摘要
针对传统方法预测结果与实际路网径向方位的平均偏差较大的问题,设计基于遥感数据的城乡路网空间演化预测方法。根据主要道路类别的路幅宽度,采用单边缘响应边缘检测方法,提取遥感影像边缘数据,并引入尺度因子在最佳尺度分割下获取路网特征。基于虚拟路线的时间阶与空间阶,构建空间演化预测基本模型,利用时空自相关函数及其偏相关函数,识别模型空间与时间自回归阶数以及空间与时间移动平均阶数,结合极大似然估计法估计的相关参数值,使预测模型性能达到最佳状态。仿真环节中,采用world view 3第四代高分辨率光学卫星,采集目标区域的路网遥感影像,结果表明所提方法能够成功完成城乡路网的空间演化预测,且精准度较高,具有一定的有效性与应用价值。
Aiming at the problem that the average deviation between the prediction results of traditional methods and the radial orientation of the actual road network is large, a spatial evolution prediction method of urban and rural road network based on remote sensing data was designed. According to the road width of the main road categories, the single edge response edge detection method was adopted to extract the edge data of remote sensing images. Under the optimal scale segmentation, the scale factor was introduced to obtain the characteristics of the road network. The basic model of spatial evolution prediction was constructed via the results of time order and space order of virtual route. The spatiotemporal autocorrelation function and its partial correlation function were applied to identify the spatial and temporal autoregressive order and the average order of spatial and temporal movement of the model. Based on the maximum likelihood estimation method, the relevant parameters were estimated to optimize the performance of the prediction model. In the simulation experiment, the fourth generation high-resolution optical satellite of world view 3 was introduced to collect the road network remote sensing image of the target area. The results show that this method can complete the spatial evolution prediction of urban and rural road network with high accuracy.
作者
于宁
卢海军
邓琳
乔雪
YU Ning;LU Hai-jun;DENG Lin;QIAO Xue(School of Architecture and Civil Engineering,Qiqihar University,Qiqhar Heilongjiang 161006,China)
出处
《计算机仿真》
北大核心
2022年第4期110-113,共4页
Computer Simulation
基金
黑龙江省省属本科高校基本科研业务费青年创新人才项目(145109236)
齐齐哈尔市哲学社会科学规划项目研究成果(QSX2021-12YB)。
关键词
遥感数据
城乡路网
空间演化预测
特征提取
虚拟路线
Remote sensing data
Urban and rural road network
Spatial evolution prediction
Feature extraction
Virtual route
