Grid pattern recognition in road networks using the C4.5 algorithm

Pattern recognition in road networks can be used for different applications, including spatiotemporal data mining, automated map generalization, data matching of different levels of detail, and other important research topics. Grid patterns are a common pattern type. This paper proposes and implements a method for grid pattern recognition based on the idea of mesh classification through a supervised learning process. To train the classifier, training datasets are selected from worldwide city samples with different cultural, historical, and geographical environments. Meshes are subsequently labeled as composing or noncomposing grids by participants in an experiment, and the mesh measures are defined while accounting for the mesh’s individual characteristics and spatial context. The classifier is generated using the C4.5 algorithm. The accuracy of the classifier is evaluated using Kappa statistics and the overall rate of correctness. The average Kappa value is approximately 0.74, which corresponds to a total accuracy of 87.5%. Additionally, the rationality of the classifier is evaluated in an interpretation step. Two other existing grid pattern recognition methods were also tested on the datasets, and comparison results indicate that our approach is effective in identifying grid patterns in road networks.     

卫星重力梯度测量数据的粗差探测

分别讨论了阈值法、Dixon检验法和小波算法应用于卫星重力梯度数据粗差探测的效果。为了克服这3种单一粗差探测方法的不足,提出了卫星重力梯度数据的粗差探测组合方案,即联合Dixon检验和小波算法,以及联合Dixon检验、阈值法和小波算法的组合方案。模拟结果表明,两种组合方案均能有效改善粗差探测的效果。     

Re-parameterization of a quasi-analytical algorithm for colored dissolved organic matter dominant inland waters

Quasi-Analytical Algorithms (QAAs) are based on radiative transfer equations and have been used to derive inherent optical properties (IOPs) from the above surface remote sensing reflectance (R_rs) in aquatic systems in which phytoplankton is the dominant optically active constituents (OACs). However, Colored Dissolved Organic Matter (CDOM) and Non Algal Particles (NAP) can also be dominant OACs in water bodies and till now a QAA has not been parametrized for these aquatic systems. In this study, we compared the performance of three widely used QAAs in two CDOM dominated aquatic systems which were unsuccessful in retrieving the spectral shape of IOPS and produced minimum errors of 350% for the total absorption coefficient (a), 39% for colored dissolved matter absorption coefficient (a_CDM) and 7566.33% for phytoplankton absorption coefficient (a_phy). We re-parameterized a QAA for CDOM dominated (hereafter QAA_CDOM) waters which was able to not only achieve the spectral shape of the OACs absorption coefficients but also brought the error magnitude to a reasonable level. The average errors found for the 400–750 nm range were 30.71 and 14.51 for a, 14.89 and 8.95 for a_CDM and 25.90 and 29.76 for a_phy in Funil and Itumbiara Reservoirs, Brazil respectively. Although QAA_CDOM showed significant promise for retrieving IOPs in CDOM dominated waters, results indicated further tuning is needed in the estimation of a(λ) and a_phy(λ). Successful retrieval of the absorption coefficients by QAA_CDOM would be very useful in monitoring the spatio-temporal variability of IOPS in CDOM dominated waters.     

整体法与序贯法在测量平差中的比较

讨论了整体法与序贯法在测量平差中的应用,并结合实例比较整体法与序贯法在相同观测情况下的平差结果和精度问题。实例表明:在测量平差中,若测量网中数据较多,并且这些数据是在不同时期所得,用序贯法具有较高的处理效率。     

基于遗传算法的RBF神经网络在GPS高程拟合中的应用

本文针对RBF神经网络中隐含层径向基中心值的确定,利用遗传算法对其进行优化,并应用于高程拟合的实验研究中。通过将遗传算法优化的RBF神经网络与K-均值优化的RBF神经网络及标准RBF神经网络进行高程拟合的误差对比分析表明:遗传算法优化的RBF神经网络提高了拟合的稳定度,改善了精度。     

基于多目标进化算法的地图注记自动配置技术的研究

在计算机数字制图过程中,地图注记的配置直接影响数字地图的效果.针对地图注记配置过程中两个或多个注记之间冲突处理的问题,本文参照格式塔原则,从地图注记的易读性、位置优先性、形状相似性、地物关联性四个格式塔因子出发对地图注记的候选位置进行评价;同时根据多目标决策理论,用进化算法模拟地图注记的自动配置过程,实现了地图注记自动配置的全局优化.     

双时间步法在二维浅水方程求解中的应用

针对传统算法效率低的问题,将隐式双时间步法应用于求解二维浅水方程,建立了非结构网格下高效的有限体积模型。在应用双时间步法时,虚拟时间层中的定常问题采用高效的隐式LU-SGS（Lower-Upper Symmetric Gauss-Seidel）方法进行迭代求解。通过模拟计算4个典型算例以及与传统显式算法进行比较,对模型精度、效率及处理实际问题能力进行检验,分析了时间步长、内迭代次数对模型性能的影响。结果表明,双时间步法放宽了稳定性对时间步长的限制,时间步长可取到显式格式10倍以上,计算耗时减少了50%以上,模型具有良好精度与适应性,具有较好的推广应用价值。     

考虑边界反射作用河流污染混合区的简化算法

以排污口为坐标原点,通过数学推证了岸边排放时P(x,y)点与中心排放时Q(x/4,±y/2)点的污染物浓度对应相等.通过计算实验和曲线拟合,分别给出了中宽河流污染混合区最大长度、最大宽度和相应的纵坐标、面积等参数以及河流中心线和两岸线沿程浓度分布的实用化公式,给出了污染混合区的近似外边界曲线方程.对中宽河流污染混合区范围的计算和环境敏感点的浓度预测,具有重要实用价值,为工程技术人员提供了准确、简便、快捷的实用化方法.     

基于改进型前馈神经网络的流域产流预报模型的研究

在分析流域产流机制、影响因素和现行产流计算方法的基础上,首次取前期影响雨量、主产流历时、全过程面平均雨量和４个代表雨强计７个因子作为神经网络输入,直接以流域产流深作为神经网络输出,并针对传统ＢＰ算法的固有缺陷,采用混合ＧＮ－ＢＦＧＳ算法训练网络。实例验证了所建模型及算法的有效性和可行性。还对神经网络隐层单位数等进行了初步研究。     

邻近算法在一维大地电磁反演中的应用

总结了大地电磁法( MT)中常用的各种反演算法,并指出其局限性。对邻近算法( NA)加以概述,并将其引入MT反演中。对一维MT合成数据进行反演分析,得到的最大似然模型十分接近理论模型。虽然NA与遗传算法( GA)抗陷入局部极小值的能力大体相等,但NA算法生成的采样点分布密度与误差函数大体一致,因此NA更有利于应用基于积分的参数估值方法。本研究表明,NA收敛速度比GA更快。这说明在一维MT反演中,NA算法比GA算法更具优势。