基于无人机三维点云的玉米植株自动计数研究

doi:10.3969/j.issn.1004-1524.2022.09.22

浙江农业学报 ›› 2022, Vol. 34 ›› Issue (9): 2032-2042.DOI: 10.3969/j.issn.1004-1524.2022.09.22

基于无人机三维点云的玉米植株自动计数研究

姜友谊¹(), 张成健¹^,²^,³, 韩少宇²^,³, 杨小冬²^,³, 杨贵军²^,³, 杨浩²^,³^,^*()

1.西安科技大学测绘科学与技术学院,陕西西安 710054
2.北京农业信息技术研究中心农业农村部农业遥感机理与定量遥感重点实验室,北京100097
3.国家农业信息化工程技术研究中心,北京100097

收稿日期:2021-03-24 出版日期:2022-09-25 发布日期:2022-09-30
通讯作者: 杨浩
作者简介:*杨浩,E-mail: yangh@nercita.org.cn
姜友谊(1974—),女,陕西西安人,硕士,副教授,主要从事空间数据处理与分析、3S集成与应用研究。E-mail: youyi_jiang1974@163.com
基金资助:
广东省重点领域研发计划(2019B020216001);国家自然科学基金(41972315);国家自然科学基金(41671411)

Automatic counting of maize plants based on unmanned aerial vehicle (UAV) 3D point cloud

JIANG Youyi¹(), ZHANG Chengjian¹^,²^,³, HAN Shaoyu²^,³, YANG Xiaodong²^,³, YANG Guijun²^,³, YANG Hao²^,³^,^*()

1. College of Geomatics, Xi’an University of Science and Technology, Xi’an 710054, China
2. Key Laboratory of Quantitative Remote Sensing in Agriculture, Ministry of Agriculture and Rural Affairs, Beijing Research Center for Information Technology in Agriculture, Beijing 100097, China
3. National Engineering Research Center for Information Technology in Agriculture, Beijing 100097, China

Received:2021-03-24 Online:2022-09-25 Published:2022-09-30
Contact: YANG Hao

摘要/Abstract

摘要：

植株计数是农民、育种专家等在整个作物生长季评估作物生长状况和管理实践的最常用方法之一,可用来进行合理的田间规划以及管理。针对高密度种植试验区高通量获取玉米自动株数方法匮乏的问题,本研究利用无人机遥感平台,获取田间314个不同基因型的玉米高密度育种小区的数码影像和激光雷达(light detection and ranging,LiDAR)点云数据,发展了一种结合玉米三维空间信息的固定窗口局部最大值算法,实现了高密度玉米育种小区成株数的自动检测,并对比了基于此两种不同数据源的检测精度。该方法以冠层高度模型(canopy height model,CHM)中包含的株高信息为基础,以玉米种植株距为固定窗口进行单株玉米种子点检测,并将检测到的种子点与目视解译的玉米位置进行空间匹配来进行精度的评估。结果表明,基于无人机数码影像构建3种空间分辨率CHM的综合检测精度分别为81.30%、83.11%和78.93%;基于无人机LiDAR的综合精度分别为82.33%、88.66%和81.46%;基于两种数据源构建的CHM,均在空间分辨率为0.05 m时,获得最佳的检测精度。此外,当空间分辨率相同时,LiDAR数据检测精度略优于无人机数码影像,无人机数码传感器由于其成本低、易于操作等优势,在大面积、高密度育种小区的玉米高通量单株检测中表现出更大的潜力。本研究实现了对密植玉米育种试验区玉米株数的自动计数,为表型筛选、田间管理和精准估产等提供依据。

关键词: 自动计数, 无人机, 高通量, 局部最大值, 冠层高度模型, 大田育种

Abstract:

Plant counting is one of the most commonly used methods for farmers and breeding experts to evaluate crop growth status and management practices throughout the crop growing season, which can be used for reasonable field planning and management. In view of the lack of high-throughput methods to obtain the number of maize plants in the high-density planting experimental area, this study used the unmanned aerial vehicle (UAV) remote sensing platform to obtain the digital images and light detection and ranging (LiDAR) point cloud data of 314 maize high-density breeding plots with different genotypes in the field, and developed a fixed window local maximum algorithm combined with maize three-dimensional spatial information. The automatic detection of the number of plants in high density maize breeding plot was realized, and the detection accuracy based on the two different data sources was compared. Based on the plant height information in the crop height model (CHM), the method detected the seed points of each maize plant by using local maximum filter with a fixed window size, and then we spatially matched the detected seed points with the position of maize interpreted visually to evaluate the accuracy. The results showed that the comprehensive detection accuracy of three spatial resolutions CHM based on UAV digital images was 81.30%, 83.11% and 78.93% respectively, and the comprehensive accuracy of UAV-LiDAR was 82.33%, 88.66% and 81.46% respectively. The CHM, based on the two different data sources had the best detection accuracy when the spatial resolution was 0.05 m. In addition, when the spatial resolution was the same, LiDAR performsed better than UAV digital images. But when the demand for detection accuracy was not high, the digital sensor showed greater potential in field management because of its cheap price and easy to operate. This study realized the automatic counting of maize plant number in dense planting maize breeding experimental area, which provided a basis for phenotypic screening, field management and accurate yield estimation.

Key words: automatic counting, unmanned aerial vehicle, high-throughput, local maximum, canopy height model, field breeding

中图分类号:

S252

姜友谊, 张成健, 韩少宇, 杨小冬, 杨贵军, 杨浩. 基于无人机三维点云的玉米植株自动计数研究[J]. 浙江农业学报, 2022, 34(9): 2032-2042.

JIANG Youyi, ZHANG Chengjian, HAN Shaoyu, YANG Xiaodong, YANG Guijun, YANG Hao. Automatic counting of maize plants based on unmanned aerial vehicle (UAV) 3D point cloud[J]. Acta Agriculturae Zhejiangensis, 2022, 34(9): 2032-2042.

图/表 11

参考文献 27

[1]	PANG Y, SHI Y Y, GAO S C, et al. Improved crop row detection with deep neural network for early-season maize stand count in UAV imagery[J]. Computers and Electronics in Agriculture, 2020, 178: 105766. DOI URL
[2]	CHEN R Z, CHU T X, LANDIVAR J A, et al. Monitoring cotton (Gossypium hirsutum L.) germination using ultrahigh-resolution UAS images[J]. Precision Agriculture, 2018, 19(1): 161-177. DOI URL
[3]	GNÄDINGER F, SCHMIDHALTER U. Digital counts of maize plants by unmanned aerial vehicles (UAVs)[J]. Remote Sensing, 2017, 9(6): 544. DOI URL
[4]	VARELA S, DHODDA P, HSU W, et al. Early-season stand count determination in corn via integration of imagery from unmanned aerial systems (UAS) and supervised learning techniques[J]. Remote Sensing, 2018, 10(3): 343. DOI URL
[5]	SHIRZADIFAR A, MAHARLOOEI M, BAJWA S G, et al. Mapping crop stand count and planting uniformity using high resolution imagery in a maize crop[J]. Biosystems Engineering, 2020, 200: 377-390. DOI URL
[6]	SHUAI G Y, MARTINEZ-FERIA R A, ZHANG J S, et al. Capturing maize stand heterogeneity across yield-stability zones using unmanned aerial vehicles (UAV)[J]. Sensors, 2019, 19(20): 4446. DOI URL
[7]	POSTMA J A, HECHT V L, HIKOSAKA K, et al. Dividing the pie: a quantitative review on plant density responses[J]. Plant, Cell & Environment, 2021, 44(4): 1072-1094.
[8]	明博, 谢瑞芝, 侯鹏, 等. 2005—2016年中国玉米种植密度变化分析[J]. 中国农业科学, 2017, 50(11): 1960-1972.
	MING B, XIE R Z, HOU P, et al. Changes of maize planting density in China[J]. Scientia Agricultura Sinica, 2017, 50(11): 1960-1972. (in Chinese with English abstract)
[9]	ZOU H W, LU H, LI Y N, et al. Maize tassels detection: a benchmark of the state of the art[J]. Plant Methods, 2020, 16: 108. DOI PMID
[10]	ZHAO B Q, ZHANG J, YANG C H, et al. Rapeseed seedling stand counting and seeding performance evaluation at two early growth stages based on unmanned aerial vehicle imagery[J]. Frontiers in Plant Science, 2018, 9: 1362. DOI PMID
[11]	ZHAO H X, WANG Y X, SUN Z B, et al. Failure detection in Eucalyptus plantation based on UAV images[J]. Forests, 2021, 12(9): 1250. DOI URL
[12]	MATIAS F I, CARAZA-HARTER M V, ENDELMAN J B. FIELDimageR: an R package to analyze orthomosaic images from agricultural field trials[J]. The Plant Phenome Journal, 2020, 3(1): e20005.
[13]	JIN X L, LIU S Y, BARET F, et al. Estimates of plant density of wheat crops at emergence from very low altitude UAV imagery[J]. Remote Sensing of Environment, 2017, 198: 105-114. DOI URL
[14]	ZHANG J, ZHAO B Q, YANG C H, et al. Rapeseed stand count estimation at leaf development stages with UAV imagery and convolutional neural networks[J]. Frontiers in Plant Science, 2020, 11: 617. DOI PMID
[15]	郑晓岚, 张显峰, 程俊毅, 等. 利用无人机多光谱影像数据构建棉苗株数估算模型[J]. 中国图象图形学报, 2020, 25(3): 520-534.
	ZHENG X L, ZHANG X F, CHENG J Y, et al. Using the multispectral image data acquired by unmanned aerial vehicle to build an estimation model of the number of seedling stage cotton plants[J]. Journal of Image and Graphics, 2020, 25(3): 520-534. (in Chinese with English abstract)
[16]	赵必权, 丁幼春, 蔡晓斌, 等. 基于低空无人机遥感技术的油菜机械直播苗期株数识别[J]. 农业工程学报, 2017, 33(19): 115-123.
	ZHAO B Q, DING Y C, CAI X B, et al. Seedlings number identification of rape planter based on low altitude unmanned aerial vehicles remote sensing technology[J]. Transactions of the Chinese Society of Agricultural Engineering, 2017, 33(19): 115-123. (in Chinese with English abstract)
[17]	PICOS J, BASTOS G, MÍGUEZ D, et al. Individual tree detection in a Eucalyptus plantation using unmanned aerial vehicle (UAV)-LiDAR[J]. Remote Sensing, 2020, 12(5): 885. DOI URL
[18]	YIN D M, WANG L. Individual mangrove tree measurement using UAV-based LiDAR data: possibilities and challenges[J]. Remote Sensing of Environment, 2019, 223: 34-49. DOI URL
[19]	李丹, 张俊杰, 赵梦溪. 基于FCM和分水岭算法的无人机影像中林分因子提取[J]. 林业科学, 2019, 55(5): 180-187.
	LI D, ZHANG J J, ZHAO M X. Extraction of stand factors in UAV image based on FCM and watershed algorithm[J]. Scientia Silvae Sinicae, 2019, 55(5): 180-187. (in Chinese with English abstract)
[20]	胡馨月, 倪海明, 戚大伟. 基于无人机影像的树木株数提取[J]. 森林工程, 2021, 37(1): 6-12.
	HU X Y, NI H M, QI D W. Tree counts extraction based on UAV imagery[J]. Forest Engineering, 2021, 37(1): 6-12. (in Chinese with English abstract)
[21]	YIN T, ZENG J, ZHANG X L, et al. Individual tree parameters estimation for Chinese fir (Cunninghamia lanceolate(Lamb.) Hook) plantations of South China using UAV oblique photography: possibilities and challenges[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 14: 827-842. DOI URL
[22]	李平昊, 申鑫, 代劲松, 等. 机载激光雷达人工林单木分割方法比较和精度分析[J]. 林业科学, 2018, 54(12): 127-136.
	LI P H, SHEN X, DAI J S, et al. Comparisons and accuracy assessments of LiDAR-based tree segmentation approaches in planted forests[J]. Scientia Silvae Sinicae, 2018, 54(12): 127-136. (in Chinese with English abstract)
[23]	ZHANG W M, QI J B, WAN P, et al. An easy-to-use airborne LiDAR data filtering method based on cloth simulation[J]. Remote Sensing, 2016, 8(6): 501. DOI URL
[24]	刘晓双, 黄建文, 鞠洪波. 高空间分辨率遥感的单木树冠自动提取方法与应用[J]. 浙江林学院学报, 2010, 27(1): 126-133.
	LIU X S, HUANG J W, JU H B. Research progress in the methods and applications of individual tree crown’s automatic extraction by high spatial resolution remote sensing[J]. Journal of Zhejiang Forestry College, 2010, 27(1): 126-133. (in Chinese with English abstract)
[25]	POULIOT D A, KING D J, BELL F W, et al. Automated tree crown detection and delineation in high-resolution digital camera imagery of coniferous forest regeneration[J]. Remote Sensing of Environment, 2002, 82(2/3): 322-334. DOI URL
[26]	WALSWORTH N A, KING D J. Image modelling of forest changes associated with acid mine drainage[J]. Computers & Geosciences, 1999, 25(5): 567-580. DOI URL
[27]	CULVENOR D S. TIDA: an algorithm for the delineation of tree crowns in high spatial resolution remotely sensed imagery[J]. Computers & Geosciences, 2002, 28(1): 33-44. DOI URL

项目Item	无人机数码UAV digital	无人机LiDAR UAV-LiDAR
无人机型号UAV model	DJI精灵4 Pro DJI Phantom 4 Pro	DJI M600
传感器及产地 Sensor and origin	1英寸CMOS/中国 1 inch CMOS/China	Riegl VUX-1/奥地利 Riegl VUX-1/Austria
像素Pixel	2×10⁷	-
测量频率/激光发射频率 Measurement frequency/laser emission frequency	前/后/下视:10/10/20 Hz Front/back/bottom view: 10/10/20 Hz	550 kHz
拍摄模式/扫描方式 Shooting mode/scanning mode	单张拍摄/多张连拍 Single shot/multiple consecutive shots	摆锤式 Pendulum type
快门速度/扫描速度Shutter speed/scanning speed	8-1/2000 s	220 Scan/s
照片尺寸/激光光斑直径Photo size/laser spot diameter	3∶2/4∶3/16∶9	0.007 5 m
是否可拆卸Detachable or not	否Not	是Yes
价格Price	9 999 yuan	1×10⁶ yuan

项目Item	无人机数码UAV digital	无人机LiDAR UAV-LiDAR
无人机型号UAV model	DJI精灵4 Pro DJI Phantom 4 Pro	DJI M600
传感器及产地 Sensor and origin	1英寸CMOS/中国 1 inch CMOS/China	Riegl VUX-1/奥地利 Riegl VUX-1/Austria
像素Pixel	2×10⁷	-
测量频率/激光发射频率 Measurement frequency/laser emission frequency	前/后/下视:10/10/20 Hz Front/back/bottom view: 10/10/20 Hz	550 kHz
拍摄模式/扫描方式 Shooting mode/scanning mode	单张拍摄/多张连拍 Single shot/multiple consecutive shots	摆锤式 Pendulum type
快门速度/扫描速度Shutter speed/scanning speed	8-1/2000 s	220 Scan/s
照片尺寸/激光光斑直径Photo size/laser spot diameter	3∶2/4∶3/16∶9	0.007 5 m
是否可拆卸Detachable or not	否Not	是Yes
价格Price	9 999 yuan	1×10⁶ yuan

指标 Measurement index	指标具体含义 Specific meaning of index
TP	被正确检测的玉米 Maize that has been correctly detected
FP	非玉米或部分玉米植株被视做整株检测出 Non-maize or part of maize plants are considered to be detected as whole plants
FN	未被检测到的玉米Undetected maize

指标 Measurement index	指标具体含义 Specific meaning of index
TP	被正确检测的玉米 Maize that has been correctly detected
FP	非玉米或部分玉米植株被视做整株检测出 Non-maize or part of maize plants are considered to be detected as whole plants
FN	未被检测到的玉米Undetected maize

样点 Sampling point	0.03 m			0.05 m			0.10 m
样点 Sampling point	正检 Correct detection	漏检 Missed detection	错检 Error detection	正检 Correct detection	漏检 Missed detection	错检 Error detection	正检 Correct detection	漏检 Missed detection	错检 Error detection
A	117	30	10	111	33	13	114	30	13
B	121	37	17	136	25	14	105	65	5
C	118	53	6	127	41	9	111	57	9
D	129	63	2	132	59	3	127	59	8
E	117	52	7	119	52	5	116	52	8
总计Total	602	235	42	625	210	44	573	263	43

基于无人机三维点云的玉米植株自动计数研究

Automatic counting of maize plants based on unmanned aerial vehicle (UAV) 3D point cloud

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 27

相关文章 15

编辑推荐

Metrics

本文评价

[1]	吕倩, 骆巧, 罗雪, 陈久兵, 马莉, 罗正中, 姚学萍, 余树民, 沈留红, 曹随忠. 基于高通量测序技术分析奶牛场垫沙和橡胶垫卧床中的菌群差异[J]. 浙江农业学报, 2022, 34(7): 1377-1385.
[2]	王钧, 陆洲, 罗明, 徐飞飞, 张序. 基于机载多光谱的冬小麦返青期土壤墒情反演[J]. 浙江农业学报, 2022, 34(6): 1297-1305.
[3]	邬宁珊, 王佳希, 张岩, 元慕田, 张琪, 高驰宇. 基于无人机可见光影像的树种和树冠信息提取——以晋西黄土区蔡家川流域为例[J]. 浙江农业学报, 2021, 33(8): 1505-1518.
[4]	涂藤, 尹清清, 张鹏飞, 王印, 杨泽晓, 姚学萍, 罗燕. 基于毛细管电泳的7种猪源性疫病的多重PCR检测方法的建立[J]. 浙江农业学报, 2021, 33(4): 618-631.
[5]	徐民民, 黄莹, 李波, 徐艳, 张帅, 姚岭芸, 王政. 生物炭对小麦根际和根内微生物群落结构的影响[J]. 浙江农业学报, 2021, 33(3): 516-525.
[6]	隋夕然, 王妍, 刘云根, 张雅洁, 吴丽芳. 典型喀斯特区云南松林土壤养分和细菌群落对海拔的响应[J]. 浙江农业学报, 2021, 33(12): 2348-2357.
[7]	陈乾丽, 汪汉成, 梁永进, 蔡刘体, 黄宇, 周浩, 李忠, 韩洁. 烤后健康烟叶和霉烂烟叶真菌群落结构分析[J]. 浙江农业学报, 2020, 32(6): 1019-1028.
[8]	尹林江, 周忠发, 黄登红, 尚梦佳. 基于无人机影像匹配点云数据的喀斯特峡谷区火龙果单株提取研究[J]. 浙江农业学报, 2020, 32(6): 1092-1102.
[9]	杨移斌, 艾晓辉, 宋怿, 董靖, 胥宁, 姜兰. 黄颡鱼溶血性腹水病初探[J]. 浙江农业学报, 2019, 31(8): 1239-1248.
[10]	刘志, 贺正, 苗芳芳, 贾彪. 基于无人机的水肥一体化玉米出苗率估算方法与试验[J]. 浙江农业学报, 2019, 31(6): 977-985.
[11]	梁长江, 吴雪梅, 王芳, 宋朱军, 张富贵. 基于无人机的田间地膜识别算法研究[J]. 浙江农业学报, 2019, 31(6): 1005-1011.
[12]	张博伟, 王海静, 宋茂勇, 谢慧君, 张建. 四溴双酚A对土壤无机氮转化的影响及其微生物学机理初探[J]. 浙江农业学报, 2019, 31(4): 639-645.
[13]	倪黎纲, 赵旭庭, 王宵燕, 宋成义, 吴信生, 甘源. 基于高通量测序分析肺炎支原体感染姜曲海猪肺组织miRNA表达谱[J]. 浙江农业学报, 2019, 31(12): 1979-1986.
[14]	徐晓锋, 郭成. 淀粉诱导奶牛乳脂下降后奶牛瘤胃细菌菌群变化[J]. 浙江农业学报, 2019, 31(10): 1591-1598.
[15]	张爱菊, 郝雅宾, 郭爱环, 刘金殿, 练青平, 周志明. 基于高通量测序技术的鱼类环境DNA研究中通用引物的筛选验证[J]. 浙江农业学报, 2019, 31(10): 1615-1623.

分辨率 Resolution/m	R_RGB	P_RGB	F_RGB
0.03	71.92	93.48	81.30
0.05	74.85	93.49	83.11
0.10	68.54	93.02	78.93

分辨率 Resolution/m	R_RGB	P_RGB	F_RGB
0.03	71.92	93.48	81.30
0.05	74.85	93.49	83.11
0.10	68.54	93.02	78.93

分辨率 Resolution/m	R_LiDAR	P_LiDAR	F_LiDAR
0.03	77.46	87.86	82.33
0.05	82.94	95.24	88.66
0.10	73.48	91.38	81.46

分辨率 Resolution/m	R_LiDAR	P_LiDAR	F_LiDAR
0.03	77.46	87.86	82.33
0.05	82.94	95.24	88.66
0.10	73.48	91.38	81.46