基于改进U-Net的珊瑚礁底栖物质信息提取方法

doi:10.19678/j.issn.1000-3428.0066091

计算机工程 ›› 2023, Vol. 49 ›› Issue (12): 231-242. doi: 10.19678/j.issn.1000-3428.0066091

基于改进U-Net的珊瑚礁底栖物质信息提取方法

傅杨淦¹^,², 朱岚巍¹^,²^,³^,⁴^,*, 吴虹蓉²^,⁵, 陈方¹^,²^,³^,⁴

1. 桂林电子科技大学计算机与信息安全学院, 广西桂林 541000
2. 海南空天信息研究院海南省地球观测重点实验室, 海南三亚 572000
3. 可持续发展大数据国际研究中心, 北京 100094
4. 中国科学院空天信息创新研究院数字地球重点实验室, 北京 100094
5. 长江大学地球科学学院, 武汉 430000

收稿日期:2022-10-24 出版日期:2023-12-15 发布日期:2023-02-08
通讯作者: 朱岚巍
作者简介:
傅杨淦（1997—），男，硕士，主研方向为深度学习、遥感图像处理
吴虹蓉，硕士
陈方，研究员、博士
基金资助:
海南省重点研发计划(ZDYF2020030); 广西创新驱动发展专项“中国-东盟地球大数据平台与应用示范”子课题(桂科AA20302022-3-2); 海南省重大科技计划项目(ZDKJ2019006)

Coral Reef Benthic Material Information Extraction Method Based on Improved U-Net

Yanggan FU¹^,², Lanwei ZHU¹^,²^,³^,⁴^,*, Hongrong WU²^,⁵, Fang CHEN¹^,²^,³^,⁴

1. School of Computer Science and Information Security, Guilin University of Electronic Technology, Guilin 541000, Guangxi, China
2. Hainan Key Laboratory of Earth Observation, Hainan Aerospace Information Research Institute, Sanya 572000, Hainan, China
3. International Research Center of Big Data for Sustainable Development Goals, Beijing 100094, China
4. Key Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
5. School of Earth Sciences, Yangtze University, Wuhan 430000, China

Received:2022-10-24 Online:2023-12-15 Published:2023-02-08
Contact: Lanwei ZHU

摘要/Abstract

摘要：

珊瑚礁底栖物质信息提取在珊瑚礁遥感监测领域具有重要意义。SVM、最大似然法等传统珊瑚礁底栖物质信息提取方法存在精度不高、不够自动化、时间成本较高等问题。目前深度学习方法在语义分割领域已有广泛应用，且取得了较好的效果，为此，利用深度学习技术设计一种基于改进U-Net的分割网络模型，以进行珊瑚礁底栖物质信息提取。为了保留分割细节，对编码器的每个层级设置一种多输入的方式。将ResNet34的残差部分结构作为网络的编码器，以提取更丰富的特征。结合分解卷积、注意力机制和通道混洗操作设计一种新的特征提取块，并将其代替编码器、底层和解码器中的普通卷积层。同时，通过注意力机制来改善U-Net模型的远跳连接，对权重进行调整，以提高分割精度。在三亚地区的GF-2多光谱遥感影像上进行实验，提取的地物类别分别为健康珊瑚礁、白化珊瑚礁、藻类混合物、沙、浪花、深海区和陆地，通过面向对象方法并结合Google Earth影像进行目视解译以修订建立数据集。实验结果表明，该模型的平均交并比和平均F1值分别达到67.17%和78.7%，与常用的分割模型相比，其在视觉效果和评价指标上更优，消融实验结果也验证了改进模块的有效性。

关键词: 珊瑚礁, 遥感影像, 图像分割, 远跳连接, 通道混洗, 注意力机制

Abstract:

The extraction of coral reef benthic material information is of great significance in coral reef remote sensing monitoring. Traditional information extraction methods for benthic organisms on coral reefs, such as Support Vector Machine(SVM) and the maximum likelihood method, have several drawbacks, including poor accuracy, low level of automation, and high time cost. At present, methods based on deep learning are being widely used in image segmentation, and satisfactory results have been obtained. Therefore, a segmentation network model based on improved U-Net and deep learning technology is designed to extract benthic material information from coral reefs. To improve segmentation details, a multi-input mode is set for each level of the encoder. The residual structure of ResNet34 is used as the encoder of the network to extract more abundant features. A new feature extraction block is designed by combining decomposition convolution, attention mechanism, and channel shuffle operation, to replace the common convolution layer in the encoder, bottom layer, and decoder; Concurrently, the attention mechanism is used to improve the far hop connection of the U-Net model, adjust the weight, and improve segmentation accuracy. Based on the GF-2 multispectral remote sensing image of Sanya, the extracted ground objects are classified as healthy coral reef, albino coral reef, algal mixture, sand, spray, deep sea area, and land. The dataset is established through visual interpretation and revision of an object-oriented method combined with Google Earth image. The experimental results show that mean Intersection over Union(mIoU)and average F1 score of the model in this study reached 67.17% and 78.7%, respectively. Compared with commonly used segmentation models, the proposed model performs better in visual effect and evaluation indicators. The ablation experimental results prove the effectiveness of the improved module.

Key words: coral reef, remote sensing image, image segmentation, skip connection, channel shuffle, attention mechanism

傅杨淦, 朱岚巍, 吴虹蓉, 陈方. 基于改进U-Net的珊瑚礁底栖物质信息提取方法[J]. 计算机工程, 2023, 49(12): 231-242.

Yanggan FU, Lanwei ZHU, Hongrong WU, Fang CHEN. Coral Reef Benthic Material Information Extraction Method Based on Improved U-Net[J]. Computer Engineering, 2023, 49(12): 231-242.

https://www.ecice06.com/CN/Y2023/V49/I12/231

图/表 18

图1 一级分割与三级分割对比结果

Fig.1 Comparison results of primary segmentation and tertiary segmentation

图2 2019年GF-2多光谱影像的三级分割结果

Fig.2 Tertiary segmentation results of GF-2 multispectral images in 2019

图3 数据集及对应的标签

Fig.3 Datasets and corresponding labels

图4 改进的U-Net网络模型

Fig.4 Improved U-Net network model

图5 Multiple-input结构

Fig.5 Multiple-input structure

图6 CBAM结构

Fig.6 CBAM structure

图7 FAS块结构

Fig.7 FAS block structure

图8 普通卷积和分解卷积

Fig.8 Ordinary convolution and decomposed convolution

图9 skip-attention结构

Fig.9 skip-attention structure

图10 网络训练过程中的指标变化曲线

Fig.10 Index change curves during network training process

图11 添加不同模块后的分割结果

Fig.11 Segmentation results after adding different modules

图12 使用不同损失函数的分割结果

Fig.12 Segmentation results using different loss functions

图13 不同模型的分割结果对比

Fig.13 Comparison of segmentation results between different models

参考文献 31

1	BRANDL S J, RASHER D B, CÔTÉ I M, et al. Coral reef ecosystem functioning: eight core processes and the role of biodiversity. Frontiers in Ecology and the Environment, 2019, 17(8): 445- 454. doi: 10.1002/fee.2088
2	MOBERG F, RÖNNBÄCK P. Ecosystem services of the tropical seascape: interactions, substitutions and restoration. Ocean & Coastal Management, 2003, 46(1/2): 27- 46.
3	CAPILI E B, IBAY A C S, VILLARIN J R T. Climate change impacts and adaptation on Philippine coasts[C]//Proceedings of OCEANS'05. Washington D. C., USA: IEEE Press, 2006: 2299-2306.
4	LANG M. Smithsonian institution coral reef research[EB/OL]. [2022-09-05]. https://ieeexplore.ieee.org/document/1640206.
5	BELLWOOD D R, HUGHES T P, FOLKE C, et al. Confronting the coral reef crisis. Nature, 2004, 429(6994): 827- 833. doi: 10.1038/nature02691
6	GLYNN P W. Coral reef bleaching: facts, hypotheses and implications. Global Change Biology, 1996, 2(6): 495- 509. doi: 10.1111/j.1365-2486.1996.tb00063.x
7	潘艳丽, 唐丹玲. 卫星遥感珊瑚礁白化概述. 生态学报, 2009, 29(9): 5076- 5080. URL
	PAN Y L, TANG D L. General introduction to satellite remote sensing of coral reef bleaching. Acta Ecologica Sinica, 2009, 29(9): 5076- 5080. URL
8	YANG J, HE Y H, CASPERSEN J. A self-adapted threshold-based region merging method for remote sensing image segmentation[C]//Proceedings of IEEE International Geoscience and Remote Sensing Symposium. Washington D. C., USA: IEEE Press, 2016: 6320-6323.
9	KABIRI K, REZAI H, MORADI M. Mapping of the corals around Hendorabi island (Persian Gulf), using WorldView-2 standard imagery coupled with field observations. Marine Pollution Bulletin, 2018, 129(1): 266- 274. doi: 10.1016/j.marpolbul.2018.02.045
10	李永丽, 王浩, 金喜子. 基于随机森林优化的自组织神经网络算法. 吉林大学学报(理学版), 2021, 59(2): 351- 358. URL
	LI Y L, WANG H, JIN X Z. Self-organizing neural network algorithm based on random forest optimization. Journal of Jilin University(Science Edition), 2021, 59(2): 351- 358. URL
11	胡蕾秋, 刘亚岚, 任玉环, 等. SPOT5多光谱图像对南沙珊瑚礁信息提取方法的探讨. 遥感技术与应用, 2010, 25(4): 493- 501. URL
	HU L Q, LIU Y L, REN Y H, et al. Research on the extraction method of coral reef at spratly islands using SPOT5. Remote Sensing Technology and Application, 2010, 25(4): 493- 501. URL
12	LEON J, WOODROFFE C D. Improving the synoptic mapping of coral reef geomorphology using object-based image analysis. International Journal of Geographical Information Science, 2011, 25(6): 949- 969. doi: 10.1080/13658816.2010.513980
13	SHELHAMER E, LONG J, DARRELL T. Fully convolutional networks for semantic segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(4): 640- 651. doi: 10.1109/TPAMI.2016.2572683
14	RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[EB/OL]. [2022-09-05]. https://arxiv.org/abs/1505.04597.
15	MILLETARI F, NAVAB N, AHMADI S A. V-Net: fully convolutional neural networks for volumetric medical image segmentation[C]//Proceedings of the 4th International Conference on 3D Vision. Washington D. C., USA: IEEE Press, 2016: 565-571.
16	ZHOU Z W, RAHMAN SIDDIQUEE M M, TAJBAKHSH N, et al. UNet++: a nested U-Net architecture for medical image segmentation[EB/OL]. [2022-09-05]. https://arxiv.org/abs/1807.10165.
17	CHEN L C, ZHU Y K, PAPANDREOU G, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation[EB/OL]. [2022-09-05]. https://arxiv.org/abs/1802.02611.
18	LIU Y J, LI X F, REN Y B. A deep learning model for oceanic mesoscale eddy detection based on multi-source remote sensing imagery[C]//Proceedings of IGARSS'20. Washington D. C., USA: IEEE Press, 2021: 6762-6765.
19	GUO M Q, LIU H, XU Y Y, et al. Building extraction based on U-Net with an attention block and multiple losses. Remote Sensing, 2020, 12(9): 1400. doi: 10.3390/rs12091400
20	LU X Y, ZHONG Y F, ZHENG Z, et al. Multi-scale and multi-task deep learning framework for automatic road extraction. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(11): 9362- 9377. doi: 10.1109/TGRS.2019.2926397
21	GOMES R, ROZARIO P, ADHIKARI N. Deep learning optimization in remote sensing image segmentation using dilated convolutions and ShuffleNet[C]//Proceedings of IEEE International Conference on Electro Information Technology. Washington D. C., USA: IEEE Press, 2021: 244-249.
22	何红术, 黄晓霞, 李红旮, 等. 基于改进U-Net网络的高分遥感影像水体提取. 地球信息科学学报, 2020, 22(10): 2010- 2022. URL
	HE H S, HUANG X X, LI H G, et al. Water body extraction of high resolution remote sensing image based on improved U-Net network. Journal of Geo-Information Science, 2020, 22(10): 2010- 2022. URL
23	赫晓慧, 宋定君, 李盼乐, 等. 融合多尺度特征的遥感影像道路提取方法. 计算机工程, 2022, 48(8): 196- 205. URL
	HE X H, SONG D J, LI P L, et al. Remote sensing image road extraction method combined with multi-scale features. Computer Engineering, 2022, 48(8): 196- 205. URL
24	胥智杰, 杨小兵, 何灵敏, 等. 多尺度遥感语义分割网络. 计算机工程与应用, 2020, 56(21): 210- 217. URL
	XU Z J, YANG X B, HE L M, et al. Multiscale remote sensing semantic segmentation network. Computer Engineering and Applications, 2020, 56(21): 210- 217. URL
25	沈骏翱, 马梦婷, 宋致远, 等. 基于深度学习语义分割模型的高分辨率遥感图像水体提取. 自然资源遥感, 2022, 34(4): 129- 135. URL
	SHEN J A, MA M T, SONG Z Y, et al. High resolution remote sensing water classification based on deep learning semantic segmentation model. Remote Sensing for Natural Resources, 2022, 34(4): 129- 135. URL
26	SZEGEDY C, VANHOUCKE V, IOFFE S, et al. Rethinking the inception architecture for computer vision[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2016: 2818-2826.
27	WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]//Proceedings of the 15th European Conference on Computer Vision. New York, USA: ACM Press, 2018: 3-19.
28	ZHANG X Y, ZHOU X Y, LIN M X, et al. ShuffleNet: an extremely efficient convolutional neural network for mobile devices[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2018: 6848-6856.
29	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2016: 770-778.
30	SRIVASTAVA N, HINTON G, KRIZHEVSKY A, et al. Dropout: a simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 2014, 15, 1929- 1958.
31	IOFFE S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift[C]//Proceedings of the 32nd International Conference on Machine Learning. Washington D. C., USA: IEEE Press, 2015: 448-456.

[1]	张天鹏, 韩晶, 吕学强. 基于多任务学习的超分辨率辅助小目标检测[J]. 计算机工程, 2024, 50(9): 304-312.
[2]	郭敏, 张熙涵, 李阳. 融合注意力的教师互一致性半监督医学图像分割[J]. 计算机工程, 2024, 50(9): 313-323.
[3]	曾钰琦, 刘博, 钟柏昌, 钟瑾. 智慧教育下基于改进YOLOv8的学生课堂行为检测算法[J]. 计算机工程, 2024, 50(9): 344-355.
[4]	李俊俊, 董建刚, 李坤. 基于Kubernetes的集群节能策略研究[J]. 计算机工程, 2024, 50(9): 82-91.
[5]	林畅, 郭伟, 任哲聪, 金海波. 基于Transformer的目标跟踪与分割统一算法[J]. 计算机工程, 2024, 50(9): 130-141.
[6]	李泽霖, 吕兆峰, 陈富强, 李克. 基于多跳信息融合的实体对齐模型[J]. 计算机工程, 2024, 50(9): 142-152.
[7]	王汝英, 马嘉骏, 董建强, 刘万龙, 张海涛, 尹凯, 赵博超. 基于MTS-BiGRU-DMHSA的工业负荷预测方法[J]. 计算机工程, 2024, 50(9): 169-178.
[8]	朱凯, 李理, 张彤, 江晟, 别一鸣. 基于Transformer的多阶段运动模糊图像修复网络[J]. 计算机工程, 2024, 50(9): 276-285.
[9]	饶日昕, 王怡文, 曾砺志, 童心恬, 赵海涛. 面向废旧电缆检测的轻量化网络模型[J]. 计算机工程, 2024, 50(8): 22-30.
[10]	李华昱, 张智康, 闫阳, 岳阳. 基于知识图谱增强的领域多模态实体识别[J]. 计算机工程, 2024, 50(8): 31-39.
[11]	王蕾, 党时鹏, 潘丰. 基于卷积神经网络的隐匿性旁路预测模型[J]. 计算机工程, 2024, 50(8): 40-49.
[12]	陈瀚, 赵春蕾, 蒋昊达, 王春东. 基于融合模型与语义网络的App用户意图识别研究[J]. 计算机工程, 2024, 50(8): 50-63.
[13]	高爽, 史轶伦, 徐巧枝, 于磊. 基于对比学习的非对称编解码结构的心脏MRI分割研究[J]. 计算机工程, 2024, 50(8): 290-300.
[14]	王夙喆, 张雪英, 陈晓玉, 李凤莲, 吴泽林. 基于有效注意力和GAN结合的脑卒中EEG增强算法[J]. 计算机工程, 2024, 50(8): 336-344.
[15]	王宇, 祁琦, 王纯, 许才. 储能变流器信号高精度故障诊断方法[J]. 计算机工程, 2024, 50(8): 389-396.

选择文件类型/文献管理软件名称

选择包含的内容

基于改进U-Net的珊瑚礁底栖物质信息提取方法

Coral Reef Benthic Material Information Extraction Method Based on Improved U-Net

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 31

相关文章 15

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于改进U-Net的珊瑚礁底栖物质信息提取方法

Coral Reef Benthic Material Information Extraction Method Based on Improved U-Net

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 31

相关文章 15

编辑推荐

Metrics

本文评价