计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (11A): 221100060-6.doi: 10.11896/jsjkx.221100060

• 交叉&应用 • 上一篇    下一篇

基于统一注意力融合网络的耕地变化检测

李滔1, 王海瑞1, 朱贵富2   

  1. 1 昆明理工大学信息工程与自动化学院 昆明 650500
    2 昆明理工大学信息化建设管理中心 昆明 650500
  • 发布日期:2023-11-09
  • 通讯作者: 朱贵富(zhuguifu@kust.edu.cn)
  • 作者简介:(3233530796@qq.com)
  • 基金资助:
    国家自然科学基金(61863016,61263023)

Detection of Farmland Change Based on Unified Attention Fusion Network

LI Tao1, WANG Hairui1, ZHU Guifu 2   

  1. 1 Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China
    2 Information Construction Management Center,Kunming University of Science and Technology,Kunming 650500,China
  • Published:2023-11-09
  • About author:LI Tao,born in 1999,postgraduate.His main research interest is remote sensing image processing.
    ZHU Guifu,born in 1984,senior engineer.His main research interests include education big data,machine lear-ning,intelligent technology,etc.
  • Supported by:
    National Natural Science Foundation of China(61863016,61263023).

PDF (PC)

146

摘要： 为了快速摸清农村乱占耕地建房底数,实现对侵占耕地房屋的检测,提出了一种统一注意力融合网络(Unified Attention Fusion Network)用于农村占用耕地建房识别。为了解决不同时相遥感影像特征相互影响的问题,首先使用孪生网络代替VGG16网络进行特征提取。其次,为了在增大网络感受野并获取更多多尺度信息的前提下减小网络模型大小,在编码阶段最底层使用了简易金字塔池化(Simple Pyramid Pooling Module,SPPM);在解码阶段,为了提高分割精度,突出有用特征,提高边缘分割精度,使用统一注意力融合模块(Unified Attention Fusion Module,UAFM)替换原始的上采样部分进行解码,获取变化检测结果。网络在占用耕地建房数据集上进行了训练和测试。实验结果表明,统一注意力融合网络在测试集上准确率(Accuracy)达到98.82%、精确率(Precision)达到89.69%、召回率(Recall)达到82.14%、F1分数(F1 Score)达到85.74%,能够快速识别不同尺度的疑似占用耕地的违建房屋,为农村乱占耕地建房整治工作提供一种技术检测方法。

关键词: 遥感影像, 变化检测, 建筑物检测, 统一注意力融合网络, 简易金字塔池化

Abstract: In order to quickly find out the number of houses built on arable land illegally occupied and realize the detection of houses built on encroached farmland,a unified attention fusion network is proposed to identify houses built on encroached farmland.In order to solve the problem of mutual influence of remote sensing image features in different phases,the network firstly uses siamese network instead of VGG16 network for feature extraction.Secondly,in order to reduce the size of network model on the premise of increasing the receptive field of network and obtaining more multi-scale information,simple pyramid pooling module is used at the bottom layer of coding stage.In order to improve the segmentation accuracy,highlight the useful features and use the unified attention fusion module to replace the original upsampling part for decoding to obtain the change detection results.The network is trained and tested on the data set of building houses on encroached farmland.Experimental results show that the unified attention fusion network has an accuracy rate of 98.82%,precision of 89.69%,recall rate of 82.14%,and F1 score of 85.74% on the test set.It can quickly identify illegal houses suspected of occupying farmland at different scales,and provide a technical detection method for the construction of houses in rural areas.

Key words: Remote sensing image, Change detection;building inspection, Unified attention fusion network, Simple pyramid pooling

中图分类号: 

  • P237
[1]GAO M,ZHOU X X,LIU Q,et al.Automatic monitoring of farmland occupation by farm house based on deep learning network[J].Bulletin of Surveying and Mapping,2022(3):47-53.
[2]ZHANG Q R,ZHAO L J,LIN G J,et al.Ecological environment assessment of three-river confluence in Yibin city using improved remote sensing ecological index[J].Remote Sensing for Natural Resources,2022,34(1):230-237.
[3]YOUSSEF S,SHAFAEY M A,SALEM A M.Egypt’s remote sensing land use classification using deep learning[J].Space Fostering African Societies,2021(5):55-67.
[4]HUANG S F,MA J W,SUN Y Y.Situation and prospect of flood disaster monitoring by remote sensing in china[J].China Water Resources,2021(15):15-17.
[5]WANG Y Q,WANG C S,WANG L H,et al.Preliminary application of night light remote sensing based on passenger aircraft in Hong Kong economic activity zone changes identification[J].Journal of Geo-information Science,2020,22(5):1153-1160.
[6]CHEN H,QI Z,SHI Z.Remote sensing image change detectionwith transformers[J].IEEE Transactions on Geoscience and Remote Sensing,2022(60):1-14.
[7]TANG P Q,YANG J Y,ZHANG C,et al.An object-oriented post-classification remote sensing change detection after the pi-xel ratio[J].Remote Sensing Information,2010,1:71-74.
[8]FENG Z Y,WANG S D.Remote sensing image change detection based on MNF transform and Canny operator[J].Computer Engineering and Applications,2019,55(14):266-270.
[9]HAO M,ZHANG H,SHI W,et al.Unsupervised change detection using Fuzzy-Means and MRF from remotely sensed image[J].Remote Sensing Letters,2013,4(12):1185-1194.
[10]GAO P X.Change detection of remote sensing images based on variable weight markov random field[J].Computer Applications and Software,2021,38(5):208-212,262.
[11]DENG X L,WANG X,AI S M,et al.Adaptive change detection algorithm of remote sensing images based on SVD and OTSU[J].Science of Surveying and Mapping,2016,41(8):38-42.
[12]YUAN J.Automatic building extraction in aerial scens usingconvolutional networks[J].arXiv,1602.06564,2016.
[13]ZHANG A,LIU X,GROS A,et al.Building detection from satel-lite images on global scale[J].arXiv:1707.;08952,2017.
[14]CHE Z J,GAO F,WU Z F,et al.Multi-source remote sensing image flood disaster information extraction and change analysis based on improved U-Net network[J].Bulletin of Surveying and Mapping,2022(1):26-32.
[15]PAN J P,LI X,SUN B W,et al.Detection of new construction land change based on attention intensive connection pyramid network [J].Bulletin of Surveying and Mapping,2022(3):41-46,59.
[16]ZHENG Z,WANG Y,ZHANG Y J,et al.CLNet:Cross-layer convolutional neural network for change detection in optical remote sensing imagery[J].ISPRS Journal of Photogrammetry and Remote Sensing,2021(175):247-267.
[17]ZHAO X,WANG T,ZHANG Y,et al.Remote sensing imagechange detection based on improved DeepLabv3＋siamese network[J].Journal of Geo-information Science,2022,24(8):1604-1616.
[18]XIANG Y,ZHAO Y D,DONG J H.Remote sensing image mi-ning area change detection based on improved UNet siamese network[J].Journal of China Coal Society,2019,44(12):3773-3780.
[19]PENG J C,LIU Y,TANG S Y,et al.PP-Liteseg:a superior real-time semantic segmentation module[J].arXiv:2204.02681,2022.
[20]CHANG Z L,YANG X G,LU R T,et al.High- Resolution remote sensing image change detection based on improved Deeplabv3+[J].Laser& Optoelectronics Progress,2022,59(12):493-504.
[21]ZHU J Z,CHENG Y,KE F Y,et al.Building change detection from high-resolution remote sensing imagery based on Siam-UNet++[J].Application Research from Computers,2021,38(11):3460-3465.
[22]LI C Y,KONG X B,YANG N,et al.A U-Net based land use/cover change detection method with high resolution image[J].Bulletin of Soil and Water Conservation,2021,41(4):133-138,144.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发