不同菌草品种叶片叶绿素含量的光谱特征参数模型研究

白妮妮; 刘凤山; 张厚喜; 蔡杨星; 卜建超; 彭实梅; 林冬梅

doi:10.19303/j.issn.1008-0384.2023.02.011

不同菌草品种叶片叶绿素含量的光谱特征参数模型研究

doi: 10.19303/j.issn.1008-0384.2023.02.011

白妮妮^{1, 2,},
刘凤山^{2, 3, ,},
张厚喜¹,
蔡杨星²,
卜建超^{2, 3},
彭实梅^{2, 3},
林冬梅²

1.
福建农林大学林学院，福建　福州　350002
2.
福建农林大学国家菌草工程技术研究中心，福建　福州　350002
3.
福建农林大学生命科学学院，福建　福州　350002

基金项目: 国家自然科学基金项目（41801020）；国家菌草工程技术研究中心基金项目（CXZX206179）

详细信息

作者简介:
白妮妮（1998−），女，硕士研究生，研究方向：无人机遥感与生态治理（E-mail：bainini0707@163.com）

通讯作者:
刘凤山（1986−），男，博士，助理研究员，硕士研究生导师，研究方向：农业生态（E-mail：liufs.11b@igsnrr.ac.cn）

中图分类号: S127；S543
计量
- 文章访问数: 366
- HTML全文浏览量: 169
- PDF下载量: 32
- 被引次数: 0
出版历程
- 收稿日期: 2022-07-24
- 修回日期: 2022-12-10
- 网络出版日期: 2023-03-28
- 刊出日期: 2023-02-28

Building of JUNCAO Leaf Chlorophyll Spectrum Model

BAI Nini^{1, 2
,},
LIU Fengshan^{2, 3
, ,},
ZHANG Houxi¹,
CAI Yangxing²,
BU Jianchao^{2, 3},
PENG Shimei^{2, 3},
LIN Dongmei²

1.
College of Forestry，Fujian Agriculture and Forestry University，Fuzhou，Fujian　350002, China
2.
National Engineering Research Center of JUNCAO Technology, Fujian Agriculture and Forestry University, Fuzhou, Fujian　350002, China
3.
College of Life Sciences, Fujian Agriculture and Forestry University, Fuzhou, Fujian　350002, China

摘要

摘要: 目的建立各种品种菌草叶片叶绿素光谱模型为快速无损地评估菌草健康状态提供理论依据。方法利用多光谱相机拍摄光谱信息，使用RGB和HSV两种颜色空间系统进行NDSI、RSI植被指数指标的建立，并采用4种回归算法进行叶绿素相对含量即SPAD值和光谱信息建模，选出合适菌草的模型。结果不同品种的菌草SPAD值差异不显著，主要是健康和干枯叶片的差异。叶片SPAD值和光谱反射值具有良好的相关性，RGB颜色系统红波段是最相关的波段，相关性达到−0.79。试验证明红波段的变化在健康叶片和干枯叶片之间更敏感。2种颜色系统综合评价，建模精度及稳定由高到低依次为：包含5个通道的RGB颜色系统，HSV颜色系统，仅包含R、G、B 3个通道的颜色系统。4种反演方法中，反演效果最好依次为随机森林机器学习方法、支持向量机机器学习和逐步线性回归方法、一元线性统计方法。预测SPAD值拟合效果最好的是RGB-NDSI-RF类型，其拟合数R²为0.95、RMSE为3.04、MRE为0.19，验证R²为0.75、RMSE为8.26、MRE为0.76。结论机器学习方法尤其是随机森林适用于菌草建模，可以取得较高的精度，具有高适应和稳定性。预测SPAD值拟合效果最好的是RGB-NDSI-RF类型，其模型可以适用于不同菌草品种的建模，敏感识别叶片的叶绿素含量以评估其健康状态。
- 菌草 /
- 多光谱 /
- 机器学习 /
- HSV颜色系统
Abstract: Objective A model based on the leaf chlorophyll was established to provide a rapid, convenient, and non-invasive means for evaluating and predicting the quality of varieties of JUNCAOs. Methods Using a multispectral camera, spectra of leaf chlorophyll were obtained from different JUNCAOs. NDSI and RSI vegetation indices were established according to the RGB and HSV color space systems. A chlorophyll meter was used to obtain SPAD readings on JUNCAO leaves . Four regression algorithms were applied to build the model to correlate the spectral and SPAD measurements. Results A significant difference on SPAD was observed between healthy and dry or dead leaves, but not between different JUNCAO. A significant correlation was found between the leaf SPAD and spectral reflection, especially, the red band of RGB color showing a coefficient of −0.79. The high sensitivity of the red band in distinguishing healthy from withering leaves allowed an accurate and stable measurements by the color systems that ranked as: 5-channel RGB＞HSV＞3-channel RGB. Among the 4 inversion methods, the rank was: random forest machine learning＞support vector machine learning＞stepwise linear regression＞univariate linear statistical. And the best fitting function in predicting SPAD was the RGB-NDSI-RF type, which had a R² of 0.95 vs. 3.04 for RMSE and 0.19 for MRE and validation R² of 0.75 vs. 8.26 for RMSE and 0.76 for MRE. Conclusion As JUNCAO is becoming increasingly popular for wide applications, a reliable, rapid, non-invasive method to determine the plant quality was much desired. The current leaf chlorophyll spectra-based RGB-NDSI-RF model built by the random forest machine learning method demonstrated an applicability for evaluating varieties of JUNCAO.
- JUNCAO /
- multispectral /
- machine learning /
- HSV color system

HTML全文

图 1 未调整位置的合成波段(a)与调整位置的合成波段(b)

Figure 1. Synthetic bands without (a) and with (b) position adjustment

下载: 全尺寸图片幻灯片

图 2 提取反射值矢量图

Figure 2. Vector diagram of reflectivity extraction

下载: 全尺寸图片幻灯片

图 3 HSV颜色空间模型

Figure 3. HSV color space model

下载: 全尺寸图片幻灯片

图 4 不同品种菌草SPAD实测值

不同颜色三角形符号表示不同种类健康叶片，不同颜色正方形符号表示不同种类干枯叶片。

Figure 4. Measured leaf SPAD of varieties of JUNCAOs

Triangles in different colors indicate healthy leaves from varieties of JUNCAO; squares in different colors, leaves in different degrees of withering from varieties of JUNCAO.

下载: 全尺寸图片幻灯片

图 5 红绿波段误差线(a)及不同波段干湿反射值(b)

不同小写字母表示不同处理间差异显著（P<0.05）。

Figure 5. Error lines in red and green bands (a) and dry and wet reflectance in different bands (b)

Different lowercase letters represent significant differences between treatments （P<0.05）.

下载: 全尺寸图片幻灯片

图 6 预测值散点图

从上到下4种类型分别为HSV-NDSI、HSV-RSI、RGB-NDSI、RGB-RSI。

Figure 6. Scatter plot of predicted values

HSV-NDSI, HSV-RSI, RGB-NDSI, and RGB-RSI types from top to bottom, respectively.

下载: 全尺寸图片幻灯片

图 7 R²正态分布图

RGB为包含了R、G、B、RE、NIR 5个波段信息的颜色系统，rgb为只包含了R、G、B 3个波段的颜色系统。

Figure 7. Normal distribution of R²

RGB: color system containing information on R, G, B, RE, and NIR bands; rgb: color system containing only R, G, and B bands.

下载: 全尺寸图片幻灯片

表 1 不同菌草品种叶片描述性分析

Table 1. Descriptive analysis on leaves of varieties of JUNCAOs

品种Varieties	样本量Sample size	最大值Maximum	最小值Minimum	差值Difference	标准差Standard deviation
串香松叶草 Silphium perfoliatum	5	58.62	43.30	15.32	6.34
菅 Themeda gigantea var. villosa	5	49.10	43.22	5.88	2.56
危地马拉草 Tripsacum laxum Nash	4	45.56	31.68	13.88	5.89
毛花雀稗 Paspalum dilatatum	5	52.90	41.83	11.07	4.91
粽叶芦 Thysanolaena maxima	5	48.32	45.57	2.75	1.15
紫象草 Penmisetum purpureum cv. Red	5	36.25	26.52	9.73	3.94
杂交狼尾草 Pennisetum americanum × P.purpureum CV.23A × N51	5	44.98	28.97	16.01	6.31
棕叶狗尾草 Setaria palmifolia (koen.) Stapf	5	51.60	28.13	23.47	9.83
绿洲三号种名待定	5	52.90	49.30	3.60	1.42
甜根子草 Saccharum spontaneum Linn.	5	47.80	39.38	8.42	3.86
台湾甜象草 Pennset um purpurcum	5	44.20	32.05	12.15	5.03
华南象草 Pennisetum purpureumcv.Huanan	5（干枯）	17.52	8.70	8.82	3.19
华南象草 Pennisetum purpureumcv.Huanan	5	50.77	40.27	10.50	4.82
巨菌草 Penmisetum giganteum	5（干枯）	9.57	4.70	4.87	2.08
巨菌草 Penmisetum giganteum	6（半干枯）	40.22	1.98	38.24	13.84
桂草一号王草 Pennisetum sinese.Guicao-1	5	46.72	36.98	9.74	4.10
桂牧一号杂交象草 Pennisetum purpureumcv.Guimu-1	5	53.62	43.47	10.15	3.85
桂牧一号杂交象草 Pennisetum purpureumcv.Guimu-1	5（干枯）	22.18	14.03	8.15	3.09
芦苇 Phragmites communis	4	42.60	39.52	3.08	1.33
高丹草 Sorghum bicolor×Sorghum sudanense	4	44.60	39.57	5.03	2.31
宽叶雀稗 Paspalum wetsfeteini Hackel	5	49.68	39.94	9.74	4.01

下载: 导出CSV

表 2 光谱特征参数与SPAD皮尔逊相关性

Table 2. Pearson correlation between spectral parameters and SPAD

光谱特征参数名称Name of spectralcharacteristicparameter	相关系数rCorrelationcoefficient	光谱特征参数名称Name of spectralcharacteristicparameter	相关系数rCorrelationcoefficient		光谱特征参数名称Name of spectralcharacteristicparameter	相关系数rCorrelationcoefficient		光谱特征参数名称Name of spectralcharacteristicparameter	相关系数rCorrelationcoefficient
光谱特征参数名称Name of spectralcharacteristicparameter	相关系数rCorrelationcoefficient	光谱特征参数名称Name of spectralcharacteristicparameter	NDSI	RSI	光谱特征参数名称Name of spectralcharacteristicparameter	NDSI	RSI	光谱特征参数名称Name of spectralcharacteristicparameter	NDSI	RSI
R	−0.79**	H_S	0.66**	0.31**	R_RE	−0.76**	−0.78**	RE_B	0.49**	0.45**
G	−0.65**	H_V	0.70**	0.25**	RE_R	0.76**	0.69**	NIR_B	0.48**	0.44**
B	−0.64**	S_V	0.59**	0.35**	NIR_R	0.76**	0.67**	B_NIR	−0.48**	−0.49**
NIR	−0.40**	S_H	0.66**	0.66**	R_NIR	−0.76**	−0.78**	G_NIR	−0.47**	−0.52**
RE	−0.23	V_H	0.70**	0.70**	R_G	−0.63**	−0.64**	NIR_G	0.47**	0.36**
H	0.38**	V_S	0.59**	0.63**	G_R	0.63**	0.57**	B_G	0.09	0.04
S	0.16	R_B	−0.76**	−0.75**	G_RE	−0.50**	−0.54**	G_B	−0.09	−0.17
V	0.73**	B_R	0.76**	0.74**	RE_G	0.50**	0.38**	RE_NIR	0.07	0.10
					B_RE	−0.49**	−0.50**	NIR_RE	−0.07	−0.02
为达到0.05 水平显著性；达到0.01 显著性水平。表中y_x格式的NDSI植被指数表示为(y-x)/(y+x)，如R_B为(R-B)/(R+B)，RSI植被指数表示为y/x，如R_B为R/B。: At 0.05 level of significance; *: at 0.01 level of significance.NDSIvegetation index iny_xformat: (y−x)/(y+x), for example, R_ B is presented as (R-B)/(R+B);RSI*vegetation index iny/x, R_ B is presented as R/B.

下载: 导出CSV

表 3 4种建模方法的训练集和验证集的精度

Table 3. Accuracy of 4 types of training sets and verifications

类型 Type		RGB								HSV
		NDSI				RSI				NDSI				RSI
		LM	GAM	SVM	RF	LM	GAM	SVM	RF	LM	GAM	SVM	RF	LM	GAM	SVM	RF
训练集 Modeling set	R²	0.77	0.87	0.81	0.95	0.88	0.93	0.82	0.95	0.73	0.79	0.81	0.94	0.77	0.84	0.82	0.94
	RMSE	6.76	5.14	6.57	3.04	5.02	3.92	6.02	3.40	7.23	5.53	6.39	3.44	7.23	5.87	6.19	3.46
	MRE	0.40	0.27	0.40	0.19	0.27	0.20	0.35	0.21	0.41	0.14	0.38	0.21	0.43	0.34	0.36	0.20
验证集 Validation set	R²	0.62	0.64	0.76	0.75	0.57	0.59	0.75	0.74	0.56	0.60	0.70	0.67	0.57	0.65	0.74	0.65
	RMSE	10.69	9.18	7.44	8.26	9.60	9.62	6.70	7.07	9.26	8.35	7.16	9.11	9.90	7.69	6.11	7.08
	MRE	0.95	0.25	0.18	0.76	0.83	0.66	0.16	0.22	0.32	0.24	0.24	0.84	0.85	0.68	0.15	0.17

下载: 导出CSV

参考文献(34)

[1]	林占熺, 苏德伟, 林辉, 等. 黄河菌草生态安全屏障建设的研究与应用 [J]. 福建农林大学学报(自然科学版), 2019, 48(6):803−812. LIN Z X, SU D W, LIN H, et al. Constructing an ecological barrier by Juncao along theYellow River and its industrial applicaton [J]. Journal of Fujian Agriculture and Forestry University(Natural Science Edition), 2019, 48(6): 803−812.(in Chinese)
[2]	丁跃, 吴刚, 郭长奎. 植物叶绿素降解机制研究进展 [J]. 生物技术通报, 2016, 32(11):1−9. DING Y, WU G, GUO C K. Research Advance on Chlorophyll Degradation in Plants [J]. Biotechnology bulletin, 2016, 32(11): 1−9.(in Chinese)
[3]	SWAIN K C, JAYASURIYA H P W, SALOKHE V M. Suitability of low-altitude remote sensing images for estimating nitrogen treatment variations in rice cropping for precision agriculture adoption [J]. Journal of Applied Remote Sensing, 2007, 1(1): 013547. doi: 10.1117/1.2824287
[4]	DOU Z G, CUI L J, LI J, et al. Hyperspectral estimation of the chlorophyll content in short-term and long-term restorations of mangrove in Quanzhou Bay Estuary, China [J]. Sustainability, 2018, 10(4): 1127. doi: 10.3390/su10041127
[5]	CARTELAT A, CEROVIC Z G, GOULAS Y, et al. Optically assessed contents of leaf polyphenolics and chlorophyll as indicators of nitrogen deficiency in wheat (Triticum aestivum L. ) [J]. Field Crops Research, 2005, 91(1): 35−49. doi: 10.1016/j.fcr.2004.05.002
[6]	刘嘉政, 王雪峰, 王甜. 基于深度学习的5种树皮纹理图像识别研究 [J]. 北京林业大学学报, 2019, 41(4):146−154. LIU J Z, WANG X F, WANG T. Research on image recognition of five bark texture images based on deep learning [J]. Journal of Beijing Forestry University, 2019, 41(4): 146−154.(in Chinese)
[7]	侯加林, 田林, 李天华, 等. 基于双侧图像识别的大蒜正芽及排种试验台设计与试验 [J]. 农业工程学报, 2020, 36(1):50−58. doi: 10.11975/j.issn.1002-6819.2020.01.006 HOU J L, TIAN L, LI T H, et al. Design and experiment of test bench for garlic bulbil adjustment and seeding based on bilateral image identification [J]. Transactions of the Chinese Society of Agricultural Engineering, 2020, 36(1): 50−58.(in Chinese) doi: 10.11975/j.issn.1002-6819.2020.01.006
[8]	杨爱萍, 张坤, 段里成, 等. 基于RGB与HSV颜色空间的水稻齐穗后叶片SPAD值估测方法研究 [J]. 江西农业学报, 2019, 31(8):106−112. doi: 10.19386/j.cnki.jxnyxb.2019.08.18 YANG A P, ZHANG K, DUAN L C, et al. Estimation method for SPAD value of rice leaves after full heading based on RGB and HSV color space [J]. Acta Agriculturae Jiangxi, 2019, 31(8): 106−112.(in Chinese) doi: 10.19386/j.cnki.jxnyxb.2019.08.18
[9]	GUO Y H, YIN G D, SUN H Y, et al. Scaling effects on chlorophyll content estimations with RGB camera mounted on a UAV platform using machine-learning methods [J]. Sensors (Basel, Switzerland), 2020, 20(18): 5130. doi: 10.3390/s20185130
[10]	ZHANG S M, ZHAO G X, LANG K, et al. Integrated satellite, unmanned aerial vehicle (UAV) and ground inversion of the SPAD of winter wheat in the reviving stage [J]. Sensors (Basel, Switzerland), 2019, 19(7): 1485. doi: 10.3390/s19071485
[11]	ELARAB M, TICLAVILCA A M, TORRES-RUA A F, et al. Estimating chlorophyll with thermal and broadband multispectral high resolution imagery from an unmanned aerial system using relevance vector machines for precision agriculture [J]. International Journal of Applied Earth Observation and Geoinformation, 2015, 43: 32−42. doi: 10.1016/j.jag.2015.03.017
[12]	KANG Y, LENZ-WIEDEMANN V, CHEN X, et al. Estimating leaf chlorophyll of barley at different growth stages using spectral indices to reduce soil background and canopy structure effects [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014, 97: 58−77. doi: 10.1016/j.isprsjprs.2014.08.005
[13]	HUNT E R Jr, DAUGHTRY C S T, EITEL J U H, et al. Remote sensing leaf chlorophyll content using a visible band index [J]. Agronomy Journal, 2011, 103(4): 1090−1099. doi: 10.2134/agronj2010.0395
[14]	KIPP S, MISTELE B, SCHMIDHALTER U. Identification of stay-green and early senescence phenotypes in high-yielding winter wheat, and their relationship to grain yield and grain protein concentration using high-throughput phenotyping techniques [J]. Functional Plant Biology:FPB, 2014, 41(3): 227−235. doi: 10.1071/FP13221
[15]	VOLLMANN J, WALTER H, SATO T, et al. Digital image analysis and chlorophyll metering for phenotyping the effects of nodulation in soybean [J]. Computers and Electronics in Agriculture, 2011, 75(1): 190−195. doi: 10.1016/j.compag.2010.11.003
[16]	LELONG C C D, BURGER P, JUBELIN G, et al. Assessment of unmanned aerial vehicles imagery for quantitative monitoring of wheat crop in small plots [J]. Sensors (Basel, Switzerland), 2008, 8(5): 3557−3585. doi: 10.3390/s8053557
[17]	贺英, 邓磊, 毛智慧, 等. 基于数码相机的玉米冠层SPAD遥感估算 [J]. 中国农业科学, 2018, 51(15):66−77. doi: 10.3864/j.issn.0578-1752.2018.15.005 HE Y, DENG L, MAO Z H, et al. Remote sensing estimation of canopy SPAD value for maize based on digital camera [J]. Scientia Agricultura Sinica, 2018, 51(15): 66−77.(in Chinese) doi: 10.3864/j.issn.0578-1752.2018.15.005
[18]	SOUDANI K, FRANÇOIS C, LE MAIRE G, et al. Comparative analysis of IKONOS, SPOT, and ETM+ data for leaf area index estimation in temperate coniferous and deciduous forest stands [J]. Remote Sensing of Environment, 2006, 102(1/2): 161−175.
[19]	李庆贺, 伍博炜, 杨珺丽, 等. 福建省城市水资源承载力综合评价研究 [J]. 水资源与水工程学报, 2014, 25(4):147−151. doi: 10.11705/j.issn.1672-643X.2014.04.030 LI Q H, WU B W, YANG J L, et al. Comprehensive evaluation on urban water resources carrying capacity in Fujian Province [J]. Journal of Water Resources and Water Engineering, 2014, 25(4): 147−151.(in Chinese) doi: 10.11705/j.issn.1672-643X.2014.04.030
[20]	INOUE Y, PEÑUELAS J, MIYATA A, et al. Normalized difference spectral indices for estimating photosynthetic efficiency and capacity at a canopy scale derived from hyperspectral and CO₂ flux measurements in rice [J]. Remote Sensing of Environment, 2008, 112(1): 156−172. doi: 10.1016/j.rse.2007.04.011
[21]	姚付启. 冬小麦高光谱特征及其生理生态参数估算模型研究[D]. 杨凌: 西北农林科技大学, 2012. YAO F Q. Hyperspectral characteristics and estimating models about physiological ecological parameters of winter wheat[D]. Yangling: Northwest A & F University, 2012. (in Chinese)
[22]	严欣荣, 张美曼, 郑亚雄, 等. 基于Sentinel-2的丛生竹林信息提取方法比较及分布特征 [J]. 生态学杂志, 2020, 39(3):1056−1066. YAN X R, ZHANG M M, ZHENG Y X, et al. Comparison of extraction methods and the distribution characteristics of cluster bamboo forest information based on Sentinel-2 [J]. Chinese Journal of Ecology, 2020, 39(3): 1056−1066.(in Chinese)
[23]	FARIFTEH J, VAN DER MEER F, ATZBERGER C, et al. Quantitative analysis of salt-affected soil reflectance spectra: A comparison of two adaptive methods (PLSR and ANN) [J]. Remote Sensing of Environment, 2007, 110(1): 59−78. doi: 10.1016/j.rse.2007.02.005
[24]	SUI J, QIN Q M, REN H Z, et al. Winter wheat production estimation based on environmental stress factors from satellite observations [J]. Remote Sensing, 2018, 10(6): 962. doi: 10.3390/rs10060962
[25]	TARAJI M, HADDAD P R, AMOS R I J, et al. Error measures in quantitative structure-retention relationships studies [J]. Journal of Chromatography A, 2017, 1524: 298−302. doi: 10.1016/j.chroma.2017.09.050
[26]	YAGER R R, ALAJLAN N. A note on mean absolute deviation [J]. Information Sciences, 2014, 279: 632−641. doi: 10.1016/j.ins.2014.04.016
[27]	JACQUEMOUD S, VERHOEF W, BARET F, et al. PROSPECT+SAIL models: A review of use for vegetation characterization [J]. Remote Sensing of Environment, 2009, 113: S56−S66. doi: 10.1016/j.rse.2008.01.026
[28]	FERET J B, FRANÇOIS C, ASNER G P, et al. PROSPECT-4 and 5: Advances in the leaf optical properties model separating photosynthetic pigments [J]. Remote Sensing of Environment, 2008, 112(6): 3030−3043. doi: 10.1016/j.rse.2008.02.012
[29]	HOEL B O, SOLHAUG K A. Effect of irradiance on chlorophyll estimation with the minolta SPAD-502 leaf chlorophyll meter [J]. Annals of Botany, 1998, 82(3): 389−392. doi: 10.1006/anbo.1998.0683
[30]	UDDLING J, GELANG-ALFREDSSON J, PIIKKI K, et al. Evaluating the relationship between leaf chlorophyll concentration and SPAD-502 chlorophyll meter readings [J]. Photosynthesis Research, 2007, 91(1): 37−46. doi: 10.1007/s11120-006-9077-5
[31]	赵英时. 遥感应用分析原理与方法 [M]. 北京: 科学出版社, 2003.
[32]	韩兆迎, 朱西存, 房贤一, 等. 基于SVM与RF的苹果树冠LAI高光谱估测 [J]. 光谱学与光谱分析, 2016, 36(3):800−805. HAN Z Y, ZHU X C, FANG X Y, et al. Hyperspectral estimation of apple tree canopy LAI based on SVM and RF regression [J]. Spectroscopy and Spectral Analysis, 2016, 36(3): 800−805.(in Chinese)
[33]	MANTERO P, MOSER G, SERPICO S B. Partially Supervised classification of remote sensing images through SVM-based probability density estimation [J]. IEEE Transactions on Geoscience and Remote Sensing, 2005, 43(3): 559−570. doi: 10.1109/TGRS.2004.842022
[34]	VERRELST J, CAMPS-VALLS G, MUÑOZ-MARÍ J, et al. Optical remote sensing and the retrieval of terrestrial vegetation bio-geophysical properties–A review [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2015, 108: 273−290. doi: 10.1016/j.isprsjprs.2015.05.005