基于矢量数据的LUCC广义转移矩阵自动挖掘方法

为了充分利用日益增多的多时态土地利用/覆被矢量数据，依据不同属性全方位挖掘提取其中隐含的要素变化转移信息，进而为土地资源科学管理与统筹规划提供技术支持和决策依据，首先提出了广义转移矩阵的概念，并进行了形式化定义，然后设计了一种基于矢量数据的广义转移矩阵自动计算生成方法，详细阐述了数据预处理、匹配关系建立、要素变化检测、矩阵设置与生成4个实施步骤中涉及的主要问题及解决策略。试验结果表明:所提方法能够大幅提高转移矩阵矢量化计算生成的准确度和灵活性，有效克服人工手动处理计算效率低、周期长、易出错、难检查的不足，以及常用栅格化计算方法精度低、结果单一、不能准确全面反映土地利用/覆盖要素在不同属性上的变化转移情况的局限。

An Automated Mining Approach of Land Use/Cover Change General Transfer Matrix Based on Vector Data

In order to take full advantage of expanding land use/cover vector data, quickly mine the underlying information of features changes, and further provide the technical support and decision-making basis for the scientific management and overall planning of land resource, an accurate approach and efficient tool to compute and generate land use/cover change transfer matrix from multiple attribu-tes or aspects of features are proposed in this paper. Firstly, the formal definition of general transfer matrix (GTM) is given and its basic characteristics and information implications are summarized. Se-condly, the main problems and the corresponding solutions during its computation and generation process are discussed, whose four steps, namely data pre-processing, feature matching, change detection and matrix production are also explained in detail. Finally, the visible software tool developed by the integrated secondary development is described briefly. It is concluded that the proposed approach and the related tool can greatly enhance the accuracy and flexibility of GTM vectorized computation and generation, and effectively overcome the shortcomings of manual processing such as low efficiency, long cycle, error-prone and difficult to check, as well as the limilations of the commonly used grid computing methods such as low accuracy, single results and inaccurate and comprehensive reflection of the change and transfer of land use/cover elements in different attributes.