考虑卡尔曼滤波的地理空间数据局部特征分布

Local Feature Distribution of Geospatial Data Considering Kalman Filter

为有效了解城市地理规划发展规律,研究考虑卡尔曼滤波的地理空间数据局部特征分布情况。依据地理空间数据频繁移动特性,把滑动窗口技术引入卡尔曼滤波的采样与数据清洗中,将探测器标签作为随机时间,在滑动窗口中读取标签数量,并满足伯努利二项分布条件,获得原始观测数据最优估计,保障特征分布结果可靠性;利用样本点局部密度与距离两个变量明确聚类中心,给各簇集加入平均密度上限,引入时间约束,修正样本属性值计算方式,完成城市地理空间数据精准特征分布类型与来源划分目标。实验结果表明,所提方法获取的地理空间数据局部特征分布情况与城市实际发展策略相符,更贴近一致性标准贴合度,所得结果具有可靠性,为城市空间布局合理规划发挥应有作用。

In order to effectively understand the development law of urban geographic planning, the local feature distribution of geospatial data considering Kalman filter is studied in this paper. According to the frequent movement characteristics of geospatial data, the sliding window technology is introduced into the sampling and data cleaning of Kalman filter. The detector label is taken as random time, and the number of labels is read in the sliding window. On condition that the Bernoulli binomial distribution is satisfied, the optimal estimation of the original observation data is obtained, and the reliability of the result of feature distribution is guaranteed. The cluster center is defined by two variables, which are local density of sample points and distance. The upper limit of average density is added to each cluster, the time constraint is introduced, and the calculation method of sample attribute value is modified. The experimental results show that the local feature distribution of geospatial data obtained by the proposed method is consistent with the actual urban development strategy and is closer to the fit degree of the consistency standard. The results are reliable which will play a role in the reasonable planning of urban spatial layout.