利用高光谱技术可估测水稻冠层叶绿素含量，为水稻的长势遥感监测与农艺决策提供科学依据. 基于地面实测水稻叶片光谱数据，提出了一种粗糙集属性简约和支持向量回归相结合的叶绿素反演方法，解决了植被光谱指数相关性高易造成计算冗余以及降低水稻叶片叶绿素高光谱反演效率的问题. 首先选择18 个与水稻叶绿素含量相关性较大的植被光谱指数作为因变量，利用粗糙集约简植被指数数据空间得到含有6 个植被光谱指数的简约核；然后采用支持向量回归方法反演叶绿素含量. 基于全部指数反演及基于简约核指数反演的R2 分别为0.858 6与0.850 6. 因此，该方法与采用全部指数进行反演的结果相比，不但具有相当的反演精度，而且有效缩短了反演算法步骤及时间，为大数据处理提供了新的技术方法.
The rice canopy chlorophyll content can be estimated by using the hyperspectral technique for rice growth monitoring and agronomic decision-making. This paper presents a chlorophyll estimation method based on rough set attribute reduction and support vector regression (SVR) using ground spectral data to solve the problem of data redundancy and low retrieve rate caused by high correlation between vegetation indices. Eighteen hyperspectral indices are selected as variables to estimate the chlorophyll content of rice canopy.The data space is reduced using the rough set algorithm. The SVR algorithm is then introduced to estimate the chlorophyll content. There are six indices reserved in the reduced kernel after attribute reduction. R2 of retrieval results based on all indices and reduced kernel are 0.858 6 and 0.850 6 respectively. The proposed method can achieve an accurate forecasting rate based on all feature attributes, and reduce processing steps and estimation time. It provides a new method for big data processing.
[1] CLEVERS J G P W, GITELSON A A. Remote sensing of crop and grass chlorophyll and nitrogen using red-edge bands on Sentinel-2 and -3 [J]. International Journal of Applied Earth Observation and Geoinformation, 2013, 23: 344-351.
[2] XUE L H, YANG L Z. Deriving leaf chlorophyll content of green-leafy vegetables from hyperspectral reflectance [J]. International Journal of Applied Earth Observation and Geoinformation, 2009, 64: 97-106.
[3] MILLIE D F，WECKMAN G R，YOUNG W A，IVEY J E，CARRICK H J，FAHNENSTIEL G L. Modeling microalgal abundance with artificial neural networks: Demonstration of a heuristic ‘Grey-Box’ to deconvolve and quantify environmental influences [J]. Environmental Modelling & Software, 2012, 38: 27-39.
[4] SUO Xingmei, JIANG Yingtao, YANG Mei, LI Shaokun, WANG Kerum, WANG Chongtao. Artificial Neural Network to Predict Leaf Population Chlorophyll Content from Cotton Plant Images [J]. Agricultural Sciences in China, 2010, 9: 38-45.
[5] ALAJLAN N, BAZI Y, MELGANI F, YAGER R R. Fusion of supervised and unsupervised learning for improved classification of hyperspectral images [J]. Information Sciences, 2012, 217: 39-55.
[6] Petropoulos G P, Arvanitis K, Sigrimis N. Hyperion hyperspectral imagery analysis combined with machine learning classifiers for land use/cover mapping [J]. Expert Systems with Applications, 2012, 39: 3800-3809.
[7] Maulik U, Chakraborty D. Learning with transductive SVM for semisupervised pixel classification of remote sensing imagery [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013, 7: 66-78.
[8] Banerjee R, Srivastava P K. Reconstruction of contested landscape: Detecting land cover transformation hosting cultural heritage sites from Central India using remote sensing [J]. Land Use Policy. 2013, 34: 193-203.
[9] Jia Xiuyi, Liao Wenhe, Tang Zhenmin, Lin Shang. Minimum cost attribute reduction in decision-theoretic rough set models [J]. Information Sciences, 2013, 219: 151-167.
[10] Wang Feng, Liang Jiye, Dang Chuangyin. Attribute reduction for dynamic data sets [J]. Applied Soft Computing, 2013, 13: 676-689.
[11] Wang Feng, Liang Jiye, Qian Yuhua. Attribute reduction: A dimension incremental strategy [J]. Knowledge-Based Systems, 2013, 39: 95-108.
[12] Wang Shiping, Zhu Qingxin, Zhu William, Min Fan. Matroidal structure of rough sets and its characterization to attribute reduction [J]. Knowledge-Based Systems, 2012, 36: 155-161.
[13] Mountrakis G, Im J, Ogole C. Support vector machines in remote sensing: A review [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2011, 66(3): 247-259.
[14] Gleason C J, Im Jungho. Forest biomass estimation from airborne LiDAR data using machine learning approaches [J]. Remote Sensing of Environment, 2012, 125: 81-90.
[15] Xu Shuo, An Xin, Qiao Xiaodong. Multi-output least-squares support vector regression machines [J]. Pattern Recognition Letters, 2013, 9(1): 1078-1084.
[16] Shao Yang, Lunetta R S. Comparison of support vector machine, neural network, and CART algorithms for the land-cover classification using limited training data points [J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2012, 70: 78-87.
[17] 杨杰，田永超，姚霞等. 水稻上部叶片叶绿素含量的高光谱估算模型 [J]. 生态学报. 2009，29（12）：6561-6571.
YANG Jie, TIAN Yongchao, YAO Xia, et al. Hyperspectral estimation model for chlorophyll concentrations in top leaves of rice [J]. Acta Ecologica Sinica, 2009, 29(12): 6561-6571
[18] 杨峰，范亚民，李建龙等. 高光谱数据估测稻麦叶面积指数和叶绿素浓度 [J]. 农业工程学报. 2010，26(2)：237-243.
YANG Feng, FAN Yamin, LI Jianlong, et al. Estimating LAI and CCD of rice and wheat using hyperspectral remote sensing data [J]. Transactions of CSAE, 2010, 26(2): 237- 243.
[19] Liu Meiling, Liu Xiangnan, Li Mi. Neural-network for estimating leaf chlorophyll concentration in rice under stress from heavy metals using four spectral indices [J]. Bio-systems engineering, 2010, 106: 223-233.
[20] Gamon J A, Peñuelas J, Field C B. A narrow-waveband spectral index that tracks diurnal changes in photosynthetic efficiency [J]. Remote Sensing of Environment, 1992, 41(1): 35-44.
[21] Gitelson A A, Kaufman Y J, MerzlyakM N. Use of a green channel in remote sensing ofglobal vegetation from EOS-MODIS [J]. Remote Sensing of Environment, 1996, 58: 289-298.
