基于全卷积网络模型的高分遥感影像内陆网箱养殖区提取

doi:10.3976/j.issn.1002-4026.2022.02.001

摘要/Abstract

摘要：

为了研究高分遥感影像的内陆网箱养殖区自动快速提取,利用福建省北部内陆水域的GF-1影像和GF-2影像,并对影像中的网箱养殖区进行人工标注,经过旋转、缩放和镜像翻转等数据增强处理后构建了2种影像的内陆网箱养殖区样本库;利用样本库训练内陆网箱养殖区提取的深度学习全卷积网络(fully convolutional networks,FCN)模型并开展精度验证。结果显示,GF-1影像提取结果的F值达到83.37%,GF-2影像提取结果的F值达到92.56%。表明基于FCN的高分影像内陆网箱养殖区提取具有较高的精度,能够进行大规模内陆网箱养殖区提取应用,为内陆水产养殖区的监测提供重要依据。

关键词: 深度学习, 全卷积网络模型, 数据增强, 高分辨率遥感影像, GF卫星, 内陆网箱养殖区, 养殖区提取

Abstract:

The extraction of cage aquacultural areas was investigated using high-resolution GF-1 and GF-2 remote sensing images from northern Fujian Province. Image enhancement was performed by correction, fusion, and cropping. The sample database of inland cage culture areas of two kinds of images was constructed; The sample bank is used to train the in-depth learning fully convolutional networks (FCN) model extracted from inland cage culture area and verify the accuracy. The results of the test experiment show that the F-measure of GF-1 and GF-2 reaches 83.37% and 92.56%,respectively. It shows that the inland cage culture area extraction based on FCN has high accuracy, and can be used for large-scale inland cage acquaculture area extraction, which provides an important basis for the monitoring of inland aquaculture area.

Key words: deep learning, FCN model, data enhancement, high-resolution remote sensing image, GF satellite, inland cage aquacultural area, aquacultural area extraction

中图分类号:

TP79

李连伟,张源榆,岳增友,薛存金,付宇轩,徐洋峰. 基于全卷积网络模型的高分遥感影像内陆网箱养殖区提取[J]. 山东科学, 2022, 35(2): 1-10.

LI Lian-wei,ZHANG Yuan-yu,YUE Zeng-you,XUE Cun-jin,FU Yu-xuan,XU Yang-feng. Extracting inland cage aquacultural areas from high-resolution remote sensing images using fully convolutional networks model[J]. Shandong Science, 2022, 35(2): 1-10.

图/表 11

图1

表1

图2

图3

图4

图5

表2

表3

图6

图7

表4

参考文献 28

[1]	YUAN J J, XIANG J, LIU D Y, et al. Rapid growth in greenhouse gas emissions from the adoption of industrial-scale aquaculture[J]. Nature Climate Change, 2019, 9(4): 318-322. DOI: 10.1038/s41558-019-0425-9. doi: 10.1038/s41558-019-0425-9
[2]	ALBERT T, MATTHIAS H, SENA S, et al. Cage aquaculture: regional reviews and global overview[R]. Rome, Italy:Food and Agriculture Organization, 2007.
[3]	LUO J H, PU R L, MA R H, et al. Mapping long-term spatiotemporal dynamics of pen aquaculture in a shallow lake: less aquaculture coming along better water quality[J]. Remote Sensing, 2020, 12(11): 1866. DOI: 10.3390/rs12111866. doi: 10.3390/rs12111866
[4]	AL-NASRAWI A, HOPLEY C, HAMYLTON S, et al. A spatio-temporal assessment of landcover and coastal changes at wandandian delta system, southeastern Australia[J]. Journal of Marine Science and Engineering, 2017, 5(4): 55. DOI: 10.3390/jmse5040055. doi: 10.3390/jmse5040055
[5]	WANG M, CUI Q, WANG J, et al. Raft cultivation area extraction from high resolution remote sensing imagery by fusing multi-scale region-line primitive association features[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2017, 123: 104-113. DOI: 10.1016/j.isprsjprs.2016.10.008. doi: 10.1016/j.isprsjprs.2016.10.008
[6]	WANG Z H, YANG X M, LIU Y M, et al. Extraction of coastal raft cultivation area with heterogeneous water background by thresholding object-based visually salient NDVI from high spatial resolution imagery[J]. Remote Sensing Letters, 2018, 9(9): 839-846. DOI: 10.1080/2150704x.2018.1468103. doi: 10.1080/2150704x.2018.1468103
[7]	KANG J M, SUI L C, YANG X M, et al. Sea surface-visible aquaculture spatial-temporal distribution remote sensing: a case study in Liaoning Province, China from 2000 to 2018[J]. Sustainability, 2019, 11(24): 7186. DOI: 10.3390/su11247186. doi: 10.3390/su11247186
[8]	ZENG Z, WANG D, TAN W X, et al. Extracting aquaculture ponds from natural water surfaces around inland lakes on medium resolution multispectral images[J]. International Journal of Applied Earth Observation and Geoinformation, 2019, 80: 13-25. DOI: 10.1016/j.jag.2019.03.019. doi: 10.1016/j.jag.2019.03.019
[9]	OTTINGER M, CLAUSS K, KUENZER C. Large-scale assessment of coastal aquaculture ponds with sentinel-1 time series data[J]. Remote Sensing, 2017, 9(5): 440. DOI: 10.3390/rs9050440. doi: 10.3390/rs9050440
[10]	ZHU M, HE Y N, HE Q Y. A review of researches on deep learning in remote sensing application[J]. International Journal of Geosciences, 2019, 10(1): 1-11. DOI: 10.4236/ijg.2019.101001. doi: 10.4236/ijg.2019.101001
[11]	YANG Z, MU X D, ZHAO F A. Scene classification of remote sensing image based on deep network grading transferring[J]. Optik, 2018, 168: 127-133. DOI: 10.1016/j.ijleo.2018.04.092. doi: 10.1016/j.ijleo.2018.04.092
[12]	SCOTT G J, ENGLAND M R, STARMS W A, et al. Training deep convolutional neural networks for land-cover classification of high-resolution imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(4): 549-553. DOI: 10.1109/LGRS.2017.2657778. doi: 10.1109/LGRS.2017.2657778
[13]	DING P, ZHANG Y, JIA P, et al. A comparison: different DCNN models for intelligent object detection in remote sensing images[J]. Neural Processing Letters, 2019, 49(3): 1369-1379. DOI: 10.1007/s11063-018-9878-5. doi: 10.1007/s11063-018-9878-5
[14]	ZHANG Z, HUANG J, JIANG T, et al. Semantic segmentation of very high-resolution remote sensing image based on multiple band combinations and patchwise scene analysis[J]. Journal of Applied Remote Sensing, 2020, 14: 016502. DOI: 10.1117/1.JRS.14.016502. doi: 10.1117/1.JRS.14.016502
[15]	KEMKER R, SALVAGGIO C, KANAN C. Algorithms for semantic segmentation of multispectral remote sensing imagery using deep learning[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 145: 60-77. DOI: 10.1016/j.isprsjprs.2018.04.014. doi: 10.1016/j.isprsjprs.2018.04.014
[16]	LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation [C]//2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA:IEEE, 2015: 3431-3440. DOI: 10.1109/CVPR.2015.7298965. doi: 10.1109/CVPR.2015.7298965
[17]	RONNEBERGER O, FISCHER P, BROX T. U-net: Convolutional networks for biomedical image segmentation [C]//Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015, 2015: 234-241. DOI: 10.1007/978-3-319-24574-4_28. doi: 10.1007/978-3-319-24574-4_28
[18]	CHEN L C, ZHU Y K, PAPANDREOU G, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation [C]//Computer Vision-ECCV 2018, 2018: 833-851. DOI: 10.1007/978-3-030-01234-2_49. doi: 10.1007/978-3-030-01234-2_49
[19]	SHI T Y, XU Q Z, ZOU Z X, et al. Automatic raft labeling for remote sensing images via dual-scale homogeneous convolutional neural network[J]. Remote Sensing, 2018, 10(7): 1130. DOI: 10.3390/rs10071130. doi: 10.3390/rs10071130
[20]	CUI B G, FEI D, SHAO G H, et al. Extracting raft aquaculture areas from remote sensing images via an improved U-net with a PSE structure[J]. Remote Sensing, 2019, 11(17): 2053. DOI: 10.3390/rs11172053. doi: 10.3390/rs11172053
[21]	SUI B K, JIANG T, ZHANG Z, et al. A modeling method for automatic extraction of offshore aquaculture zones based on semantic segmentation[J]. ISPRS International Journal of Geo-Information, 2020, 9(3): 145. DOI: 10.3390/ijgi9030145. doi: 10.3390/ijgi9030145
[22]	ZENG Z, WANG D, TAN W X, et al. RCSANet: A full convolutional network for extracting inland aquaculture ponds from high-spatial-resolution images[J]. Remote Sensing, 2020, 13(1): 92. DOI: 10.3390/rs13010092. doi: 10.3390/rs13010092
[23]	农业农村部渔业渔政管理局. 中国渔业年鉴2018[M]. 北京: 中国农业出版社, 2019: 169-300.
[24]	林晓平. 福建省闽江水口库区古田段渔业生产SWOT分析及发展对策[J]. 水产养殖, 2015, 36(8): 28-30. DOI: 10.3969/j.issn.1004-2091.2015.08.011. doi: 10.3969/j.issn.1004-2091.2015.08.011
[25]	白照广. 高分一号卫星的技术特点[J]. 中国航天, 2013(8): 5-9.
[26]	潘腾. 高分二号卫星的技术特点[J]. 中国航天, 2015(1): 3-9.
[27]	孙伟伟, 杨刚, 陈超, 等. 中国地球观测遥感卫星发展现状及文献分析[J]. 遥感学报, 2020, 24(5): 479-510.
[28]	LIU Y M, WANG Z H, YANG X M, et al. Satellite-based monitoring and statistics for raft and cage aquaculture in China's offshore waters[J]. International Journal of Applied Earth Observation and Geoinformation, 2020, 91: 102118. DOI: 10.1016/j.jag.2020.102118. doi: 10.1016/j.jag.2020.102118

卫星名称	成像时间	幅宽/km	传感器	波谱范围/μm	空间分辨率/m
GF-1	2018-10-30	60	全色	0.45~0.90	2
			多光谱	0.45~0.52	8
				0.52~0.59
				0.63~0.69
				0.77~0.89
GF-2	2019-03-20	45	全色	0.45~0.90	1
			多光谱	0.45~0.52	4
				0.52~0.59
				0.63~0.69
				0.77~0.89

软件环境		硬件环境
项目	内容	项目	内容
系统	Windows server 2016	CPU	InterXeno W-2125
语言	Python 3.6	内存	16 G
运算平台	CUDA10.0	GPU	Quadro P6000
加速器	CUDNN7.5	显存	24 G
学习框架	TensorFlow1.8

		真值
		正确	错误
提取	正确	TP	FP
提取	错误	FN	TN

影像		GF-1				GF-2
区域		A		B		A		B
区域		养殖区	非养殖区	养殖区	非养殖区	养殖区	非养殖区	养殖区	非养殖区
提取	养殖区	9 977	1 529	11 363	1 690	72 852	6 441	101 203	10 559
提取	非养殖区	1 970	986 524	3 417	143 530	2 965	3 917 742	9 021	3 879 217
精确率		86.71%		87.05%		91.88%		90.55%
召回率		83.51%		76.88%		96.09%		91.82%
F值		85.08%		81.65%		93.94%		91.18%
准确率		99.65%		96.81%		99.76%		99.51%
Kappa系数		0.85		0.8		0.94		0.91