作物杂志,2023, 第1期: 226–232 doi: 10.16035/j.issn.1001-7283.2023.01.034

• 农业信息技术 • 上一篇    下一篇

基于无人机RGB遥感的甘蔗株高估测

梁永检1(), 吴文志2, 施泽升1, 唐利球1, 宋修鹏3, 颜梅新3, 郭强1, 秦昌鲜1, 何洪良1, 张小秋3()   

  1. 1广西南亚热带农业科学研究所,532415,广西龙州
    2广州极飞科技股份有限公司,510000,广东广州
    3广西农业科学院甘蔗研究所/广西甘蔗遗传改良重点实验室/农业农村部广西甘蔗生物技术与遗传改良重点实验室,530004,广西南宁
  • 收稿日期:2021-09-25 修回日期:2022-06-15 出版日期:2023-02-15 发布日期:2023-02-22
  • 通讯作者: 张小秋,主要从事甘蔗病害防控研究,E-mail:ZhangXiaoQiuXHD@163.com
  • 作者简介:梁永检,主要从事甘蔗栽培与遥感研究,E-mail:Yongjianliang_605@163.com
  • 基金资助:
    桂西南甘蔗育种与栽培技术研究团队(桂农科2021YT161);广西农业科学院科技发展基金(桂农科2020YM145);广西农业科学院科技发展基金(桂农科2021ZX03);广西农业科学院科技发展基金(桂农科2022YM29);国家自然科学基金(32001606);广西自然科学基金(2020GXNS FBA297040);国家现代农业产业技术体系广西甘蔗创新团队(nycytxgxcxtd-2021-03-04)

Estimation of Sugarcane Plant Height Based on UAV RGB Remote Sensing

Liang Yongjian1(), Wu Wenzhi2, Shi Zesheng1, Tang Liqiu1, Song Xiupeng3, Yan Meixin3, Guo Qiang1, Qin Changxian1, He Hongliang1, Zhang Xiaoqiu3()   

  1. 1Guangxi South Subtropical Agricultural Science Research Institute, Longzhou 532415, Guangxi, China
    2Guangzhou Jifei Technology Co., Ltd., Guangzhou 510000, Guangdong, China
    3Sugarcane Research Institute, Guangxi Academy of Agricultural Sciences/Guangxi Key Laboratory of Sugarcane Genetic Improvement/Key Laboratory of Sugarcane Biotechnology and Genetic Improvement (Guangxi), Ministry of Agriculture and Rural Affairs, Nanning 530004, Guangxi, China
  • Received:2021-09-25 Revised:2022-06-15 Online:2023-02-15 Published:2023-02-22

摘要:

为快速准确地估测甘蔗不同生育期株高,探讨了无人机RGB系统遥感估算株高的可行性及效果。利用无人机RGB遥感平台,获取苗期、分蘖期、伸长中、后期和工艺成熟期的影像,通过Pix4D mapper生成数字表面模型(digital surface model,DSM),采用Eris Arcmap提取株高,基于DSM提取的株高与实测株高建立各生育期的估测模型,采用决定系数(R2)、均方根误差(RMSE)和平均相对误差(MRE)对模型进行评价。结果表明,基于DSM提取甘蔗各生育期的株高高于实测株高;全生育期模型拟合性最好,预测精度较高(验证集R2、RMSE和MRE分别为0.9611、0.1623和0.1102),苗期株高模型预测精度最高。其他各生育期模型的拟合性不及全生育期和苗期,精度较低,工艺成熟期模型的预测精度最低,拟合性最差。因此,基于无人机RGB遥感平台获取甘蔗不同生育期影像后通过DSM提取株高并运用于甘蔗重要生育期株高的估测时,注意不同生育期模型的适用性。

关键词: 无人机, 遥感, 甘蔗, 株高

Abstract:

The purpose of this study is to investigate the feasibility and efficacy of sugarcane plant height estimation based on UAV RGB remote sensing, and get the estimate sugarcane plant height quickly and correctly at various growth stages. Images of sugarcane at several growth phases, including seedling, tillering, mid- elongation, late-elongation and maturation were collected using the UAV RGB remote sensing platform. The digital surface model (DSM) was generated by the software of Pix4D mapper and the plant height was extracted by the software of Eris Arcmap. The estimation models of plant height based on DSM and actual measurement at different growth stages were established. Coefficient of determination (R2), root mean square error (RMSE) and mean relative error (MRE) were used to evaluate the models. The results showed that the plant height extracted based on DSM at each growth stage was higher than that of actual measurement. The estimation model of whole growth season had the best fit and high prediction accuracy with R2 of 0.9611, RMSE of 0.1623 and MRE of 0.1102 in validation data set. The estimation model of seedling stage had the highest prediction accuracy, whereas the fits of other estimation models less than that of whole growth season and seedling stage, and their accuracies were also lower. The estimation model of maturity stage had the lowest accuracy and the worst fit. Therefore, it should be paid an attention to the applicability of plant height prediction model at different growth stages when estimate sugarcane plant height via extracting plant height from DSM images acquired based on UAV RGB remote sensing.

Key words: UAV, Remote sensing, Sugarcane, Plant height

表1

极侠-XMISSION无人机主要技术参数

参数Parameter 数值/类型Value/pattern
质量(含桨和电池)
Weight (includes paddle and battery) (kg)
2.25
轴距Wheel base (mm) 492
续航时间
Time of endurance (min)
35
定位系统Positioning system GNSS
定位精度PositionaI accuracy

垂直1.5cm+1ppm(RMS),水平1cm+ 1ppm(RMS)(1ppm是指飞行器每移动1km误差增加1mm)

表2

无人机遥感采集数据及Pix4D mapper拼接主要参数

参数Parameter 数值Value
仿地飞行高度Ground-like altitude (m) 45
航向、旁向重叠率Rate of course, lateral overlap (%) 75
影像数量Number of images 145
空间分辨率Spatial resolution (cm) 1.38
影像比例Image scale factor 1/2(多比例)
彩色影像传感器尺寸
Size of color image sensor (mm×mm)
13.133×8.755

图1

遥感与人工塔尺测量甘蔗株高示意图(基于三维点云数据构建) a:塔尺测量的株高,b:遥感测量的株高

图2

基于DSM提取的甘蔗株高和DOM图 a~f:DOM图,g~l:DSM图。a、g:种植期;b、h:苗期;c、i:分蘖期;d、j:伸长中期;e、k:伸长后期;f、l:工艺成熟期

表3

甘蔗不同生育期实测株高的描述性统计分析

生育期
Growth stage
样本数
Number of samples
最小值
Minimum (m)
最大值
Maximum (m)
平均值
Mean (m)
标准差
Standard deviation (m)
苗期(5月)Seedling stage (May) 150 0.22 0.51 0.35 0.0737
分蘖期(6月)Tillering stage (June) 150 0.63 1.94 1.23 0.2468
伸长中期(8月)Mid-elongation stage (August) 150 0.89 2.07 1.51 0.2737
伸长后期(9月)Late-elongation stage (September) 150 1.43 3.12 2.24 0.3258
工艺成熟期(11月)Maturity stage (November) 150 1.58 3.15 2.47 0.3550
全生育期Whole growth stage 750 0.22 3.15 1.56 0.8074

表4

甘蔗不同生育期基于DSM提取株高的描述性统计分析

生育期
Growth stage
样本数
Number of samples
最小值
Minimum (m)
最大值
Maximum (m)
平均值
Mean (m)
标准差
Standard deviation (m)
苗期(5月)Seedling stage (May) 150 0.34 0.89 0.68 0.1088
分蘖期(6月)Tillering stage (June) 150 1.21 2.39 1.74 0.2885
伸长中期(8月)Mid-elongation stage (August) 150 1.30 2.72 2.03 0.2912
伸长后期(9月)Late-elongation stage (September) 150 2.01 3.83 2.75 0.3657
工艺成熟期(11月)Maturity stage (November) 150 1.82 3.59 3.00 0.3427
全生育期Whole growth stage 750 0.34 3.83 2.04 0.8714

表5

甘蔗各生育期基于DSM取株高与实测株高的回归模型

生育期
Growth stage
一元线性回归模型
Simple linear
regression model
建模集
Calibration set
验证集
Validation set
R2 RMSE (m) MRE R2 RMSE (m) MRE
苗期(5月)Seedling stage (May) y=0.6203x-0.0782 0.8729 0.3419 0.3386 0.7387 0.0411 0.0344
分蘖期(6月)Tillering stage (June) y=0.7327x-0.0414 0.7804 0.5286 0.5113 0.8288 0.2854 0.1280
伸长中期(8月)Mid-elongation stage (August) y=0.8940x-0.2867 0.7813 0.5110 0.5042 0.7693 0.1554 0.1290
伸长后期(9月)Late-elongation stage (September) y=0.8666x-0.1058 0.7642 0.5059 0.4750 0.7846 0.2256 0.1621
工艺成熟期(11月)Maturity stage (November) y=0.8726x-0.1467 0.6863 0.5675 0.5380 0.6681 0.2066 0.1444
全生育期Whole growth season y=0.9229x-0.3131 0.9670 0.4980 0.4738 0.9611 0.1623 0.1102

图3

基于DSM提取的株高与实测值散点图

图4

甘蔗不同生育期株高实测值与估测值散点图

[1] Li Y R, Yang L T. Sugarcane agriculture and sugar industry in China. Sugar Tech, 2015, 17(1):1-8.
doi: 10.1007/s12355-014-0342-1
[2] 陆中华. 甘蔗产量构成因素与产量的关系. 种子, 2002(3):38-39,65.
[3] 邓祖湖, 徐良年, 韦先明, 等. 经济遗传值在甘蔗选育种的应用研究Ⅰ. 经济遗传值及性状经济权重的确定. 中国糖料, 2011(1):39-43.
[4] 李纯佳, 徐超华, 覃伟, 等. 甘蔗株高相关性状数量研究. 中国糖料, 2015, 37(3):1-3.
[5] 张建, 谢田晋, 尉晓楠, 等. 无人机多角度成像方式的饲料油菜生物量估算研究. 作物学报, 2021, 47(9):1816-1823.
doi: 10.3724/SP.J.1006.2021.04211
[6] 谢国雪, 马灿达, 张秀龙, 等. 无人机精准监测甘蔗长势技术研究与应用. 国土资源信息化, 2021(1):6-11.
[7] 张建, 谢田晋, 杨万能, 等. 近地遥感技术在大田作物株高测量中的研究现状与展望. 智慧农业(中英文), 2021, 3(1):1-15.
[8] Tirado S B, Hirsch C N, Springer N M. UAV-based imaging platform for monitoring maize growth throughout development. Plant Direct, 2020, 4(6):e00230.
[9] Hu P C, Chapman S C, Wang X M, et al. Estimation of plant height using a high throughput phenotyping platform based on unmanned aerial vehicle and self-calibration:example for sorghum breeding. European Journal of Agronomy, 2018, 95:24-32.
doi: 10.1016/j.eja.2018.02.004
[10] Malambo L, Popescu S C, Murray S C, et al. Multitemporal field-based plant height estimation using 3D point clouds generated from small unmanned aerial systems high-resolution imagery. International Journal of Applied Earth Observation and Geoinformation, 2018, 64:31-42.
doi: 10.1016/j.jag.2017.08.014
[11] Maesano M, Khoury S, Nakhle F, et al. UAV-based LiDAR for high-throughput determination of plant height and above-ground biomass of the bioenergy grass arundo donax. Remote Sensing, 2020, 12(20):1-20.
doi: 10.3390/rs12010001
[12] Zhou L, Gu X, Cheng S, et al. Analysis of plant height changes of lodged maize using UAV-LiDAR data. Agriculture, 2020, 10(5):146.
doi: 10.3390/agriculture10050146
[13] 张宏鸣, 谭紫薇, 韩文霆, 等. 基于无人机遥感的玉米株高提取方法. 农业机械学报, 2019, 50(5):241-250.
[14] 郭海, 樊江川, 李英伦, 等. 基于RGB-D点云的田间原位玉米株高测量试验研究. 农机化研究, 2021, 43(10):102-109.
[15] 付虹雨, 崔国贤, 李绪孟, 等. 基于无人机遥感图像的苎麻产量估测研究. 作物学报, 2020, 46(9):1448-1455.
doi: 10.3724/SP.J.1006.2020.04020
[16] 郭涛, 颜安, 耿洪伟. 基于无人机影像的小麦株高与LAI预测研究. 麦类作物学报, 2020, 40(9):1129-1140.
[17] Yuan W A, Li J T, Bhatta M, et al. Wheat height estimation using LiDAR in comparison to ultrasonic sensor and UAS. Sensors, 2018, 18(11):3731.
doi: 10.3390/s18113731
[18] 颜安, 郭涛, 陈全家, 等. 基于无人机影像的棉花株高预测. 新疆农业科学, 2020, 57(8):1493-1502.
doi: 10.6048/j.issn.1001-4330.2020.08.014
[19] 王庆, 车荧璞, 柴宏红, 等. 基于无人机可见光与激光雷达的甜菜株高定量评估. 农业机械学报, 2021, 52(3):178-184.
[20] Sumesh K C, Ninsawat S, Som-ard J. Integration of RGB-based vegetation index,crop surface model and object-based image analysis approach for sugarcane yield estimation using unmanned aerial vehicle. Computers and Electronics in Agriculture, 2021, 180:105903.
doi: 10.1016/j.compag.2020.105903
[21] 杨琦, 叶豪, 黄凯, 等. 利用无人机影像构建作物表面模型估测甘蔗LAI. 农业工程学报, 2017, 33(8):104-111.
[22] Sofonia J, Shendryk Y, Phinn S, et al. Monitoring sugarcane growth response to varying nitrogen application rates:A comparison of UAV SLAM LiDAR and photogrammetry. International Journal of Applied Earth Observation and Geoinformation, 2019, 82:101878.
doi: 10.1016/j.jag.2019.05.011
[23] 牛庆林, 冯海宽, 杨贵军, 等. 基于无人机数码影像的玉米育种材料株高和LAI监测. 农业工程学报, 2018, 34(5):73-82.
[24] Weiss M, Baret F. Using 3D point clouds derived from UAV RGB imagery to describe vineyard 3D macro-structure. Remote Sensing, 2017, 9(2):111.
doi: 10.3390/rs9020111
[25] 李燕强, 张娟娟, 熊淑萍, 等. 不同冬小麦品种株高的高光谱估算模型. 麦类作物学报, 2012, 32(3):523-529.
[26] 鲁冬冬, 邹进贵. 三维激光点云的降噪算法对比研究. 测绘通报, 2019(增2):102-105.
[27] Bendig J, Bolten A, Bennertz S, et al. Estimating biomass of barley using crop surface models (CSMs) derived from UAV- based RGB imaging. Remote Sensing, 2014, 6(11):10395-10412.
doi: 10.3390/rs61110395
[28] 毛智慧, 邓磊, 赵晓明, 等. 利用无人机遥感提取育种小区玉米倒伏信息. 中国农学通报, 2019, 35(3):62-68.
[29] 赵立成, 段玉林, 史云, 等. 基于无人机DSM的小麦倒伏识别方法. 中国农业信息, 2019, 31(4):36-42.
[1] 苏利荣, 谭裕模, 秦芳, 李琴, 曾成城, 李忠义, 韦彩会, 董文斌, 梁俊, 何铁光. 绿豆/黑豆压青还田下减量施肥对宿根甘蔗产量和主要农艺性状的影响[J]. 作物杂志, 2022, (6): 105–110
[2] 闫鹏, 董学瑞, 卢霖, 房孟颖, 李毅杰, 王维赞, 董志强. NAA/KT浸种对新植蔗产量、根系发育及抗倒伏能力的影响[J]. 作物杂志, 2022, (4): 99–106
[3] 谭秦亮, 程琴, 潘成列, 朱鹏锦, 李佳慧, 宋奇琦, 农泽梅, 周全光, 庞新华, 吕平. 干旱胁迫对甘蔗新品种桂热2号生理指标的影响[J]. 作物杂志, 2022, (3): 161–167
[4] 程琴, 谭秦亮, 李佳慧, 朱鹏锦, 周全光, 欧克纬, 卢业飞, 吕平, 庞新华. 不同宿根年限甘蔗品种内源激素及酶活性分析[J]. 作物杂志, 2022, (3): 181–186
[5] 刘丽敏, 刘红坚, 李傲梅, 何为中. 光温处理对甘蔗试管苗光合自养生根的影响[J]. 作物杂志, 2022, (2): 153–157
[6] 罗含敏, 熊发前, 丘立杭, 刘菁, 段维兴, 高轶静, 覃夏燕, 吴建明, 李杨瑞, 刘俊仙. 性状相关的分子标记在甘蔗分子育种中的应用研究[J]. 作物杂志, 2022, (2): 35–43
[7] 李佳慧, 程琴, 欧克纬, 谭秦亮, 庞新华, 周全光, 吕平, 宋奇琦, 唐毓玮, 朱鹏锦. 不同蔗区甘蔗品种(系)分蘖性状比较及其对产量和产量构成因子的影响[J]. 作物杂志, 2021, (5): 79–86
[8] 郭强, 马文清, 秦昌鲜, 施泽升, 彭崇, 闭德金, 何洪良, 梁永检, 唐利球. 甘蔗新品系的DTOPSIS法综合评价[J]. 作物杂志, 2021, (4): 32–37
[9] 范业赓, 陈荣发, 闫海锋, 周慧文, 翁梦苓, 黄杏, 罗霆, 周忠凤, 丘立杭, 吴建明. 甘蔗轮作青饲玉米和花生对甘蔗生长和土壤性状的影响[J]. 作物杂志, 2021, (1): 104–111
[10] 徐林, 吴凯朝, 庞天, 邓智年, 张荣华, 黄成丰, 黄海荣, 李毅杰, 刘晓燕, 覃文宪, 王维赞. 促根剂对甘蔗单芽种茎的生长及产量影响[J]. 作物杂志, 2020, (6): 132–136
[11] 纪景纯, 刘建立, 牛玉洁, 宣可凡, 蒋一飞, 邓皓东, 李晓鹏. 基于全波段高光谱的冬小麦生长参数估算方法比较[J]. 作物杂志, 2020, (6): 180–188
[12] 田婷, 张青, 张海东. 无人机遥感在作物监测中的应用研究进展[J]. 作物杂志, 2020, (5): 1–8
[13] 谢金兰, 林丽, 李长宁, 罗霆, 莫璋红. 氮肥减量条件下间作绿豆压青对甘蔗生长及氮代谢的影响[J]. 作物杂志, 2020, (4): 164–169
[14] 王猛, 隋学艳, 梁守真, 侯学会, 梁永全. 利用无人机遥感技术提取农作物植被覆盖度方法研究[J]. 作物杂志, 2020, (3): 177–183
[15] 范业赓,闫海锋,陈荣发,丘立杭,周慧文,黄杏,翁梦苓,吴建明,李杨瑞,韦生满. 甘蔗脱毒种苗第三代种茎不同体积单芽育苗差异及其移栽效果[J]. 作物杂志, 2020, (2): 194–199
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!