基于自组织网络多源异构信息场演化融合的贵州山体蠕滑-拉裂型地质滑坡监测及预警应用研究

According to the missing of landslide evolution model in researching landslide monitoring and pre-warning control system in the mountain area of Guizhou Province. There are following problems: the existing monitoring system of low efficiency of data transmission and data processing means of a single problem; automatic landslide monitoring and pre-warning ability low, intends to use the method of ad hoc network heterogeneous information fusion based on field evolution in Guizhou mountain area, the creep crack type the geological model of landslide monitoring and pre-warning research: first, the deployment of self-organizing network through the sensor to different sources, different structure, the typical Guizhou mountain creep crack geological monitoring, seeking information fusion algorithm is efficient, multi field information to determine the characteristics of Guizhou Landslide Evolution; secondly, the establishment of creep mutual response association rules between the crack and deformation of landslide outside the main controlling factors of the construction of landslide deformation field threshold judgment According to the established model, landslide pre-warning and effective; finally, select the Guizhou Bijie disaster area typical creep crack landslide example, through the ad hoc network technology research and development of landslide disaster monitoring and pre-warning system prototype, improve Guizhou mountain creep crack type geological model of landslide disaster prediction accuracy and time efficiency. The project is expected to promote the disaster reduction and disaster prevention in Guizhou Province, and provide theoretical support for improving the ability of automatic monitoring and pre-warning of disasters in Guizhou Province.

针对贵州省山区滑坡监测预警防治研究中存在的滑坡渐进演化模型缺失；现有监测系统数据传输效率低、数据处理手段单一；滑坡自动化监测和预警能力低等问题，拟采用基于自组网多源异构信息场演化融合的方法，针对贵州山体的蠕滑-拉裂型地质模式滑坡监测预警开展研究：首先，通过对不同来源、异种结构的传感器进行自组织组网部署，对典型贵州山区蠕滑-拉裂式地质地貌进行监测，寻求高效的信息融合算法，以确定贵州山体滑坡演化的多场信息特征；其次，建立蠕滑-拉裂式滑坡外在主控因素与变形之间的互响应关联规则，构建滑坡变形演化多场阀值判据，建立有效的滑坡预警模型；最后，选取贵州毕节灾区等典型的蠕滑-拉裂式滑坡实例，通过自组网技术研发滑坡灾害监测预警系统原型，提高贵州山体蠕滑-拉裂型地质模式滑坡灾害预报的准确性和时间效率。项目预期能对贵州省减灾防灾和滑坡灾害治理工作起到推动作用，对提高贵州省灾害自动化监测和预警能力，提供理论支持。

本项目将自组网多源信息融合理论首次与贵州省山体滑坡监测预警系统相结合，建立了适用于贵州山体地质地貌的滑坡演化模型与监测预警系统，在理论方法和关键技术方面取得了如下创新成果：.（1）采用机器学习和数据挖掘技术，对贵州地质滑坡的敏感性因素进行了建模分析，绘制了地质滑坡及滑坡演化敏感性评价地图，解决了滑坡预警监测因子如何选的问题；.（2）采用多源信息融合理论，建立了适用于贵州山区蠕滑-拉裂式地质地貌的滑坡预警的数据级、特征集、决策级三级融合方法，从时间空间融合的角度，解决了贵州山体滑坡灾害如何预测的问题.（3）研究并测试了多源异构灾害监测无线传感器网络在贵州山地的部署、覆盖、传输、通信等问题，选取贵州毕节大方县典型地貌，开发了原型示范系统，确定了高程、坡度、地层岩性、降雨量等敏感评价因子作为监测对象，在原型系统上进行了验证测试，并给出了滑坡预防等级，解决了贵州地质滑坡监测预警系统如何实施的问题。.项目研究目标的实现，验证了该方法能够提高贵州山体蠕滑-拉裂型地质模式滑坡灾害决策的准确性和时间效率，后期将通过与相关部门合作，进一步推广项目研究成果，以持续提高贵州省乃至周边省份滑坡自动化监测和滑坡预警能力。