网络、典型相关分析等方法对土地利用变化驱动力进行了系统的研究，揭示了驱动因素在不同时空尺度和区域背景条件下的多样性以及空间联系的复杂性。基于小尺度典型城市边缘区，分别采用“自上而下”和“自下而上”模式思路，构建了“典型相关+逻辑回归+CLUE-S”耦合模型和“MARKOV+多评价规则+CA”耦合模型，并成功运用到历城区的实证研究中，重建了土地利用空间格局的动态演变过程，并对未来土地的空间结构动态发展进行了预测，加深了对土地利用空间扩展过程和规律的理解与认知，为小尺度区域土地利用规划和宏观调控提供了可操作性的思路和模式。鉴于耕地保护和粮食安全的严峻形势，提出了基于GIS技术的耕地质量自动化评价流程和方法，研发了“耕地分等信息系统”，完成了海量数据的空间计算和处理。在耕地分等的基础上，将综合评价算法模型移植到基本农田空间配置过程中，并基于松散耦合模式，开发了耕地入选基本农田决策系统，实现了基本农田空间配置的自动化，为土地利用规划的编制和耕地、基本农田的保护提供了科学的依据，具有较强的技术可行性和现实意义。本文的研究内容结合“信息模式识别理论研究及其在地学中的应用”(40074001)和“多元数据的模式信息研究及其在测量中的应用”(40574001)国家自然科学基金项目、国土资源部“山东省土地利用基础图件更新和信息系统建设”和“山东省农用地分等定级”大调查项目、省计委“山东省土地资源遥感综合调查”等课题，并研制开发“耕地分等信息系统”和“耕地入选基本农田决策支持系统”，获得了业内人士的关注和好评。更多还原

【Abstract】 Researches on LUCC (Land use/ cover change) now are the frontier problem for discussion in the researches on the global environmental change. However, studies on the dynamic change case of LUCC in typical areas are one of the most important ways to understand global changes. Urbanization has already become one of the major trends for human to change the Earth and the LUCC. Urbanization is also turning into one typically interactional process between human and the Earth.In the paper, Jinan city, which is the east-coast developed city in China, is selected as a typical area for this case study. Based on landsat TM remoted data, a model of decision tree is proposed on the basis of spectral characteristics and geosciences information, so classified precision and reliability are hence improved. And databases of land use in three periods are built. Combined with geography, landscape biocenology, statistics, fractal, data mining, science of land evaluation, a series of model, this thesis, by making use of the combination of qualitative and quantitative, analyzes the character, machine, process, change tendency and regulation of dynamic change of land use from medium scale and small scale. The results can not only provide Jinan governmental departments with the basis for the decision-making, but also a reference for a similar area.By studying on macroscopic process of regional land use change in medium scale, along with "temporal process" and "spatial pattern" axes, we analyze the spatial and temporal change of land use of Jinan, combined with land use dynamic degree, the model of composite degree index, Lorenz’s Curve, transformation matrix, the model of landscape pattern indices, distribution index model, half-variation triangle function model. As for the study on microscopic process of regional land use change in small scale, the author chooses Jingshi Road from the east of Lixia Area to the west of Licheng Area, and indicates characteristic and regularities about urban expansion occupying arable land.This paper, combined quantitative statistical calculation with qualitative comparative analysis, studies systematically driving force of land use change used by buffer analysis, principal component analysis, ridge regression analysis, BP neural network and canonical correlation analysis, et al. Then the study above indicates diversity of driving factors, which is between different spatial and temporal scale and regional background condition, and complexity of spatial connection.As for the typical urban border district in small scale, the author here uses top-. down thought and bottom-up thought separately, builds the "logic regression+canonical correlation+CLUE-S" coupling model and the "MARKOV+Multi-criteria Evaluation+CA" coupling model, and applicated empirical accounting research of Licheng successfully. Then the author simulates the dynamic development course of land use spatial pattern, and predicts the dynamic development of land use spatial land structure in the future. By comparison, space development and mechanism of land use can be apprehended and recognized deeply. The foundation and reference were provided to study on land use dynamic change. The ideas and patterns for land use planning and macroeconomic regulation were provided to study on land use dynamic change.Based on the importance of arable land protection and food security, the author puts forward automatic evaluation process and techniques of the quality of arable land based on GIS. Arable land grading evaluation information system is developed based on GIS assembly. Mass data processing and spatial calculation can be achieved. Based on classification of arable land, overall evaluation model is transplanted successfully in automatic spatial allocation of basic farmland, so we build the model of decision indication system and algorithm about arable land selected as basic farmland. Priority result will be gained by comprehensive evaluation and ideal point method, so we can carry through space orientation based on the result and conservation index. Software system of arable land selected basic farmland and decision-making is developed by loosely coupled model. It achieves automation and scientification of spatial allocation of basic farmland. The scientific basis is provided for compilation about planning of land use and the arable land and basic farmland protection. It is more feasible in technique and has realistic meaning.The research combines with the projects of "The study for pattern recognition and tis application to geoscience" and "The study for pattern recognition of multivariate data and tis application to survey" from National Natural Foundation Fund,"A Software System for Updating Land Use Maps and Information System Construction in Shandong Province", "Farmland Grading in Shandong Province" and "Integrated Remote Sensing Investigation of Shandong Provincial Land Resources", and builds "the Arable Land’s Grading Information System" and "Software System of Arable Land Selected Basic Farmland and Decision-making", and the systems attract the experts ’attention and favourable comment.更多还原