基于目标检测算法YOLOv 9的滑坡隐患识别——以永新县为例

doi:10.13474/j.cnki.11-2246.2025.0607

测绘通报 ›› 2025, Vol. 0 ›› Issue (6): 37-42,102.doi: 10.13474/j.cnki.11-2246.2025.0607

• 学术研究 • 上一篇

基于目标检测算法YOLOv 9的滑坡隐患识别——以永新县为例

涂梨平^1,2, 陈美球¹, 冷鹏³

1. 江西农业大学国土资源与环境学院, 江西南昌 330045;
2. 江西省核工业地质调查院, 江西南昌 330038;
3. 江西核工业测绘院集团有限公司, 江西南昌 330038

收稿日期:2025-01-02 发布日期:2025-07-04
通讯作者: 陈美球。E-mail:cmq12@263.net
作者简介:涂梨平(1981—),女,硕士,正高级工程师,主要从事遥感技术、地质灾害监测预警和国土空间规划等相关工作。E-mail:21924581@qq.com
基金资助:
国家自然科学基金(42461041)

Landslide hazard identification based on the object detection algorithm YOLOv9:taking Yongxin county as an example

TU Liping^1,2, CHEN Meiqiu¹, LENG Peng³

1. School of Land Resources and Environment, Jiangxi Agricultural University, Nanchang 330045, China;
2. Jiangxi Nuclear Industry Geology Survey Institute, Nanchang 330038, China;
3. Jiangxi Nuclear Industry Surveying and Mapping Institute Group Co., Ltd., Nanchang 330038, China

Received:2025-01-02 Published:2025-07-04

摘要/Abstract

摘要： 滑坡灾害是最为严重的地质灾害之一,每年因滑坡灾害造成的财产损失与人员伤亡巨大,传统的基于影像人工排查工作量大、效率低。本文以永新县为研究区,首先基于高分辨率航空影像构建的207个滑坡样本,采用YOLOv 9目标检测算法构建滑坡识别模型,然后对模型精度进行评价,最后识别全县滑坡,并对识别的滑坡结果进行分析。结果表明,模型的精度为0.98,召回率为0.97,mAP为0.95;全县共识别滑坡312处(模型误判46处),经人工内业比对和外业调查验证,模型识别准确率为85.26%。研究表明,基于目标检测算法YOLOv9能有效识别南方地区滑坡,为大范围识别南方小规模滑坡提供了一种有效解决方法。

关键词: 滑坡, 目标检测, 遥感, YOLOv9, 自动识别

Abstract: Landslide disaster is one of the most serious geological disasters,which causes huge property losses and casualties every year.Traditional image-based manual investigation is heavy in workload and low in efficiency.This study takes Yongxin county as the research area,firstly,uses the YOLOv9 object detection algorithm to build a landslide recognition model based on 207 landslide samples constructed by high-resolution aerial images,and then evaluates the accuracy of the model.Finally,the landslide of the whole county is identified and the landslide results identified are analyzed.The results show that the accuracy of the model is 0.98,the recall rate is 0.97,and the mAP is 0.95.There are 312 common landslides in the county,and 46 are misjudged by the model through comparison and field investigation,and the accuracy of model recognition is 85.26%.It can be seen that YOLOv9,an object detection algorithm,can effectively identify landslides in the southern region,providing an effective solution for large-scale identification of small-scale landslides in the south.

Key words: landslide, target detection, remote sensing, YOLOv9, automatic recognition

中图分类号:

P237

涂梨平, 陈美球, 冷鹏. 基于目标检测算法YOLOv 9的滑坡隐患识别——以永新县为例[J]. 测绘通报, 2025, 0(6): 37-42,102.

TU Liping, CHEN Meiqiu, LENG Peng. Landslide hazard identification based on the object detection algorithm YOLOv9:taking Yongxin county as an example[J]. Bulletin of Surveying and Mapping, 2025, 0(6): 37-42,102.

参考文献

[1] 周晓亭.基于多源数据的滑坡识别及其易发性动态评价[D].抚州:东华理工大学,2022.
[2] 魏丽.面向村镇的泥石流灾害风险定量评估研究[D].成都:中国科学院大学(中国科学院水利部成都山地灾害与环境研究所),2020.
[3] 李永鑫,王德富,马志刚,等.知识图谱驱动下的多源遥感滑坡隐患识别[J].测绘通报,2024(1):12-18.
[4] 丁永辉,张勤,杨成生,等.基于高分遥感的金沙江流域滑坡识别:以巴塘县王大龙村为例[J].测绘通报,2022(4):51-55.
[5] 简小婷,赵康,左小清,等.基于Faster R-CNN目标检测的滑坡隐患识别——以福贡县城区为例[J].化工矿物与加工,2022,51(12):19-24.
[6] FENG Q,XU X,WANG Z.Deep learning-based small object detection:a survey[J].Mathematical Biosciences & Engineering,2023,20(4):6551-6590.
[7] 梁峰.基于遥感技术与深度学习的四川高陡山区典型地质灾害识别[D].成都:成都理工大学,2021.
[8] LI Zheng,WANG Yongcheng,ZHANG Ning,et al.Deep learning-based object detection techniques for remote sensing images:a survey[J].Remote Sensing,2022,14(10):2385.
[9] 王晶晶.基于深度学习的泸定县滑坡隐患识别与易发性评估方法研究[D].武汉:中国地质大学,2023.
[10] LONG Leijin,HE Feng,LIU Hongjiang.The use of remote sensing satellite using deep learning in emergency monitoring of high-level landslides disaster in Jinsha River[J].The Journal of Supercomputing,2021,77(8):8728-8744.
[11] ZHANG Pengfei,XU Chong,MA Siyuan,et al.Automatic extraction of seismic landslides in large areas with complex environments based on deep learning:an example of the 2018 iburi earthquake,Japan[J].Remote Sensing,2020,12(23):3992.
[12] LI Huajin,HE Yusen,XU Qiang,et al.Detection and segmentation of loess landslides via satellite images:a two-phase framework[J].Landslides,2022,19(3):673-686.
[13] YU Zhengbo,CHANG Ruichun,CHEN Zhe.Automatic detection method for loess landslides based on GEE and an improved YOLOX algorithm[J].Remote Sensing,2022,14(18):4599.
[14] LI Yunlong,DING Mingtao,ZHANG Qian,et al.Old landslide detection using optical remote sensing images based on improved YOLOv8[J].Applied Sciences,2024,14(3):1100.
[15] QI Jiangtao,LIU Xiangnan,LIU Kai,et al.An improved YOLOv5 model based on visual attention mechanism:Application to recognition of tomato virus disease[J].Computers and Electronics in Agriculture,2022,194:106780.
[16] WANG G,CHEN Y,AN P,et al.UAV-YOLOv8:a small-object-detection model based on improved YOLOv8 for UAV aerial photography scenarios[J].Sensors (Basel),2023,23(16):7190.
[17] WANG C Y,YEH I H,MARK LIAO H Y.YOLOv9:learning what you want to learn using programmable gradient information[M]//Computer Vision-ECCV 2024.Cham:Springer Nature Switzerland,2024:1-21.
[18] HOU Heyi,CHEN Mingxia,TIE Yongbo,et al.A universal landslide detection method in optical remote sensing images based on improved YOLOX[J].Remote Sensing,2022,14(19):4939.

基于目标检测算法YOLOv 9的滑坡隐患识别——以永新县为例

Landslide hazard identification based on the object detection algorithm YOLOv9:taking Yongxin county as an example

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	刘甜甜, 张明, 冯继锋, 李倩楠, 朱雨芯. 基于GF-2影像的农村黑臭水体遥感识别与动态监测[J]. 测绘通报, 2025, 0(5): 8-14.
[2]	王珍香, 肖海, 全思湘, 陈铸, 张泰. 长株潭绿心中央公园核心区林地生态系统碳储量估算与分析[J]. 测绘通报, 2025, 0(5): 15-20.
[3]	李娟, 蔡哲理, 向娟, 袁方芳, 袁光碧, 雷邦俊. 基于中高分辨率遥感影像的植被综合盖度时空演变分析[J]. 测绘通报, 2025, 0(5): 21-26.
[4]	罗建强, 贺秋华, 夏乐, 梅金华, 黄杰, 余姝辰. 东洞庭湖藕池河东支三角洲冲淤时空分析[J]. 测绘通报, 2025, 0(5): 34-39.
[5]	张锋, 刘锋, 佘延娣, 赵媛媛, 马涛. 融合多源遥感数据识别青藏高原草地生产力恢复机制[J]. 测绘通报, 2025, 0(5): 40-46,93.
[6]	王靖凯, 葛星彤, 李兆博, 丁翔, 彭玲. 基于改进HRNet的高速公路路域内光伏板信息提取[J]. 测绘通报, 2025, 0(5): 74-78,99.
[7]	季辰, 刘磊, 蔡栋, 程亮. 街景影像交通标志牌限速限高信息提取[J]. 测绘通报, 2025, 0(4): 39-44.
[8]	尚依炜, 熊俊楠, 贾倩, 罗思远, 王启盛, 曹依帆. 融合升降轨SBAS-InSAR技术的冕宁县滑坡隐患识别与分析[J]. 测绘通报, 2025, 0(4): 58-62.
[9]	王克晓, 李波. 基于粒子群优化和互信息的特征优化与油菜倒伏识别[J]. 测绘通报, 2025, 0(4): 63-67,74.
[10]	马于博, 张爱国, 王浩宇, 刘帅琪, 靳镜宇, 沈占锋, 李均力. 基于U-ConvHDNet模型的戈壁砾幕层提取[J]. 测绘通报, 2025, 0(4): 68-74.
[11]	杨国柱, 孙诗睿, 田茂杰, 孙华敏, 胡伟, 李俊磊. 基于高分影像和改进YOLOv7模型在输电线路走廊的建筑物识别[J]. 测绘通报, 2025, 0(4): 82-89.
[12]	陈子谋, 王新驰, 郭靖. 黄石市生态环境质量时空变化及驱动因子分析[J]. 测绘通报, 2025, 0(4): 145-151.
[13]	田丹, 白晓, 秦坤. 基于Landsat 8 OLI的2013—2020年祁连山区草地退化时空分布特征[J]. 测绘通报, 2025, 0(4): 152-157.
[14]	郝佳, 姚国英, 周剑, 王斯远, 肖进胜. 基于图像和点云融合的三维小目标检测方法[J]. 测绘通报, 2025, 0(3): 33-38.
[15]	谢巴图, 胡佳睿, 潘俊. 基于双路全局信息优化网络的遥感影像海陆分割算法[J]. 测绘通报, 2025, 0(3): 52-58,86.