首页|基于多特征融合与对象边界联合约束网络的建筑物提取

基于多特征融合与对象边界联合约束网络的建筑物提取

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
针对现有全卷积神经网络因光谱混杂造成建筑物漏检、误检以及边界缺失的问题,设计了 一种基于多特征融合与对象边界联合约束网络的高分辨率遥感影像建筑物提取方法.所提方法基于编解码结构,并在编码阶段末端融入连续空洞空间金字塔模块,以在不损失过多有效信息的前提下进行多尺度特征提取和融合;在解码阶段,通过实现基于对象和边界的多输出融合约束结构,为网络融入更多准确的建筑物特征并细化边界;在编码与解码阶段间的横向跳级连接中引入卷积块注意力机制模块,以增强有效特征.此外,解码阶段的多层级输出结果还被用于构建分段多尺度加权损失函数,实现对网络参数的精细化更新.在WHU和Massachusetts建筑物数据集上进行对比试验分析,其中交并比和F1分数分别达到了 90.44％、94.98％和72.57％、84.10％,且模型的复杂度与效率均优于MFCNN与BRRNet.
Building Extraction Based on Multi-feature Fusion and Object-Boundary Joint Constraint Network
Objectives:Accurately and automatically extracting buildings from high-resolution remote sensing images is of great significance in many aspects,such as urban planning,map data updating,emer-gency response,etc.The problems of missing and wrong detection of buildings and missing boundaries caused by spectrum confusion still exist in the existing full convolution neural networks.Methods:In order to overcome the limitations,a multi-feature fusion and object-boundary joint constraint network is pro-posed based on an encoding and decoding structure.In the encoding stage,the continuous-atrous spatial pyramid module is positioned at the end to extract and combine multi-scale features without sacrificing too much useful information.In the decoding stage,more accurate building features are integrated into the net-work and the boundary is refined by implementing the multi-output fusion constraint structure based on ob-ject and boundary.In the skip connection between the encoding and decoding stages,the convolutional block attention module is introduced to enhance the effective features.Furthermore,the multi-level output results from the decoding stage are used to build a piecewise multi-scale weighted loss function for fine net-work parameter updating.Results:Comparative experimental analysis is performed on the WHU and Mas-sachusetts building datasets.The results show that the building extraction results of the proposed method are close to the ground truth.The quantitative evaluation result is higher than the other five state-of-the-art approaches.Specifically,intersection over union and F1-score on WHU and Massachusetts building datasets reach 90.44％,94.98％,and 72.57％,84.10％,respectively.The proposed model outperforms MFCNN and BRRNet in both complexity and efficiency.Conclusions:The proposed method not only improves the accuracy and integrity of extraction results in spectral obfuscation buildings,but also maintains a good boundary with strong robustness in scale.

building extractionfully convolutional neural networkmulti-scale featuresattention mecha-nismjoint constraint

高贤君、冉树浩、张广斌、杨元维

展开 >

长江大学地球科学学院,湖北 武汉,430100

东华理工大学自然资源部环鄱阳湖区域矿山环境监测与治理重点实验室,江西 南昌,330013

湖南科技大学测绘遥感信息工程湖南省重点实验室,湖南 湘潭,411201

武汉大学水工程科学研究院,湖北 武汉,430072

展开 >

建筑物提取 全卷积神经网络 多尺度特征 注意力机制 联合约束

城市轨道交通数字化建设与测评技术国家工程实验室开放课题基金城市轨道交通数字化建设与测评技术国家工程实验室开放课题基金自然资源部环鄱阳湖区域矿山环境监测与治理重点实验室开放基金湖南科技大学测绘遥感信息工程湖南省重点实验室开放基金

2023ZH012021ZH02MEMI-2021-2022-08E22205

2024

10.13203/j.whugis20210520

武汉大学学报(信息科学版)
武汉大学

武汉大学学报(信息科学版)

CSTPCD北大核心
影响因子：1.072
ISSN：1671-8860
年,卷(期)：2024.49(3)
  • 24