[1]王 颖,吴 旭,冷小鹏,等.基于改进 ResU-Net 的中分辨遥感影像滑坡检测方法[J].计算机技术与发展,2023,33(11):182-188.[doi:10. 3969 / j. issn. 1673-629X. 2023. 11. 027]
　WANG Ying,WU Xu,LENG Xiao-peng,et al.Landslide Detection Method Using Improved ResU-Net of Medium Resolution Remote Sensing Images[J].,2023,33(11):182-188.[doi:10. 3969 / j. issn. 1673-629X. 2023. 11. 027]

点击复制

基于改进 ResU-Net 的中分辨遥感影像滑坡检测方法()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年11期

页码:

182-188

栏目:

人工智能

出版日期:

2023-11-10

文章信息/Info

Title:

Landslide Detection Method Using Improved ResU-Net of Medium Resolution Remote Sensing Images

文章编号:

1673-629X(2023)11-0182-07

作者:

王 颖; 吴 旭; 冷小鹏; 余 戈

成都理工大学 计算机与网络安全学院(牛津布鲁克斯学院),四川 成都 610059

Author(s):

WANG Ying; WU Xu; LENG Xiao-peng; YU Ge

School of Computer Science and Cyber Security ( Oxford Brookes College) ,Chengdu University of Technology,Chengdu 610059,China

关键词:

滑坡检测; 多光谱; 图像语义分割; 注意力机制; ResU-Net

Keywords:

landslide detection; multispectral; image semantic segmentation; attention mechanism; ResU-Net

分类号:

TP753

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 11. 027

摘要:

针对基于中分辨率遥感影像滑坡检测精度低的问题,提出了一种基于注意力机制的改进 ResU-Net 模型,并且基于多光谱遥感影像数据集得出了有益于滑坡检测的多特征模型输入

组合。 本研究所用的原始数据集共 14 个特征,首先剔除无效特征,并加入归一化植被指数和归一化水体指数,生成新数据集。 然后将新数据集应用于改进的 ResU-Net 与 U-Net,　ResU-Net,Attention U-Net,BiSeNet,Semantic FPN,U-Net++的对比实验,结果表明改进的 ResU-Net 在测试集上可获得 76. 91% 的 F1 分数,同时精确率和召回率分别为 77. 34%?

和 76. 49% ,在该任务中优于其他对比模型,且比 ResU-Net 模型的 F1 分数高了 0. 43 百分点,有效提高了中分辨率遥感影像的滑坡检测精度。 最后,再向数据集中依次加入归一化

湿度指数和坡向特征,对比不同特征组合数据集产生的检测效果,结果发现加入坡向特征可最大化提升滑坡检测精度,F1 分数可达 77. 03% 。

Abstract:

Aiming at the problem of low accuracy of landslide detection based on medium resolution remote sensing images,we proposean improved ResU-Net model based on attention mechanism,and multi-feature model input combination that is beneficial to landslidedetection is obtained based on multispectral remote sensing imagery dataset. The original dataset used has a total of 14 features. Firstly,the invalid features are removed,and the normalized difference vegetation index and the normalized difference water index are added togenerate a new dataset. Secondly,The new dataset is applied in the comparative experiments of the improved ResU-Net with U-Net,ResU-Net,Attention U-Net,BiSeNet,　Semantic FPN,U-Net+ +. It is showed that the improved ResU-Net can obtain the F1 score of76. 91% on the test set, while the precision and recall are 77. 34% and 76. 49% , respectively, which are better than that of othercomparison models in this task, and it is 0. 43 percentage points higher than the F1 score of the ResU-Net model, which effectively　improves the landslide detection accuracy based on medium resolution remote sensing images. Finally,the normalized difference moistureindex and aspect features are added to the dataset in turn,and the detection accuracy of different feature combinations is compared. Theresults show that adding aspect features can maximize the accuracy of landslide detection,and the F1 score reaches 77. 03% .

常用功能