许强. 对地质灾害隐患早期识别相关问题的认识与思考[J]. 武汉大学学报 ( 信息科学版), 2020, 45(11): 1651-1659. DOI: 10.13203/j.whugis20200043
引用本文: 许强. 对地质灾害隐患早期识别相关问题的认识与思考[J]. 武汉大学学报 ( 信息科学版), 2020, 45(11): 1651-1659. DOI: 10.13203/j.whugis20200043
XU Qiang. Understanding and Consideration of Related Issues in Early Identification of Potential Geohazards[J]. Geomatics and Information Science of Wuhan University, 2020, 45(11): 1651-1659. DOI: 10.13203/j.whugis20200043
Citation: XU Qiang. Understanding and Consideration of Related Issues in Early Identification of Potential Geohazards[J]. Geomatics and Information Science of Wuhan University, 2020, 45(11): 1651-1659. DOI: 10.13203/j.whugis20200043

对地质灾害隐患早期识别相关问题的认识与思考

Understanding and Consideration of Related Issues in Early Identification of Potential Geohazards

  • 摘要: 中国地质灾害点多面广,目前通过人工排查已发现近30万处隐患点,但近年来发生的多起重大地质灾害并不在已发现的隐患点范围内,应该还有大量的灾害隐患没被发现,尽可能全面识别和发现灾害隐患仍是中国防灾减灾最重要的工作内容之一。就如何进一步推动地质灾害隐患早期识别工作提出了自己的认识和建议:(1)近年来,各种遥感技术在地质灾害隐患识别中发挥了重要作用,但每种技术都有各自的长处和短处,所能识别的隐患类型和特征也不尽相同,只有将各种技术手段综合应用,相互补充和校验,才能最大限度地识别已存在的地质灾害隐患,有效破解隐患识别难题。(2)对于识别难度最大的不稳定斜坡,需要将传统地质勘测与现代技术激光雷达(light detection and ranging, LiDAR)、航空或半航空物探等有机结合,才能提升识别效率和准确性。(3)利用深度机器学习可望实现地质灾害隐患的智能化自动识别,但目前其仅对光谱和纹理特性显著的新生地质灾害具有较好的自动识别能力,而对其他类型如古老滑坡体、一般地质灾害隐患点而言,自动识别的正确率还不高,应加大力度开展相关方面的深入研究。

     

    Abstract:
      Objectives  In China, geohazards are wide-ranging. Traditional artificial investigations have found nearly three hundred thousand locations of potential geohazards. However, the recent occurred catastrophic geohazards are not within these determined locations. Widely identification of potential geohazards become one of the most important jobs for geohazard prevention and mitigation.
      Methods  We propose some suggestions to promote early identification for potential geohazards.
      Results  (1) Recently, various remote sensing techniques play an significant role in geohazard identification, but each technique has its limitation to recognize geohazards with different types and characteristics. Only integrated technologies, mutual complementation and verification, can effectively solve the problem. (2) Combination between traditional geological surveys and modern technologies (LiDAR, aerial and semi-aerial geophysical exploration, etc.) can improve the efficiency and accuracy for identification of the most difficult and unstable slops.(3) The deep machine learning is expected to realize the intelligent automatic identification of geohazards. Currently, it shows good performance in new geohazards with significant spectral and texture characteristics, while the accuracy of automatic identification for other types, such as ancient landslides and normal potential geohazards, is still not enough.
      Conclusions  More efforts are in urgent need for further research in related fields.

     

/

返回文章
返回