Please wait a minute...
img

官方微信

遥感技术与应用  2019, Vol. 34 Issue (4): 694-703    DOI: 10.11873/j.issn.1004-0323.2019.4.0694
CNN 专栏     
基于CNN和农作物光谱纹理特征进行作物分布制图
周壮1,2,3(),李盛阳1,2,张康1,2,3,邵雨阳1,2
1. 中国科学院空间应用工程与技术中心 北京 100094
2. 中国科学院太空应用重点实验室 北京 100094
3. 中国科学院大学 北京 100049
Crop Mapping Using Remotely Sensed Spectral and Context Features based on CNN
Zhuang Zhou1,2,3(),Shengyang Li1,2,Kang Zhang1,2,3,Yuyang Shao1,2
1. Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences, Beijing 100094, China
2. Key Laboratory of Space Utilization, Chinese Academy of Sciences, Beijing 100094, China
3. University of Chinese Academy of Sciences, Beijing 100049, China
 全文: PDF(4024 KB)   HTML
摘要:

以卷积神经网络(Convolutional Neural Network, CNN)为代表的深度学习技术,在农作物遥感分类制图领域具有广阔的应用前景。以多时相Landsat 8 多光谱遥感影像为数据源,搭建CNN模型对农作物进行光谱特征提取与分类,并与支撑向量机(SVM)常规分类方法进行对比。进一步引入影像纹理信息,利用CNN对农作物光谱和纹理特征进行提取,优化作物分布提取结果。实验表明:① 基于光谱特征的农作物分布提取,验证结果对比显示,CNN对应各类别精度、总体精度均优于SVM,其中二者总体精度分别为95.14%和91.77%;② 引入影像纹理信息后,基于光谱和纹理特征的CNN农作物分类总体精度提高至96.43%,Kappa系数0.952,且分类结果的空间分布更为合理,可有效区分花生、道路等精细地物,说明纹理特征可用于识别不同作物。基于光谱和纹理信息的CNN特征提取,可面向种植结构复杂区域实现农作物精准分类与分布制图。

关键词: 农作物遥感分类CNN纹理信息    
Abstract:

Deep learning algorithms such as Convolutional Neural Network (CNN) can learn the representative and discriminative features in a hierarchical manner from the remote sensing data. Considering the low-level features as the bottom level, the output feature representation from the top level of the network can be directly fed into a subsequent classifier for pixel-based classification, the CNN has a broad application prospect in the field of agricultural remote sensing. The advantage of CNN in feature extraction can obtain the crop classification in complex planting structure area from multi-spectral remote sensing data, which is difficult in conventional methods. In this paper, a crop mapping method using remotely sensed spectral and context features based on CNN from Landsat OLI data is proposed and applied in Yuanyang county.The architecture of the proposed CNN classifier contains eight layers with weights which are the input layer, two convolution layers, two max pooling layers, two full connection layers and output layer. These eight layers are implemented on spectral and context signatures from 4 different phase Landsat OLI images to discriminate different crops against others. Experimental results demonstrate that the proposed CNN classifier can achieve better classification performance than support vector machines in spectral domain. The context features calculated by the gray level co-occurrence matrix method from Landsat OLI data can enhance the proposed CNN method to achieve the best results.In terms of verification accuracy, the proposed CNN classifier is superior than SVM in spectral domain. The overall accuracy of the two methods is 95.14% and 91.77%, respectively. The accuracy of the proposed classifier is further improved by adding spatial context features on the basis of spectral information. The overall accuracy and Kappa coefficient of the proposed method is 96.43% and 0.952.Furthermore, the crop mapping using spectral and context features based on CNN achieves better spatial representation especially for peanut and roads which is easy to form mixed-pixel. The context features can be extracted by the CNN to enhance the feature representation of these small objects.The CNN-based method from remotely sensed spectral and context features for crop mapping can achieve outstanding performance especially for the fine ground objects in complex planting structure area such as peanuts and roads.

Key words: Crop    Remote sensing    Classification    CNN    Context features
收稿日期: 2018-06-25 出版日期: 2019-10-16
ZTFLH:  TP79  
基金资助: 国家重点研发计划(2018YFD1100405)
作者简介: 周 壮(1990— ),男,江苏徐州人,助理工程师,主要从事遥感技术应用研究。E?mail:zhouzhuang@csu.ac.cn
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
周壮
李盛阳
张康
邵雨阳

引用本文:

周壮,李盛阳,张康,邵雨阳. 基于CNN和农作物光谱纹理特征进行作物分布制图[J]. 遥感技术与应用, 2019, 34(4): 694-703.

Zhuang Zhou,Shengyang Li,Kang Zhang,Yuyang Shao. Crop Mapping Using Remotely Sensed Spectral and Context Features based on CNN. Remote Sensing Technology and Application, 2019, 34(4): 694-703.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2019.4.0694        http://www.rsta.ac.cn/CN/Y2019/V34/I4/694

图1  研究区概况
编号时相传感器空间分辨率/m云量/%
12015年7月28日OLI301.9
22015年8月13日OLI304.7
32015年9月14日OLI300
42015年10月16日OLI300
52015年8月28日PMS0.80
表1  遥感影像列表
日期07-2808-1309-1410-16
水稻分蘖抽穗成熟
玉米拔节成熟
花生开花下针成熟
表2  影像对应农作物物候历
编号类别训练集验证集测试集
1水稻13345267
2玉米12542251
3花生9230183
4其他15451307
表3  样本数据列表
图2  技术流程图
图3  图像转化流程
图4  CNN模型结构
图6  研究区农作物分类结果空间分布
类别水稻玉米花生其他总计UA/%
PA/%93.4588.8486.0396.04
水稻25744226796.25
玉米722314725188.84
花生917154318384.15
其他27729130794.79
总计2752511793031 008
表4  SVM混淆矩阵结果
类别水稻玉米花生其他总计UA/%
PA/%95.2493.2894.1597.11
水稻26011526797.38
玉米82366125194.02
花生316161318387.98
其他20330230798.37
总计2732531713111 008
表5  基于光谱信息的CNN分类混淆矩阵结果
类别水稻玉米花生其他总计UA/%
PA/%96.6994.8094.9498.38
水稻26301326798.50
玉米72376125194.42
花生112169118392.35
其他11230330798.70
总计2722501783081 008
表6  基于光谱和空间信息的CNN分类混淆矩阵结果
分类方法总体精度/%Kappa系数
SVM分类91.770.889
基于光谱的CNN分类95.140.934
基于光谱+空间的CNN分类96.430.952
表7  三种方法的分类精度和Kappa系数
图5  农作物近红外反射率变化曲线
图7  研究区农作物分类结果局部对比
1 TangHuajun, WuWenbin, YangPeng, et al. Recent Progresses in Monitoring Crop Spatial Patterns by Using Remote Sensing Technologies[J]. Scientia Agricultura Sinica, 2010, 43(14): 2879-2888.
1 唐华俊, 吴文斌, 杨鹏, 等. 农作物空间格局遥感监测研究进展[J]. 中国农业科学, 2010, 43(14): 2879-2888.
2 XiaoX, BolesS, FrolkingS, et al. Mapping Paddy Rice Agriculture in South and Southeast Asia Using Multi-temporal MODIS Images[J]. Remote Sensing of Environment, 2006, 100(1): 95-113.
3 CaoWeibin, YangBangjie, SongJinpeng. Spectral Information based Model for Cotton Identification on Landsat TM Image[J]. Transactions of the CSAE, 2004, 20(4): 112-116.
3 曹卫彬, 杨邦杰, 宋金鹏. TM影像中基于光谱特征的棉花识别模型[J]. 农业工程学报, 2004, 20(4):112-116.
4 JiaKun, LiQiangzi, TianYichen, et al. Accuracy Improvement of Spectral Classification of Crop Using Microwave Backscatter Data[J]. Spectroscopy and Spectral Analysis, 2011, 31(2):483-487.
4 贾坤, 李强子, 田亦陈,等. 微波后向散射数据改进农作物光谱分类精度研究[J]. 光谱学与光谱分析, 2011,31(2): 483-487.
5 YangC, EverittJ H, MurdenD. Evaluating High Resolution SPOT 5 Satellite Imagery for Crop Identification[J]. Computers and Electronics in Agriculture, 2011, 75(2): 347-354.
6 LiuKebao, LiuShubin, LiZhongjun. Extraction on Cropping Structure based on High Spatial Resolution Remote Sensing Data[J]. Chinese Journal of Agricultural Resources and Regional Planning, 2014, 35(1): 21-26.
6 刘克宝, 刘述彬, 陆忠军,等. 利用高空间分辨率遥感数据的农作物种植结构提取[J]. 中国农业资源与区划, 2014, 35(1):21-26.
7 ZhuDengsheng, PanJiazhi, HeYong. Identification Methods of Crop and Weeds based on VIs/NIR spectroscopy and RBF-NN model[J]. Spectroscopy and Spectral Analysis, 2008, 28(5): 1102-1106.
7 朱登胜, 潘家志, 何勇. 基于光谱和神经网络模型的作物与杂草识别方法研究[J]. 光谱学与光谱分析, 2008, 28(5): 1102-1106.
8 PengGuangxiong, GongAdu, CuiWeihong. Study on Methods Comparision of Typical Remote Sensing Classification based on Multi-temporal Images[J]. Journal of Geo-Information Science, 2012, 11(2): 20-26.
8 彭光雄, 宫阿都, 崔伟宏, 等. 多时相影像的典型区农作物识别分类方法对比研究[J]. 地球信息科学学报, 2012, 11(2): 225-230.
9 XiongQinxue, HuangJingfeng. Estimation of Autumn Harvest Crop Planting Area based on NDVI Sequential Characteristics[J]. Transactions of the Chinese Society of Agricultural Engineering, 2009, 25(1): 144-148.
9 熊勤学, 黄敬峰. 利用 NDVI 指数时序特征监测秋收作物种植面积[J]. 农业工程学报, 2009, 25(1): 144-148.
10 FoersterS, KadenK, FoersterM, et al. Crop Type Mapping Using Spectral-temporal Profiles and Phenological Information[J]. Computers and Electronics in Agriculture, 2012, 89: 30-40.
11 WangLianxi, XuShengnan, LiQi, et al. Extraction of Winter Wheat Planted Area in Jiangsu Province Using Decision tree and Mixed-pixel Methods[J]. Transactions of the Chinese Society of Agricultural Engineering, 2016, 32(5): 182-187.
11 王连喜, 徐胜男, 李琪,等. 基于决策树和混合像元分解的江苏省冬小麦种植面积提取[J]. 农业工程学报, 2016(5):182-187.
12 WangWenjing, ZhangXia, ZhaoYinde, et al. Cotton Extraction Method of Integrated Multi-features based on Multi-temporal Landsat 8 Images[J]. Journal of Remote Sensing, 2017, 21(1):115-124.
12 王文静, 张霞, 赵银娣, 等. 综合多特征的 Landsat 8 时序遥感图像棉花分类方法[J]. 遥感学报, 2017, 21(1): 115-124.
13 LiuJikai, ZhongShiquan, LiangWenhai,et al. Extraction on Crops Planting Structure based on Multi-temporal Landsat 8 OLI Images[J]. Remote Sensing Technology and Application, 2015, 30(4): 775-783.
13 刘吉凯, 钟仕全, 梁文海. 基于多时相Landsat 8 OLI影像的作物种植结构提取[J]. 遥感技术与应用, 2015, 30(4):775-783.
14 LiXiaohui, WangHong, LiXiaobing, et al. Study on Crops Remote Sensing Classification based on Multi-temporal Landsat 8 OLI Images[J]. Remote Sensing Technology and Application, 2019, 34(2):384-397.
14 李晓慧, 王宏, 李晓兵, 等.基于多时相Landsat 8 OLI 影像的农作物遥感分类研究[J]. 遥感技术与应用,2019,34(2):387-397.
15 ZhangRi, MaJianwen. A Feature Selection Algorithm for Hyperspectual Data with SVM-RFE[J]. Geomatics and Information Science of Wuhan University, 2009, 34(7):834-837.
15 张睿, 马建文. 一种SVM-RFE高光谱数据特征选择算法[J]. 武汉大学学报(信息科学版), 2009, 34(7):834-837.
16 ShiFeifei, GaoXiaohong, YangLingyu, et al. Research on Jypical Crop Classification based on HJ-1A Hyperspectral Data in the Huangshui River Basin[J]. Remote Sensing Technology and Application, 2017, 32(2): 206-217.
16 史飞飞,高小红,杨灵王,等.基于HJ-1A高光谱遥感数据的湟水流域典型农作物分类研究[J]. 遥感技术与应用,2017,32(2):201-217.
17 PingYaopeng, ZangShuying. Crop Identification based on MODIS NDVI Time-series Data and Phenological Characteristics[J]. Journal of Natural Resources, 2016, 31(3): 503-513.
17 平跃鹏, 臧淑英. 基于 MODIS 时间序列及物候特征的农作物分类[J]. 自然资源学报, 2016, 31(3): 503-513.
18 KhatamiR, MountrakisG, StehmanS V. A Meta-analysis of Remote Sensing Research on Supervised Pixel-based Land-cover Image Classification Processes: General Guidelines for Practitioners and Future Research[J]. Remote Sensing of Environment, 2016, 177: 89-100.
19 KrizhevskyA, SutskeverI, HintonG E. Imagenet Classification with Deep Convolutional Neural Networks[C]⫽Advances in Neural Information Processing Systems. 2012: 1097-1105.
20 MakantasisK, KarantzalosK, DoulamisA, et al. Deep Supervised Learning for Hyperspectral Data Classification Through Convolutional Neural Networks[C]⫽2015 IEEE International Geoscience and Remote Sensing Symposium(IGARSS), 2015: 4959-4962.
21 ZhangF, DuB, ZhangL. Saliency-Guided Unsupervised Feature Learning for Scene Classification[J]. IEEE Transactions on Geoscience & Remote Sensing, 2015, 53(4):2175-2184.
22 ZhangKang, BaoqingHei, ZhouZhuang, et al. CNN with Coefficient of Variation-based Dimensionality Reduction for Hyperspectral Remote Sensing Images Classification[J]. Journal of Remote Sensing, 2018, 22(1):87-96.
22 张康,黑保琴, 周壮,等. 变异系数降维的CNN高光谱遥感图像分类[J]. 遥感学报, 2018, Journal of Remote Sensing, 2018, 22(1):87-96.
23 HuangYun, TangLinbo, LiZhen, et al. Research on Peanut Planting Area Classification Technology Using Remote Sensing Image based Deep Learning [J]. Journal of Signal Processing, 2019,35(4):617-622.
23 黄云,唐林波,李震,等.采用深度学习的遥感图像花生种植区域分类技术研究[J].信号处理,2019,35(4):617-622.
24 MaLi. Extracting Corn Planting Area by Multi-source Data with SVM Mixed-field Decomposed Method[D]. Xi'an: Xi'an University of Science and Technology, 2009.
24 马丽. 多源信息复合的 SVM 混合地块分解法提取玉米种植面积[D]. 西安: 西安科技大学, 2009.
25 LecunY, BottouL, BengioY, et al. Gradient-based Learning Applied to Document Recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
26 HintonG E, SalakhutdinovR R. Reducing the Dimensionality of Data with Neural Networks[J]. Science, 2006, 313(5786): 504-507.
27 HuW, HuangY, WeiL, et al. Deep Convolutional Neural Networks for Hyperspectral Image Classification[J]. Journal of Sensors, 2015: 1-12.
28 YueJ, ZhaoW, MaoS, et al. Spectral–spatial Classification of Hyperspectral Images Using Deep Convolutional Neural Networks[J]. Remote Sensing Letters, 2015, 6(6): 468-477.
29 LiYandong, HaoZongbo, LeiHang. Survey of Convolutional Neural Network[J]. Journal of Computer Applications, 2016(9): 2508-2515.
29 李彦冬, 郝宗波, 雷航. 卷积神经网络研究综述[J]. 计算机应用, 2016, 36(9): 2508-2515.
[1] 王思恒, 黄长平, 张立福, 高显连, 付安民. 陆地生态系统碳监测卫星远红波段叶绿素荧光反演算法设计[J]. 遥感技术与应用, 2019, 34(3): 476-487.
[2] 林中立, 徐涵秋. 近20年来新旧“火炉城市”热岛状况对比研究[J]. 遥感技术与应用, 2019, 34(3): 521-530.
[3] 丛铭, 段晨曦, 许妙忠, 陶翊婷. 基于格式塔形状分析的高分辨率遥感影像道路提取[J]. 遥感技术与应用, 2019, 34(3): 595-601.
[4] 卢静, 贾立, 郑超磊, 胡光成. 遥感水分收支对区域水资源估算潜能[J]. 遥感技术与应用, 2019, 34(3): 630-638.
[5] 林沂, 周国清, 童庆禧. 偏振激光雷达对地观测遥感 [J]. 遥感技术与应用, 2019, 34(2): 232-242.
[6] 徐凡, 张雪红, 石玉立. 基于激光雷达和航拍影像的城市地物分类研究[J]. 遥感技术与应用, 2019, 34(2): 253-262.
[7] 李伟, 唐伶俐, 吴昊昊, 腾格尔, 周梅. 轻小型无人机载激光雷达系统研制及电力巡线应用[J]. 遥感技术与应用, 2019, 34(2): 269-274.
[8] 丁海宁. 黄土高原土壤铁元素含量遥感反演方法 [J]. 遥感技术与应用, 2019, 34(2): 275-283.
[9] 李晓慧, 王宏, 李晓兵, 迟登凯, 汤曾伟, 韩重远. 基于多时相Landsat 8 OLI影像的农作物遥感分类研究[J]. 遥感技术与应用, 2019, 34(2): 389-397.
[10] 马鹏飞, 厉青, 陈辉, 张丽娟, 张玉环, 王桥, 周春艳, 毛慧琴, 陈翠红, 王中挺. 京津冀及周边地区大气污染防治重点关注区域遥感综合分析[J]. 遥感技术与应用, 2019, 34(2): 404-411.
[11] 向嘉敏, 祝善友, 张桂欣, 刘祎, 周洋. 灰霾遥感监测研究进展[J]. 遥感技术与应用, 2019, 34(1): 12-20.
[12] 鲁军景, 孙雷刚, 黄文江.  作物病虫害遥感监测和预测预警研究进展[J]. 遥感技术与应用, 2019, 34(1): 21-32.
[13] 迟文峰, 匡文慧, 党晓宏, 潘涛, 刘正佳. 基于遥感的内蒙古地级市土地覆盖结构时空变化特征分析[J]. 遥感技术与应用, 2019, 34(1): 33-45.
[14] 刘亲亲, 崔耀平, 刘素洁, 李楠. 中国不同土地利用类型分光辐射地表反照率研究[J]. 遥感技术与应用, 2019, 34(1): 46-56.
[15] 谷晓天, 高小红, 马慧娟, 史飞飞, 刘雪梅, 曹晓敏. 复杂地形区土地利用/土地覆被分类机器学习方法比较研究[J]. 遥感技术与应用, 2019, 34(1): 57-67.