Participatory sensing-based geospatial localization of distant objects for disaster preparedness in urban built environments
Participatory sensing-based geospatial localization of distant objects for disaster preparedness in urban built environments
复制标题
DOI：
10.1016/j.autcon.2019.102960
复制
发表时间：
2019-11
影响因子：
10.3
通讯作者：
Hongjo Kim;Youngjib Ham
中科院分区：
文献类型：
--
作者：
Hongjo Kim;Youngjib Ham
虽然参与式遥感在大范围收集当地数据方面的好处早已得到承认，但由于缺乏对遥远物体的地理空间定位能力，它还没有被广泛用于各种应用，如灾害准备。在这类应用中，需要对感兴趣的对象进行强有力的本地化和记录，以支持现场检查和资源调动中的数据驱动决策。然而，参与式传感由于缺乏公民移动设备中的测距传感器，不适合对远处的物体进行定位，因此定位精度有很大的差异。针对这一问题，本研究提出了一种新的基于参与式感知的远距离物体空间定位方法。提出的地理空间定位过程包括多个计算模块--地理坐标转换、均值漂移聚类、基于深度学习的目标检测、磁偏角调整、视线方程公式和Moore-Penrose广义逆方法--以提高参与式传感环境中的定位精度。在休斯顿和德克萨斯州的大学站进行了实验，实验结果表明该方法具有合理的定位精度，当观察者到感兴趣对象的距离为17 m到296 m时，记录的距离误差为1.5 m到27.8 m。该方法有望为大城市地区的快速数据收集做出贡献，从而为需要识别远距离危险对象位置的灾难准备工作提供帮助。
Although the benefit of participatory sensing for collecting local data over large areas has long been recognized, it has not been widely used for various applications such as disaster preparation due to a lack of geospatial localization capability with respect to a distant object. In such applications, objects of interest need to be robustly localized and documented for supporting data-driven decision-making in site inspection and resource mobilization. However, participatory sensing is inappropriate to localize a distant object due to the absence of ranging sensors in citizens' mobile devices; thus, the localization accuracy varies to a large extent. To address this issue, this study presents a novel geospatial localization method for distant objects based on participatory sensing. The proposed geospatial localization process consists of multiple computational modules—a geographic coordinate conversion, mean-shift clustering, deep learning-based object detection, magnetic declination adjustment, line of sight equation formulation, and the Moore-Penrose generalized inverse method—to improve the localization accuracy in participatory sensing environments. The experiments were conducted in Houston and College Station in Texas to evaluate the proposed method, and the experimental results demonstrated a reasonable localization accuracy, recording the distance errors of 1.5 m to 27.8 m when the distance from observers to the objects of interest were 17 m to 296 m. The proposed method is expected to contribute to rapid data collection over large urban areas, thereby facilitating disaster preparedness that needs to identify locations of distant objects at risk.