Vegetation coverage extraction model of winter wheat based on pixel dichotomy
-
摘要:目的
快速准确提取冬小麦返青期植被覆盖度信息。
方法利用无人机获取田间冬小麦可见光图像,提取图像中4种常见可见光植被指数;在像元二分法原理的基础上,分别构建基于差异植被指数(Visible-band difference vegetation index,VDVI)、过绿指数(Excess green,EXG)、归一化绿蓝差异指数(Normalized green-blue difference index,NGBDI)和归一化绿红差异指数(Normalized green-red difference index,NGRDI)的植被覆盖度提取模型,采用支持向量机(Support vector machine,SVM)监督分类结果作为真值对各模型进行精度验证。
结果4种模型中,利用VDVI植被覆盖度提取模型获取的植被覆盖度精度最高,提取效果较好。与监督分类结果对比,4种植被覆盖度提取模型的提取误差(EF)分别为3.36%、15.68%、8.74%和15.46%,R2分别为0.946 1、0.934 4、0.695 3和0.746 0,均方根误差(RMSE)分别为0.021 9、0.059 5、0.042 0和0.055 9。
结论采用可见光植被指数结合像元二分法构建植被覆盖度提取模型实现了冬小麦返青期植被覆盖度准确快速提取,为植被覆盖度提取提供了一种新途径,可为无人机遥感监测提供参考。
Abstract:ObjectiveTo quickly and accurately extract vegetation coverage information of winter wheat in turning green period.
MethodThe field visible light images of winter wheat were obtained by UAV, and four common visible light vegetation indices were extracted. According to the principle of pixel dichotomy model, the vegetation coverage extraction models were established based on visible-band difference vegetation index(VDVI), excess green (EXG), normalized green-blue difference index (NGBDI) and normalized green-red difference index (NGRDI) respectively. The accuracies of four models were verified using the support vector machine (SVM) supervised classification results as the truth values.
ResultThe VDVI vegetation coverage extraction model had the highest accuracy and the best extraction effect to extract vegetation coverage among the four models. Compared with the supervised classification results, the extraction errors (EF) of four vegetation coverage extraction models were 3.36%, 15.68%, 8.74% and 15.46% respectively, the values of R2 were 0.946 1, 0.934 4, 0.695 3 and 0.746 0 respectively, and the values of root mean square error (RMSE) were 0.021 9, 0.059 5, 0.042 0 and 0.055 9 respectively.
ConclusionThe vegetation coverage extraction model based on visible vegetation index and pixel dichotomy has realized accurate and rapid extraction of vegetation coverage of winter wheat in turning green period, which provides a new way to extract vegetation coverage and a reference for UAV remote sensing monitoring vegetation coverage information.
-
Keywords:
- UAV remote sensing /
- winter wheat /
- pixel dichotomy /
- visible vegetation index /
- vegetation coverage
-
-
表 1 冬小麦分类精度评价
Table 1 Evaluation of classification accuracy of winter wheat
类别
Category小麦/像素
Wheat土壤/像素
Soil样本总数/像素
Total sample size用户精度/%
User accuracy小麦/像素 Wheat 20 779 3 20 782 99.99 土壤/像素 Soil 92 25 806 25 898 99.64 样本总数/像素 Total sample size 20 871 25 809 46 680 用户精度/% User accuracy 99.56 99.99 表 2 植被覆盖度精度评价
Table 2 Accuracy evaluation of vegetation coverage
植被指数模型
Vegetation index model覆盖度 Vegetation coverage 提取误差/%
Extraction error像元二分法
Pixel dichotomy监督分类
Supervised classification差值
Difference valueVDVI 0.321 452 0.332 623 0.011 171 3.36 NGBDI 0.303 564 0.332 623 0.029 059 8.74 NGRDI 0.384 043 0.332 623 0.051 420 15.46 EXG 0.384 780 0.332 623 0.052 157 15.68 表 3 植被覆盖度等级分布统计结果
Table 3 Statistical results of vegetation coverage grade distribution
植被覆盖度
Vegetation coverage像元数
Pixel number百分比/%
Percentage0~0.10 1 960 500 8.15 0.10~0.30 12 265 775 51.01 0.30~0.45 4 166 478 17.33 0.45~0.60 2 460 841 10.23 0.60~1.00 3 192 675 13.28 总计 Total 24 046 269 100 -
[1] 赵广才, 常旭虹, 王德梅, 等. 中国小麦生产发展潜力研究报告[J]. 作物杂志, 2012(3): 1-5. [2] 刘治开, 牛亚晓, 王毅, 等. 基于无人机可见光遥感的冬小麦株高估算[J]. 麦类作物学报, 2019, 39(7): 859-866. [3] PUREVDORJ T, ISHIYAMA T, HONDA Y, et al. Relationships between percent vegetation cover and vegetation indices[J]. Int J Remote Sens, 1998, 19(18): 3519-3535. doi: 10.1080/014311698213795
[4] GITELSON A A, KAUFMAN Y J, STARK R, et al. Novel algorithms for remote estimation of vegetation fraction[J]. Remote Sens Environ, 2002, 80(1): 76-87. doi: 10.1016/S0034-4257(01)00289-9
[5] 牛亚晓, 张立元, 韩文霆, 等. 基于无人机遥感与植被指数的冬小麦覆盖度提取方法 [J]. 农业机械学报, 2018, 49(4): 212-221. [6] SONG W, MU X, RUAN G, et al. Estimating fractional vegetation cover and the vegetation index of bare soil and highly dense vegetation with a physically based method[J]. Int J Appl Earth Obs, 2017, 58: 168-176. doi: 10.1016/j.jag.2017.01.015
[7] RUNDQUIST B C. The influence of canopy green vegetation fraction on spectral measurements over native tallgrass prairie[J]. Remote Sens Environ, 2002, 81(1): 129-135. doi: 10.1016/S0034-4257(01)00339-X
[8] ZHOU Q, ROBSON M, PILESJO P. On the ground estimation of vegetation cover in Australian Rangelands[J]. Int J Remote Sens, 1998, 19(9): 1815-1820. doi: 10.1080/014311698215261
[9] 陈虹兵, 黄贝贝, 彭道黎. 基于混合像元分解的植被覆盖度模型比较研究[J]. 西北林学院学报, 2018, 33(3): 203-207. doi: 10.3969/j.issn.1001-7461.2018.03.31 [10] 佟斯琴, 包玉海, 张巧凤, 等. 基于像元二分法和强度分析方法的内蒙古植被覆盖度时空变化规律分析[J]. 生态环境学报, 2016, 25(5): 737-743. [11] 李苗苗, 吴炳方, 颜长珍, 等. 密云水库上游植被覆盖度的遥感估算[J]. 资源科学, 2004, 26(4): 153-159. doi: 10.3321/j.issn:1007-7588.2004.04.022 [12] 何勇, 岑海燕, 李艺健, 等. 基于无人机遥感的农田信息获取技术及其机载装备[J]. 农业工程技术: 农产品加工, 2018, 38(9): 28-32. [13] 汪小钦, 王苗苗, 王绍强, 等. 基于可见光波段无人机遥感的植被信息提取[J]. 农业工程学报, 2015, 31(5): 152-159. doi: 10.3969/j.issn.1002-6819.2015.05.022 [14] 赵静, 杨焕波, 兰玉彬, 等. 基于无人机可见光图像的夏季玉米植被覆盖度提取方法[J]. 农业机械学报, 2019, 50(5): 232-240. doi: 10.6041/j.issn.1000-1298.2019.05.027 [15] TORRES-SÁNCHEZ J, PEÑA J M, DE CASTRO A I, et al. Multi-temporal mapping of the vegetation fraction in early-season wheat fields using images from UAV[J]. Comput Electron Agr, 2014, 103: 104-113. doi: 10.1016/j.compag.2014.02.009
[16] 龙满生, 何东健. 玉米苗期杂草的计算机识别技术研究[J]. 农业工程学报, 2007, 23(7): 139-144. doi: 10.3321/j.issn:1002-6819.2007.07.027 [17] HUNT E R, CAVIGELLII M, DAUGHTRY C S T, et al. Evaluation of digital photography from model aricraft for remote sensing of crop biomass and nitrogen status[J]. Precis Agric, 2005, 6(4): 359-378. doi: 10.1007/s11119-005-2324-5
[18] MEYER G E, NETO J C. Verification of color vegetation indices of automated crop image application[J]. Comput Electron Agr, 2008, 63(2): 282-293. doi: 10.1016/j.compag.2008.03.009
[19] 陈晋, 陈云浩, 何春阳, 等. 基于土地覆盖分类的植被覆盖率估算亚像元模型与应用[J]. 遥感学报, 2001, 5(6): 416-422. doi: 10.11834/jrs.20010603 [20] ZRIBI M, HÉGARAT-MASCLE S L, TACONET O, et al. Derivation of wild vegetation cover density in semi-arid regions: ERS2/SAR evaluation[J]. Int J Remote Sens, 2003, 24(6): 1335-1352. doi: 10.1080/01431160210146668
[21] 张军国, 韩欢庆, 胡春鹤, 等. 基于无人机多光谱图像的云南松虫害区域识别方法[J]. 农业机械学报, 2018, 49(5): 249-255. doi: 10.6041/j.issn.1000-1298.2018.05.029