欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2019, Vol. 36 ›› Issue (3): 354-362.DOI: 10.7523/j.issn.2095-6134.2019.03.009

• 环境科学与地理学 • 上一篇    下一篇

结合质量控制的震后房屋倒塌众包评估模型

贾莹玉1,2, 刘士彬2, 段建波2, 谢帅2   

  1. 1. 中国科学院大学, 北京 100049;
    2. 中国科学院遥感与数字地球研究所, 北京 100094
  • 收稿日期:2018-02-09 修回日期:2018-04-20 发布日期:2019-05-15
  • 通讯作者: 段建波
  • 基金资助:
    中国科学院遥感与数字地球研究所所长基金(ZZCEODE2015HT013)和中国科学院遥感与数字地球研究所数据管理部资助

Crowdsourcing assessment model combined with quality control for collapsed buildings after the earthquake

JIA Yingyu1,2, LIU Shibin2, DUAN Jianbo2, XIE Shuai2   

  1. 1. University of Chinese Academy of Sciences, Beijing 100049, China;
    2. Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094, China
  • Received:2018-02-09 Revised:2018-04-20 Published:2019-05-15

摘要: 在遥感灾害评估领域,众包可用于集合网络大众的智慧完成灾情评估任务,但在现实的众包平台中存在着工作者的答案质量不够可靠的问题。提出一种新型的结合质量控制的震后房屋倒塌众包评估模型。以玉树地震为例搭建众包平台,对收集到的数据,首先使用黄金数据和一致性检验方法过滤掉低质量的数据,随后根据提出的房屋倒塌众包模型得到可靠的灾情解译结果。实验结果显示,质量控制前后,试验区房屋倒塌评估精度分别为70.59%和85.29%,表明本文提出的评估模型有效地提高了众包灾情解译的准确度,可在震后初期提供可信度较高的遥感目视解译结果。

关键词: 众包, 质量控制, 遥感灾害评估, 建筑物倒塌评估

Abstract: In the field of remote sensing disaster assessment, crowdsourcing assembles the intelligence of the network users to complete the disaster assessment tasks. However, there is a problem that the quality of the workers' answers is not reliable in the real crowdsourcing platform. In this study, a new model combined with quality control for crowdsourcing assessment of collapsed buildings after the earthquake is proposed. A case study of Yushu earthquake was presented and the experimental platform was set up to collect crowdsourcing disaster assessment data. First, the gold data and consistency test method were used to filter out low-quality crowdsourcing data. Then, the reliable disaster interpretation results were derived by the crowdsourcing assessment model. The experimental results show that the assessment accuracies of the collapsed buildings in the test area were 70.59% without the quality control and 85.29% with the quality control, which indicates that the proposed model effectively increases the accuracy of the crowdsourcing disaster interpretation. Therefore, crowdsourcing provides reliable visual interpretation results based on the proposed method, which is of importance for remote sensing disaster assessment.

Key words: crowdsourcing, quality control, remote sensing disaster assessment, building collapse assessment

中图分类号: