Estimation of Crop Biomass Using GF-3 Polarization SAR Data Based on Genetic Algorithm Feature Selection

doi:10.11947/j.JGGS.2020.0413

Abstract

Abstract:

In recent years, Polarization SAR (PolSAR) has been widely used in the filed of crop biomass estimation. However, high dimensional features extracted from PolSAR data will lead to information redundancy which will result in low accuracy and poor transfer ability of the estimation model. Aiming at this problem, we proposed a estimation method of crop biomass based on automatic feature selection method using genetic algorithm (GA). Firstly, the backscattering coefficient, the polarization parameters and texture features were extracted from PolSAR data. Then, these features were automatically pre-selected by GA to obtain the optimal feature subset. Finally, based on this subset, a support vector regression machine (SVR) model was applied to estimate crop biomass. The proposed method was validated using the GaoFen-3 (GF-3) QPSΙ (C-band, quad-polarization) SAR data. Based on wheat and rape biomass samples acquired from a synchronous field measurement campaign, the proposed method achieve relative high validation accuracy (over 80%) in both crop types. For further analyzing the improvement of proposed method, validation accuracies of biomass estimation models based on several different feature selection methods were compared. Compared with feature selection based on linear correlation, GA method has increased by 5.77% in wheat biomass estimation and 11.84% in rape biomass estimation. Compared with the method of recursive feature elimination (RFE) selection, the proposed method has improved crops biomass estimation accuracy by 3.90% and 5.21%, respectively.

Key words: Polarization SAR; estimation of crop biomass; genetic algorithm; feature selection; GaoFen-3

Kunpeng XU,Lei ZHAO,Kun LI,Erxue CHEN,Wangfei ZHANG,Hao YANG. Estimation of Crop Biomass Using GF-3 Polarization SAR Data Based on Genetic Algorithm Feature Selection[J]. Journal of Geodesy and Geoinformation Science, 2020, 3(4): 126-136.

Figures/Tables 26

Fig.1

Fig.2

Fig.3

Tab.1

Fig.4

Tab.2

Indicators of estimation model accuracy*"

Accuracy indicators	Symbol	Function
Root mean squared error	RMSE	$RMSE = ∑ i = 1 n (y^i - y i) 2 / n$
Accuracy	Acc.	$Acc. = 1 - RMSE y ? = 1 - ∑ i = 1 n (y^i - y i) 2 / n y ?$
Coefficient of determination	R²	$R 2 = 1 - ∑ i = 1 n (y^i - y i) 2 / ∑ i = 1 n (y i - y ?) 2$

Tab.2

Tab.2

Indicators of estimation model accuracy*"

Accuracy indicators	Symbol	Function
Root mean squared error	RMSE	$RMSE = ∑ i = 1 n (y^i - y i) 2 / n$
Accuracy	Acc.	$Acc. = 1 - RMSE y ? = 1 - ∑ i = 1 n (y^i - y i) 2 / n y ?$
Coefficient of determination	R²	$R 2 = 1 - ∑ i = 1 n (y^i - y i) 2 / ∑ i = 1 n (y i - y ?) 2$

Tab.2

Fig.5

Tab.3

Fig.6

Fig.7

Fig.8

Fig.9

Fig.10

References 21

[1]	MCNAIRN H, BRISCO B. The application of C-band polarimetric SAR for agriculture: a review[J]. Canadian Journal of Remote Sensing, 2004,30(3):525-542.
[2]	MATTIA F, TOAN T L, PICARD G, et al. Multitemporal C-band radar measurements on wheat fields[J]. IEEE Transactions on Geoscience & Remote Sensing, 2003,41(7):1551-1560.
	MATTIA F, TOAN T L, PICARD G, et al. Multitemporal C-band radar measurements on wheat fields[J]. IEEE Transactions on Geoscience & Remote Sensing, 2003,41(7):1551-1560.
[3]	TAO L, LI J, JIANG J, et al. Leaf Area index inversion of winter wheat using RADARSAT-2 data and modified water-cloud model[J]. Journal of Triticeae Crops, 2016,36(2):236-242.
	TAO L, LI J, JIANG J, et al. Leaf Area index inversion of winter wheat using RADARSAT-2 data and modified water-cloud model[J]. Journal of Triticeae Crops, 2016,36(2):236-242.
[4]	HOSSEINI M, MCNAIRN H. Using multi-polarization C- and L-band synthetic aperture radar to estimate biomass and soil moisture of wheat fields[J]. International Journal of Applied Earth Observation & Geoinformation, 2017. DOI: 10.1016/j.jag.2017.01.006. doi: 10.1016/j.jag.2004.07.001 pmid: 22545030
	HOSSEINI M, MCNAIRN H. Using multi-polarization C- and L-band synthetic aperture radar to estimate biomass and soil moisture of wheat fields[J]. International Journal of Applied Earth Observation & Geoinformation, 2017. DOI: 10.1016/j.jag.2017.01.006. doi: 10.1016/j.jag.2004.07.001 pmid: 22545030
[5]	HOSSEINI M, MCNAIRN H, MERZOUKI A, et al. Estimation of Leaf Area Index (LAI) in corn and soybeans using multi-polarization C- and L-band radar data[J]. Remote Sensing of Environment, 2015. DOI: 10.1016/j.rse.2015.09.002. pmid: 31359890
	HOSSEINI M, MCNAIRN H, MERZOUKI A, et al. Estimation of Leaf Area Index (LAI) in corn and soybeans using multi-polarization C- and L-band radar data[J]. Remote Sensing of Environment, 2015. DOI: 10.1016/j.rse.2015.09.002. pmid: 31359890
[6]	YANG H, XIE L, CHEN E, et al. Biomass estimation of oilseed rape using simulated compact polarimtric SAR imagery [C]//GARSS 2016—2016 IEEE International Geoscience and Remote Sensing Symposium. Beijing: IEEE, 2016.
	YANG H, XIE L, CHEN E, et al. Biomass estimation of oilseed rape using simulated compact polarimtric SAR imagery [C]//GARSS 2016—2016 IEEE International Geoscience and Remote Sensing Symposium. Beijing: IEEE, 2016.
[7]	ZHANG W, LI Z, CHEN E, et al. Compact polarimetric response of rape (Brassica napus L.) at C-band: analysis and growth parameters inversion[J]. Remote Sensing, 2017,9(6):591.
	ZHANG W, LI Z, CHEN E, et al. Compact polarimetric response of rape (Brassica napus L.) at C-band: analysis and growth parameters inversion[J]. Remote Sensing, 2017,9(6):591.
[8]	SASAN V, JAVAD S, KAMRAN A, et al. Improving accuracy estimation of forest aboveground biomass based on incorporation of ALOS-2 PALSAR-2 and Sentinel-2A imagery and machine learning: a case study of the Hyrcanian forest area (Iran)[J]. Remote Sensing, 2018,10(2):172.
	SASAN V, JAVAD S, KAMRAN A, et al. Improving accuracy estimation of forest aboveground biomass based on incorporation of ALOS-2 PALSAR-2 and Sentinel-2A imagery and machine learning: a case study of the Hyrcanian forest area (Iran)[J]. Remote Sensing, 2018,10(2):172.
[9]	CHANDRASHEKAR G, SAHIN F. A survey on feature selection methods[J]. Computers & Electrical Engineering, 2014,40(1):16-28.
	CHANDRASHEKAR G, SAHIN F. A survey on feature selection methods[J]. Computers & Electrical Engineering, 2014,40(1):16-28.
[10]	GUYON I, WESTON J, BARNHILL S, et al. Gene selection for cancer classification using support vector machines[J]. Machine Learning, 2002,46(1-3):389-422.
	GUYON I, WESTON J, BARNHILL S, et al. Gene selection for cancer classification using support vector machines[J]. Machine Learning, 2002,46(1-3):389-422.
[11]	LIU XIAOXIAO, WANG LIANG, XU SHENGHUA, et al. A remote sensing feature selection method of forest biomass estimation based on RF-RFE[J]. Science of Surveying and Mapping, 2017,42(05):100-105.
	LIU XIAOXIAO, WANG LIANG, XU SHENGHUA, et al. A remote sensing feature selection method of forest biomass estimation based on RF-RFE[J]. Science of Surveying and Mapping, 2017,42(05):100-105.
[12]	MARYAM S, SAHEBI M R, YASSER M. Improving the accuracy of urban land cover classification using Radarsat-2 PolSAR data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014,7(4):1394-1401.
	MARYAM S, SAHEBI M R, YASSER M. Improving the accuracy of urban land cover classification using Radarsat-2 PolSAR data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014,7(4):1394-1401.
[13]	ATAOLLAH H G, SAHEBI M R, MANSOURIAN A. Polarimetric SAR feature selection using a genetic algorithm[J]. Canadian Journal of Remote Sensing, 2011,37(1):10.
	ATAOLLAH H G, SAHEBI M R, MANSOURIAN A. Polarimetric SAR feature selection using a genetic algorithm[J]. Canadian Journal of Remote Sensing, 2011,37(1):10.
[14]	ZHAO L, CHEN E, LI Z, et al. Three-step semi-empirical radiometric terrain correction approach for PolSAR data applied to forested areas[J]. Remote Sensing, 2017,9(3):269.
	ZHAO L, CHEN E, LI Z, et al. Three-step semi-empirical radiometric terrain correction approach for PolSAR data applied to forested areas[J]. Remote Sensing, 2017,9(3):269.
[15]	FREEMAN A, DURDEN S L. A three-component scattering model for polarimetric SAR data[J]. IEEE Transactions on Geoscience & Remote Sensing, 1998,36(3):963-973.
	FREEMAN A, DURDEN S L. A three-component scattering model for polarimetric SAR data[J]. IEEE Transactions on Geoscience & Remote Sensing, 1998,36(3):963-973.
[16]	CLOUDE S R, POTTIER E. An entropy based classication scheme for land applications of polarimetric SAR[J]. IEEE Transactions on Geoscience and Remote Sensing, 1997. DOI: 10.1109/36.551935.
	CLOUDE S R, POTTIER E. An entropy based classication scheme for land applications of polarimetric SAR[J]. IEEE Transactions on Geoscience and Remote Sensing, 1997. DOI: 10.1109/36.551935.
[17]	ARII M, ZYL J J V, KIM Y. Adaptive Model-Based Decomposition of Polarimetric SAR covariance matrices[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011,49(3):1104-1113.
	ARII M, ZYL J J V, KIM Y. Adaptive Model-Based Decomposition of Polarimetric SAR covariance matrices[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011,49(3):1104-1113.
[18]	YAMADA Y. Preliminary Study on the Radar Vegetation Index (RVI) Application to actual paddy fields by ALOS/PALSAR full-polarimetry SAR data[J]. The International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, 2015,40(7):129.
	YAMADA Y. Preliminary Study on the Radar Vegetation Index (RVI) Application to actual paddy fields by ALOS/PALSAR full-polarimetry SAR data[J]. The International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, 2015,40(7):129.
[19]	LUNEBURG E. Foundations of the mathematical theory of polarimetry[R]. Final Report PhaseI. N00014-00-M-0152, EML Consultants, 2001.
	LUNEBURG E. Foundations of the mathematical theory of polarimetry[R]. Final Report PhaseI. N00014-00-M-0152, EML Consultants, 2001.
[20]	DE SIQUEIRA F R, SCHWARTZ W R, PEDRINI H. Multi-scale gray level co-occurrence matrices for texture description[J]. Neurocomputing, 2013. DOI: 10.1016/j.neucom.2012.09.042.
	DE SIQUEIRA F R, SCHWARTZ W R, PEDRINI H. Multi-scale gray level co-occurrence matrices for texture description[J]. Neurocomputing, 2013. DOI: 10.1016/j.neucom.2012.09.042.
[21]	PEDREGOSA F, GRAMFORT A, MICHEL V, et al. Scikit-learn: machine learning in Python[J]. Journal of Machine Learning Research, 2013,12(10):2825-2830.
	PEDREGOSA F, GRAMFORT A, MICHEL V, et al. Scikit-learn: machine learning in Python[J]. Journal of Machine Learning Research, 2013,12(10):2825-2830.

Feature type	Feature name	Description	Number
Polarization features	\|HH\|,\|HV\|,\|VV\|,\|HH+ VV\|,\|HH-VV\|,SPAN	Backscattering coefficients parameters	16
	Freeman_Odd, Freeman_Vol, Freeman_Dbl	Freeman-Durden decom- position parameters
	Entropy, alpha	H-Alpha decompo- sition parameters
	ANNED_Odd, ANNED_Vol, ANNED_Dbl	Adaptive model-based decomposition
	RVI, Lueneburg entropy	Radar vegetation index and Lueneburg entropy
Textural features	Mean,Homogeneity,Unifor- mity,Dissimilarity,Contrast	GLCM textural features	5

Feature type	Feature name	Description	Number
Polarization features	\|HH\|,\|HV\|,\|VV\|,\|HH+ VV\|,\|HH-VV\|,SPAN	Backscattering coefficients parameters	16
	Freeman_Odd, Freeman_Vol, Freeman_Dbl	Freeman-Durden decom- position parameters
	Entropy, alpha	H-Alpha decompo- sition parameters
	ANNED_Odd, ANNED_Vol, ANNED_Dbl	Adaptive model-based decomposition
	RVI, Lueneburg entropy	Radar vegetation index and Lueneburg entropy
Textural features	Mean,Homogeneity,Unifor- mity,Dissimilarity,Contrast	GLCM textural features	5

Feature selection method	Crop type	Selected feature combinations	Quantity
Feature selection based on linear correlation	Wheat	\|HH\|, \|VV\|, \|HH+VV\|, \|HH-VV\|, SPAN, Mean, Contrast	7
Feature selection based on linear correlation	Rape	\|HH\|, \| HH-VV \|, SPAN, FreemanDbl, ANNEDDbl, Uniformity, Contrast	7
Feature selection of GA	Wheat	\|HV\|, alpha, RVI, FreemanVol, ANNEDOdd, ANNEDVol, Contrast	7
Feature selection of GA	Rape	\|HV\|, SPAN, Lueneburg Entropy, FreemanVol, ANNEDDbl, ANNEDVol, Mean, Contrast	9
RFE	Wheat	\|VV\|, \|HH+VV\|, \|HH-VV\|, FreemanOdd, ANNEDOdd, SPAN, Homogeneity, Dissimilarity, Contrast	9
RFE	Rape	FreemanDbl, FreemanOdd, Homogeneity, Dissimilarity, Contrast	5

Feature selection method	Crop type	Selected feature combinations	Quantity
Feature selection based on linear correlation	Wheat	\|HH\|, \|VV\|, \|HH+VV\|, \|HH-VV\|, SPAN, Mean, Contrast	7
Feature selection based on linear correlation	Rape	\|HH\|, \| HH-VV \|, SPAN, FreemanDbl, ANNEDDbl, Uniformity, Contrast	7
Feature selection of GA	Wheat	\|HV\|, alpha, RVI, FreemanVol, ANNEDOdd, ANNEDVol, Contrast	7
Feature selection of GA	Rape	\|HV\|, SPAN, Lueneburg Entropy, FreemanVol, ANNEDDbl, ANNEDVol, Mean, Contrast	9
RFE	Wheat	\|VV\|, \|HH+VV\|, \|HH-VV\|, FreemanOdd, ANNEDOdd, SPAN, Homogeneity, Dissimilarity, Contrast	9
RFE	Rape	FreemanDbl, FreemanOdd, Homogeneity, Dissimilarity, Contrast	5