
基于无人机高清影像的棉花单产预测
Cotton yield estimation based on UAV high-resolution images
【目的】在苗期对棉花产量进行预测,及早为棉花的田间管理提供技术手段和主要依据。【方法】基于无人机高清影像数据,首先利用绿叶指数(green leaf index, GLI)对3~4叶期棉花幼苗进行准确识别并提取;然后根据试验区域内棉花幼苗直径的相对大小以自然断点法对棉花进行等级划分,并在收获期分别统计单株结铃数和铃重;最后结合试验区内不同等级棉花的数量以及对应的单株结铃数和铃重构建棉花估产模型对棉花产量进行预测,并将该模型命名为基于苗期状态的估产模型,简称NDCS(number and diameter of cotton seedlings)。【结果】在34个植被指数中,图像分割效果最好的是GLI,通过不同尺度上的重复检验,棉花幼苗的平均提取精度为96.2%。试验区内共提取棉花380 715株,其中甲等苗2 657株,乙等苗103 753株,丙等苗214 691株,丁等苗59 614株。经验证,估产模型的决定系数为0.919 2,估产模型的均方根误差为0.168 7,经与实际产量对比,模型的估产精度为94.7%。【结论】利用棉花苗期图像数据结合与产量密切相关的指标实现了棉花产量的预测,为棉花估产提供了一种新的思路和方法。
[Objective] This study aims to build a cotton yield prediction model at seedling stage for cotton field management. [Method] This study was conducted based on UAV (unmanned aerial vehicle) high-resolution image data. Firstly, the cotton seeding at 3-4 leaf stage were identified and extracted on the UAV images using green leaf index (GLI). Then, cotton seedlings were graded according to the diameters of cotton in the trial area by the natural breakpoint method, and the number of cotton bolls per plant and the boll weight were evaluated at harvest time. Finally, the cotton yield estimation model NDCS (number and diameter of cotton seedlings) based on the growth state at the seedling stage was constructed to predict the cotton yield by combining the grade of cotton seedlings in the trial area and the number of cotton bolls per plant and the boll weight. [Results] Among the 34 vegetation indices, GLI showed the best image segmentation effect. The average rate of precision extraction of cotton seedlings on multiple repeated trial at different scales was 96.2%. A total of 380 715 cotton plants were extracted from the trial area, including 2 657 seedlings of A grade, 103 753 seedlings of B grade, 214 691 seedlings of C grade and 59 614 seedlings of D grade. The NDCS model showed relative high accuracy on yield prediction, with the coefficient of determination of 0.919 2. The root mean squared error(RMSE) of the yield estimation model was 0.168 7. The estimated accuracy of the yield prediction model was 94.7% when compared with the actual yields. [Conclusion] This study used the high resolution UAV image of cotton seedling combined with the performance of yield-related traits to achieve cotton yield prediction, which provided a new route and method for cotton yield estimation.
无人机 / 棉花 / 苗期 / 分级 / 估产模型 {{custom_keyword}} /
unmanned aerial vehicle / cotton / seedling / grading / yield estimation model {{custom_keyword}} /
表1 植被指数列表Table 1 List of vegetation indices |
植被指数 Vegetation index | 计算公式 Calculation formula | 文献 Literature |
---|---|---|
绿叶指数 Green leaf index | [24] | |
修正三角植被指数 Modified triangular vegetation index 1 | [25] | |
转换差异植被指数 Transformed difference vegetation index | [26] | |
比值植被指数 Simple ratio | [27] | |
土壤调节植被指数 Soil adjusted vegetation index | [28] | |
重归一化植被指数 Renormalized difference vegetation index | [29] | |
调整土壤亮度植被指数 Optimized soil adjusted vegetation index | [30] | |
归一化植被指数 Normalized difference vegetation index | [31] | |
改进比值植被指数 Modified simple ratio | [32] | |
绿色土壤调整植被指数 Green soil adjusted vegetation index | [33] | |
绿色比值植被指数 Green ratio vegetation index | [34] | |
绿色优化土壤调整植被指数 Green optimized soil adjusted vegetation index | [33] | |
绿色归一化差异植被指数 Green normalized difference vegetation index | [35] | |
绿色差异植被指数 Green difference vegetation index | [36] | |
绿色叶绿素指数 Green chlorophyll index | [37] | |
可见大气阻抗植被指数 Visible atmospherically resistant index | [38] | |
绿波大气阻抗指数 Green atmospherically resistant index | [39] | |
差值植被指数 Difference vegetation index | [40] | |
大气阻抗植被指数 Atmospherically resistant vegetation index | [41] |
注:NIR、R、G、B、Rx分别表示近红外波段、红光波段、绿光波段、蓝光波段和x波长的反射率;y代表大气光路订正系数,本次试验中取y =1;γ是1个加权函数,取决于大气中气溶胶条件,本次试验中取Envi 5.6中的默认值γ =1.7。 | |
Note: NIR, R, G, B and Rx indicate the reflectance at NIR band, red band, green band, blue band, and x wavelength respectively; y represents the atmospheric light path revision factor, and y=1 was set in this study; γ is a weighting function that depends on the aerosol conditions, and the default value γ=1.7 in Envi 5.6 was taken in this study. |
表2 不同田块尺度棉花幼苗提取精度变化Table 2 Variation of extraction accuracy of cotton seedlings at different scales |
面积 Area/m2 | 影像提取棉苗数量 Number of seedlings by image extraction | 人工统计棉苗数量 Number of seedlings by manual count | 精度 Accuracy/% | 平均精度 Average accuracy/% | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 3 | 1 | 2 | 3 | 1 | 2 | 3 | ||||
3.94×3.94 | 301 | 291 | 313 | 328 | 320 | 342 | 91.7 | 90.9 | 91.5 | 91.3 | ||
6.06×6.06 | 601 | 582 | 621 | 646 | 629 | 669 | 93.1 | 92.6 | 92.8 | 92.8 | ||
8.18×8.18 | 1 056 | 1 006 | 1 093 | 1 106 | 1 062 | 1 160 | 95.5 | 94.7 | 94.2 | 94.8 | ||
10.3×10.3 | 1 690 | 1 575 | 1 741 | 1 754 | 1 649 | 1 831 | 96.3 | 95.5 | 95.1 | 95.6 | ||
12.42×12.42 | 2 426 | 2 289 | 2 489 | 2 506 | 2 382 | 2 601 | 96.8 | 96.1 | 95.7 | 96.2 |
[1] |
戴建国, 薛金利, 赵庆展, 等. 利用无人机可见光遥感影像提取棉花苗情信息[J/OL]. 农业工程学报, 2020, 36(4): 63-71 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2020.04.008.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[2] |
周亮, 慕号伟, 马海姣, 等. 基于卷积神经网络的中国北方冬小麦遥感估产[J/OL]. 农业工程学报, 2019, 35(15): 119-128 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2019.15.016.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[3] |
解毅, 王鹏新, 王蕾, 等. 基于作物及遥感同化模型的小麦产量估测[J/OL]. 农业工程学报, 2016, 32(20): 179-186 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2016.20.023.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[4] |
谭昌伟, 杜颖, 童璐, 等. 基于开花期卫星遥感数据的大田小麦估产方法比较[J/OL]. 中国农业科学, 2017, 50(16): 3101-3109 [2021-01-30]. https://doi.org/10.3864/j.issn.0578-1752.2017.16.005.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[5] |
刘姣娣, 曹卫彬, 李华, 等. 基于植被指数的新疆棉花遥感估产模型研究[J/OL]. 石河子大学学报(自然科学版), 2011, 29(2): 153-157 [2021-01-30]. https://doi.org/10.13880/j.cnki.65-1174/n.2011.02.027.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[6] |
程乙峰, 贡璐, 张雪妮, 等. 新疆北疆棉花多元复合遥感估产模型研究[J/OL]. 新疆农业科学, 2012, 49(8): 1497-1502 [2021-01-30]. https://doi.org/10.6048/j.issn.1001-4330.2012.08.020.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[7] |
温鹏飞, 朱鹏, 张强, 等. 不同水氮耦合条件下棉花冠层NDVI分析与产量估算[J/OL]. 中国农业大学学报, 2016, 21(1): 33-38 [2021-01-30]. https://doi.org/10.11841/j.issn.1007-4333.2016.01.04.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[8] |
白丽, 王进, 蒋桂英, 等. 干旱区基于高光谱的棉花遥感估产研究[J/OL]. 中国农业科学, 2008(8): 2499-2505 [2021-01-30]. https://doi.org/10.3864/j.issn.0578-1752.2008.08.041.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[9] |
黄健熙, 黄海, 马鸿元, 等. 遥感与作物生长模型数据同化应用综述[J/OL]. 农业工程学报, 2018, 34(21): 144-156 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2018.21.018.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[10] |
吴蕾, 柏军华, 肖青, 等. 作物生长模型与定量遥感参数结合研究进展与展望[J/OL]. 农业工程学报, 2017, 33(9): 155-166 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2017.09.020.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[11] |
戴建国, 蒋楠, 薛金利, 等. 基于CNN-BiLSTM的棉花产量预测方法[J/OL]. 农业工程学报, 2021, 37(17): 152-159 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2021.17.017.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[12] |
柏军华, 李少昆, 王克如, 等. 棉花产量遥感预测的L-Y模型构建[J]. 作物学报, 2006, 32(6): 840-844.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[13] |
何亚娟, 潘学标, 裴志远, 等. 基于SPOT遥感数据的甘蔗叶面积指数反演和产量估算[J/OL]. 农业机械学报, 2013, 44(5): 226-231 [2021-01-30]. https://doi.org/10.6041/j.issn.1000-1298.2013.05.039.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[14] |
徐春萌, 田芷源, 陈威, 等. 基于DSSAT作物模型的中美大豆主产区单产模拟与验证[J/OL]. 农业工程学报, 2021, 37(3): 132-139 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2021.03.016.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[15] |
王鹏新, 齐璇, 李俐, 等. 基于粒子群优化投影寻踪的玉米单产估测[J/OL]. 农业工程学报, 2019, 35(13): 145-153 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2019.13.016.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[16] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[17] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[18] |
刘国敏, 刘玉宝, 陈辉, 等. 油茶F1代苗木分级对盛果期生长量的影响初探[J/OL]. 亚热带农业研究, 2008(2): 110-114 [2021-01-30]. https://doi.org/10.13321/j.cnki.subtrop.agric.res.2008.02.017.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[19] |
赵建华, 叶力勤, 安巍, 等. 宁夏枸杞苗木分级及不同质量苗木定植初期生长特征[J]. 浙江农林大学学报, 2011, 28(3): 450-454.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[20] |
王生华. 草珊瑚实生苗年生长规律和分级标准研究[J/OL]. 福建农业学报, 2013, 28(10): 993-998 [2021-01-30]. https://doi.org/10.19303/j.issn.1008-0384.2013.10.012.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[21] |
简才源, 韦小丽, 段如雁, 等. 贵州不同区域闽楠裸根苗的生长评价与苗木分级[J]. 贵州农业科学, 2015, 43(6): 153-157.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[22] |
徐权, 郭鹏, 祁佳峰, 等. 基于无人机影像的SEGT棉花估产模型构建[J/OL]. 农业工程学报, 2020, 36(16): 44-51 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2020.16.006.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[23] |
郭鹏, 武法东, 戴建国, 等. 基于无人机可见光影像的农田作物分类方法比较[J/OL]. 农业工程学报, 2017, 33(13): 112-119 [2021-01-30]. https://doi.org/10.11975/j.issn.1002-6819.2017.13.015.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[24] |
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/OL]. 农业工程学报, 2015, 31(5): 152-157 [2021-01-30]. https://doi.org/10.3969/j.issn.1002-6819.2015.05.022.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[25] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[26] |
康悦, 文军, 张堂堂, 等. 卫星遥感数据评估黄土高原陆面干湿程度研究[J/OL]. 地球物理学报, 2014, 57(8): 2473-2483 [2021-01-30]. https://doi.org/10.6038/cjg20140809.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[27] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[28] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[29] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[30] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[31] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[32] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[33] |
王建光, 吕小东, 姚贵平, 等. 苜蓿和无芒雀麦混播草地高光谱遥感估产研究[J]. 中国草地学报, 2013, 35(1): 35-41.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[34] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[35] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[36] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[37] |
肖汉, 陈秀万, 杨振宇, 等. 基于光谱分析的草地叶绿素含量估测植被指数[J/OL]. 光谱学与光谱分析, 2014, 34(11): 3075-3078 [2021-01-30]. http://doi.org/10.3964/j.issn.1000-0593(2014)11-3075-04.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[38] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[39] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[40] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[41] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[42] |
林滢, 邵怀勇. 基于随机森林算法的河南省冬小麦产量预测最佳时间窗和影响因子研究[J/OL]. 麦类作物学报, 2020, 40(7): 874-880 [2021-01-30]. https://doi.org/10.7606/j.issn.1009-1041.2020.07.13.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[43] |
朱炯, 杜鑫, 李强子, 等. 关键时相长势-环境和景观特征对河北省县级尺度冬小麦单产估算精度影响分析[J]. 遥感学报, 2022, 26(7): 1354-1367 [2021-01-30]. https://doi.org/10.11834/jrs.20221019.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[44] |
闫春雨, 赵静, 兰玉彬, 等. 面向对象和多分类器的棉花出苗信息快速提取方法[J/OL]. 山东理工大学学报(自然科学版), 2021, 35(3): 55-59 [2021-01-30]. https://doi.org/10.13367/j.cnki.sdgc.2021.03.010.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[45] |
魏文丽, 陈圣波, 田粉粉. 基于GF-1全色多光谱传感器数据的地块级玉米出苗情况提取[J]. 科学技术与工程, 2018, 18(23): 21-25.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[46] |
杨亚莉, 向祖恒, 黄正明, 等. 华榛1年生实生苗质量分级标准的研究[J/OL]. 园艺与种苗, 2018, 38(1):34-37 [2021-01-30]. https://doi.org/10.16530/j.cnki.cn21-1574/s.2018.01.012.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[47] |
李玛, 苏钛, 李云, 等. 滇重楼种苗质量分级标准研究[J/OL]. 种子, 2020, 39(5): 141-143 [2021-01-30]. https://doi.org/10.16590/j.cnki.1001-4705.2020.05.141.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[48] |
周新华, 黄拯, 厉月桥, 等. 杉木容器苗分级标准研究[J/OL]. 中南林业科技大学学报, 2017, 37(9): 68-73 [2021-01-30]. https://doi.org/10.14067/j.cnki.1673-923x.2017.09.012.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[49] |
李亚麒, 孙继伟, 李建华, 等. 不同等级云南松幼苗生物量估测模型[J/OL]. 云南大学学报(自然科学版), 2019, 41(5): 1073-1082 [2021-01-30]. https://doi.org/10.7540/j.ynu.20190024.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
[50] |
关参政, 申培增. 甘肃枸杞种苗分级标准及其田间验证试验研究[J/OL]. 林业科技通讯, 2017(6): 29-32 [2021-01-30]. https://doi.org/10.13456/j.cnki.lykt.2017.06.008.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
文章所在专题
/
〈 |
|
〉 |