Rape Identification at Seedling Stage Based on UAV RGB Image

doi:10.13304/j.nykjdb.2021.0913

Abstract

Abstract:

UAV visible light remote sensing has the advantages of low economic cost and flexible take-off and landing mode， and has good application prospect in monitoring rape seedling shortage and plant quantity calculation. The visible light images of rape planting area were obtained by DJI phantom 4 Pro v2.0 UAV with 4 rotors. 1 experimental area and 4 verification areas were set up to compare and analyze DN values in red（R），green（G），blue（B） bands of soil， rape seedlings， weeds and other ground objects. According to the normal distribution characteristics and ratio relationship of DN values in each band， the green-blue-red difference index （GBRDI） was constructed. the GBRDI was used to calculate the visible light image of rape planting area. The histogram of different ground objects was drawn and the intersection point was as the image segmentation threshold to extract rape seedlings and plants； finally， the extraction results of the GBRDI was compared with those of common excess green index （ExG）， visible band difference vegetation index （VDVI） and normalized green blue difference index （NGBDI）. The results showed that： compared with ExG， VDVI and NGBDI， GBRDI had better effect on extracting rape seedlings from weeds and other background features， with accuracy of 92.93% and integrity of 83.63%； in the 4 verification areas with complex weed growth， the accuracy and integrity of extracting rape seedlings by GBRDI index were higher than those of VDVI， ExG and NGBDI. Above results showed that GBRDI could quickly and accurately extract rape seedlings and separate weeds， soil and other background features， which provided technical reference for rape precision farming.

Key words: UAV remote sensing, visible light image, GBRDI, rape seedling identification, weed identification

摘要：

无人机可见光遥感具有经济成本低、起降方式灵活等优点，在监测苗期油菜缺苗和植株数量方面具有较好的应用前景。采用大疆Phantom 4 Pro V2.0四旋翼无人机获取油菜种植区的可见光影像，设置1个实验区和4个验证区，对比分析土壤、油菜苗、杂草等地物目标红（R）、绿（G）、蓝（B）波段的DN值，根据各波段DN值的正态分布特性和比值关系构建绿蓝红差异指数（green-blue-red difference index， GBRDI）。利用GBRDI对油菜种植区可见光影像进行计算，绘制不同地物的直方图，将其交点作为图像分割阈值对油菜苗植株进行提取，并与常见的过绿指数（excess green， ExG）、差异植被指数（visible-band difference vegetation index， VDVI）、归一化绿蓝差异指数（normalized green-bule difference index， NGBDI）的提取结果进行对比。结果表明：相较于ExG、VDVI、NGBDI，GBRDI从杂草等背景地物中提取油菜苗效果更好，精度92.93%，完整性83.63%；在4个杂草长势复杂的验证区，GBRDI指数提取油菜苗的精度和完整性均比VDVI、ExG、NGBDI更高；GBRDI能快速准确地提取油菜苗，分离杂草和土壤等背景地物，可为油菜精细化耕作提供技术参考。

关键词: 无人机遥感, 可见光影像, GBRDI指数, 油菜苗识别, 杂草识别

CLC Number:

S127

Lingwen HU, Zhongfa ZHOU, Linjiang YIN, Meng ZHU, Denghong HUANG. Rape Identification at Seedling Stage Based on UAV RGB Image[J]. Journal of Agricultural Science and Technology, 2022, 24(9): 116-128.

胡灵炆, 周忠发, 尹林江, 朱孟, 黄登红. 基于无人机RGB影像的苗期油菜识别[J]. 中国农业科技导报, 2022, 24(9): 116-128.

Figures/Tables 14

References 42

1	冷博峰,李先容,陈雪婷,等. 2008—2019年中国油菜生产性状变化趋势[J].中国油料作物学报,2021, 43(2): 171-185.
	LENG B F, LI X R, CHEN X T, et al. Change trend of rape production characters in China from 2008 to 2019 [J]. Oilseed Crops China, 2021, 43(2): 171-185.
2	刘成,冯中朝,肖唐华,等.我国油菜产业发展现状、潜力及对策[J].中国油料作物学报,2019, 41(4): 485-489.
	LIU C, FENG Z C, XIAO T H, et al.. Development status, potential and countermeasures of rape industry in China [J]. Oilseed Crops China, 2019, 41(4): 485-489.
3	李中元,吴炳方,张淼,等.利用物候差异与面向对象决策树提取油菜苗种植面积[J].地球信息科学学报, 2019, 21(5): 720-730.
	LI Z Y, WU B F, ZHANG M, et al.. Extraction of rape seedling planting area by phenological difference and object-oriented decision tree [J]. J. Earth Inform. Sci., 2019, 21(5): 720-730.
4	王东,方圣辉,王政.基于光谱特征和颜色特征的油菜提取研究[J].农业机械学报, 2018, 49(3): 158-165.
	WANG D, FANG S H, WANG Z. Research on rape extraction based on spectral and color features [J]. J. Agric. Mach., 2018, 49(3): 158-165..
5	WEI C, HUANG J, Lamin M, et al.. Estimation and mapping of winter oilseed rape LAI from high spatial resolution satellite data based on a hybrid method [J/OL]. Remote Sens., 2017, 2017(5): 488 [2022-08-09]. .
6	邹长慧,周忠发.喀斯特高原山区无人机低空遥感影像数据的获取与处理[J].测绘通报, 2012(): 421-423.
	ZOU C H, ZHOU Z F. Acquisition and processing of UAV low altitude remote sensing image data in Karst Plateau Mountain area [J]. Survey Map Bull., 2012 (Supp.1): 421-423.
7	黄登红,周忠发,吴跃,等.基于无人机可见光影像的高原丘陵盆地区山药植株识别[J].热带地理, 2019, 39(4): 571-582.
	HUANG D H, ZHOU Z F, WU Y, et al.. Identification of yam plants in Plateau Hilly Basin based on UAV visible light image [J]. Tropical Geography, 2019, 39(4): 571-582.
8	董金玮,吴文斌,黄健熙.农业土地利用遥感信息提取的研究进展与展望[J].地球信息科学学报,2020, 22(4): 772-783.
	DONG J W, WU W B, HUANG J X. Research progress and Prospect of remote sensing information extraction of agricultural land use [J]. J. Earth Inform. Sci., 2020, 22(4): 772-783.
9	汪沛,罗锡文,周志艳,等.基于微小型无人机的遥感信息获取关键技术综述[J].农业工程学报, 2014, 30(18): 1-12.
	WANG P, LUO X W, ZHOU Z Y, et al.. Key technologies of remote sensing information acquisition based on micro-UAV [J]. J. Agric. Eng., 2014, 30(18): 1-12.
10	林娜,陈宏,赵健,等.轻小型无人机遥感在精准农业中的应用及展望[J].江苏农业科学, 2020, 48(20): 43-48.
	LIN N, CHEN H, ZHAO J, et al.. Application and prospect of light and small UAV remote sensing in precision agriculture [J]. Jiangsu Agric. Sci., 2020, 48(20): 43-48.
11	毕凯,李英成,丁晓波,等.轻小型无人机航摄技术现状及发展趋势[J].测绘通报, 2015(3): 27-31.
	BI K, LI Y C, DING X B, et al.. Status quo and development trend of aerial photography technology for light and small UAV [J]. Survey. Map Bull., 2015 (3): 27-31.
12	COMB L, BIGLIA A, AIMONINO D R, et al.. Unsupervised detection of vineyards by 3D point-cloud UAV photogrammetry for precision agriculture [J]. Comp. Electron. Agric., 2018, 155: 84-95.
13	尹林江,周忠发,黄登红,等.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.
	YIN L J, ZHOU Z F, HUANG D H, et al.. Single plant extraction of Pitaya in Karst Canyon Area Based on UAV image matching point cloud data [J]. Zhejiang Agric. J., 2020, 32(6): 1092-1102.
14	HYYPPÄ E, HYYPPÄ J, HAKALA T, et al.. Under-canopy UAV laser scanning for accurate forest field measurements [J]. ISPRS J. Photogram. Remote Sens., 2020, 164: 41-60.
15	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].农业工程学报,2017, 33(13): 112-119.
	GUO P, WU F D, DAI J G, et al.. Comparison of crop classification methods based on UAV visible light image [J]. J. Agric. Eng., 2017, 33(13): 112-119.
16	LI B, XU X, HAN J, et al.. The estimation of crop emergence in potatoes by UAV RGB imagery [J]. Plant Methods, 2019, 15(1): 1-13.
17	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].农业工程学报, 2020, 36(4): 63-71.
	DAI J G, XUE J L, ZHAO Q Z, et al.. Extracting cotton seedling information from UAV visible light remote sensing image [J]. J. Agric. Eng., 2020, 36(4): 63-71.
18	高永刚,林悦欢,温小乐,等.基于无人机影像的可见光波段植被信息识别[J].农业工程学报, 2020, 36(3): 178-189.
	GAO Y G, LIN Y H, WEN X L, et al.. Identification of vegetation information in visible light band based on UAV image [J]. J. Agric. Eng., 2020, 36(3): 178-189.
19	黄登红,周忠发,彭睿文,等.西南高原山区作物低空遥感挑战与研究进展——以贵州省为例[J].贵州师范大学学报(自然科学版), 2021, 39(5): 53-61.
	HUANG D H, ZHOU Z F, PENG R W, et al. Challenges and research progress of crop low altitude remote sensing in mountainous areas of southwest plateau—Taking Guizhou Province as an example [J]. J. Guizhou Norm. Univ. (Nat. Sci.), 2021, 39(5): 53-61.
20	任明强, 张家德,卢正艳.贵州喀斯特与非喀斯特农业生态地质环境质量对比研究[J].中国岩溶,2009, 28(4): 397-401.
	REN M Q, ZHANG J D, LU Z Y. Comparative study on agroecological geology environment quality between karst and non karst in Guizhou [J]. Karst China, 2009, 28(4): 397-401.
21	何腾兵.贵州喀斯特山区水土流失状况及生态农业建设途径探讨[J].水土保持学报, 2000, 14(Z1): 28-34.
	HE T B. Soil and water loss in karst mountainous area of Guizhou Province and approach to ecological agriculture construction [J]. J. Soil Water Conserv., 2000, 14(Z1): 28-34.
22	高雯晗.无人机载遥感数据估测油菜长势参数方法研究[D].武汉:华中农业大学,2019.
	GAO W H. Research on method of estimating oilseed rape growth parameters based on UAV borne remote sensing data [D]. Wuhan: Huazhong Agricultural University, 2019.
23	韩文霆,李广,苑梦婵,等.基于无人机遥感技术的玉米种植信息提取方法研究[J].农业机械学报,2017, 48(1): 139-147.
	HAN W T, LI G, YUAN M C, et al.. Research on corn planting information extraction method based on UAV remote sensing technology [J]. J. Agric. Mach., 2017, 48(1): 139-147.
24	WOEBBECKE D M, MEYER G E, VON BARGEN K, MORTENSEN D A. Color indices for weed identification under various soil, residue, and lighting conditions [J]. Trans. ASAE, 1995, 38(1): 259-269.
25	HUNT E R, CAVIGELLI M, DAUGHTRY C S T, et al.. Evaluation of digital photography from model aircraft for remote sensing of crop biomass and nitrogen status [J]. Precision Agric., 2005, 6(4): 359-378.
26	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.
	WANG X Q, WANG M M, WANG S Q, et al.. Vegetation information extraction based on visible band UAV remote sensing [J]. J. Agric. Eng., 2015, 31(5): 152-159.
27	VERRELST, SCHAEPMANM.E., KOETZB.,M, et al.. Angula sensitivity analysis of vegetation indices derived from CHRIS/PROBA data [J]. Remote Sens. Environ., 2008, 112(5): 2341-2353.
28	王猛,隋学艳,梁守真,等.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.
	WANG M, SUI X Y, LIANG S Z, et al.. Extraction of crop vegetation coverage using UAV remote sensing technology [J]. Crops, 2020(3): 177-183.
29	DU M, NOBORU N. Monitoring of wheat growth status and mapping of wheat yield’s within-field spatial variations using color images acquired from UAV-camera system [J/OL]. Remote Sens., 2017, 9(3):289[2022-08-09]. .
30	SHUFELT J A. Performance evaluation and analysis of monocular building extraction from aerial imagery [J]. IEEE Trans. Pattern Anal. Mach. Intell.,1999,21(4): 311-326.
31	李青,李玉,王玉,等.利用格式塔的高分辨率遥感影像建筑物提取[J].中国图象图形学报, 2017, 22(8): 1162-1174.
	LI Q, LI Y, WANG Y, et al.. Building extraction from high resolution remote sensing image using Gestalt [J]. Chin. J. Image Graphics, 2017, 22(8): 1162-1174.
32	GÓMEZ-DÉNIZ E, SARABIA J M, CALDERín-OJEDA E. Bimodal normal distribution: extensions and applications [J/OL]. J. Compu. Appl. Mathematics, 2021, 388: 113292 [2022-08-09]. .
33	TOMASZ G, LAJOS H, PIOTR K. Tests of normality of functional data [J]. Int. Stat. Rev., 2020, 88(3): 677-697.
34	YANG C H. Application of remote sensing and precision agriculture technology in crop disease detection and management [J]. Engineering, 2020, 6(5): 102-112.
35	刘建刚,赵春江,杨贵军,等.无人机遥感解析田间作物表型信息研究进展[J].农业工程学报,2016, 32(24): 98-106.
	LIU J G, ZHAO C J, YANG G J, et al.. Research progress of UAV Remote Sensing Analysis of field crop phenotypic information [J]. J. Agric. Eng., 2016, 32(24): 98-106.
36	王正兴,刘闯, HUETE Alfredo. 植被指数研究进展:从AVHRR-NDVI到MODIS-EVI[J].生态学报, 2003, 23(5): 979-987.
	WANG Z X, LIU C, HUETE A. Research progress of vegetation index: from AVHRR-NDVI to MODIS-EVI [J]. Ecol. Learned J., 2003, 23(5): 979-987.
37	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.
	ZHAO J, YANG H B, LAN Y B, et al.. Extraction method of summer maize vegetation coverage based on UAV visible light image [J]. J. Agric. Mach., 2019, 50(5): 232-240.
38	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.
	LONG M S, HE D J. Computer identification of weeds in maize seedling [J]. J. Agric. Eng., 2007, 23(7): 139-144.
39	周涛,胡振琪,韩佳政,等.基于无人机可见光影像的绿色植被提取[J].中国环境科学, 2021, 41(5): 2380-2390.
	ZHOU T, HU Z Q Han J Z, et al.. Green vegetation extraction based on UAV visible light image [J]. China Environ. Sci., 2021, 41(5): 2380-2390.
40	尹林江,周忠发,李韶慧,等.基于无人机可见光影像对喀斯特地区植被信息提取与覆盖度研究[J].草地学报, 2020, 28(6): 1664-1672.
	YIN L J, ZHOU Z F, LI S H, et al.. Vegetation information extraction and coverage in Karst Area based on UAV visible light image [J]. Grassland Learned J., 2020, 28(6): 1664-1672.
41	李冰,刘镕源,刘素红,等.基于低空无人机遥感的冬小麦覆盖度变化监测[J].农业工程学报, 2012, 28(13): 160-165.
	LI B, LIU R Y, LIU S H, et al.. Monitoring of winter wheat coverage change based on low altitude UAV Remote Sensing [J]. J. Agric. Eng., 2012, 28(13): 160-165.
42	田振坤,傅莺莺,刘素红,等.基于无人机低空遥感的农作物快速分类方法[J].农业工程学报, 2013, 29(7): 109-116.
	TIAN Z K, FU Y Y, LIU S H, et al.. Rapid crop classification method based on UAV low altitude remote sensing [J]. J. Agric. Eng., 2013, 29(7): 109-116.

名称 Name	平均值 Mean value			标准差 Standard deviation
名称 Name	R	G	B	R	G	B
土壤 Soil	181.89	168.83	142.51	11.30	14.31	12.33
油菜苗 Rape	160.62	210.22	179.33	11.76	11.90	13.60
杂草 Weeds	156.24	182.16	135.85	18.91	24.45	17.31

名称 Name	平均值 Mean value			标准差 Standard deviation
名称 Name	R	G	B	R	G	B
土壤 Soil	181.89	168.83	142.51	11.30	14.31	12.33
油菜苗 Rape	160.62	210.22	179.33	11.76	11.90	13.60
杂草 Weeds	156.24	182.16	135.85	18.91	24.45	17.31

指数 Index	土壤Soil		油菜苗Rape		杂草Weeds
指数 Index	均值 Mean	标准差 Standard deviation	均值 Mean	标准差 Standard deviation	均值 Mean	标准差 Standard deviation
GBRDI	197.54	20.82	311.93	19.27	232.77	33.02
ExG	13.33	14.99	80.40	13.11	94.62	25.10
VDVI	0.02	0.02	0.11	0.02	0.15	0.03
NGBDI	0.08	0.03	0.08	0.02	0.18	0.03

指数 Index	土壤Soil		油菜苗Rape		杂草Weeds
指数 Index	均值 Mean	标准差 Standard deviation	均值 Mean	标准差 Standard deviation	均值 Mean	标准差 Standard deviation
GBRDI	197.54	20.82	311.93	19.27	232.77	33.02
ExG	13.33	14.99	80.40	13.11	94.62	25.10
VDVI	0.02	0.02	0.11	0.02	0.15	0.03
NGBDI	0.08	0.03	0.08	0.02	0.18	0.03

指数 Index	正确提取 TP	误提取 FP	漏提取 FN	真实株数 TN	分支因子BF	检测率 DR/%	完整性 QP/%
GBRDI	710	84	54	849	0.12	92.93	83.63
ExG	668	92	89	849	0.14	88.24	78.68
VDVI	611	136	102	849	0.22	85.69	71.97
NGBDI	17	780	52	849	45.88	24.64	2.00