Kernel-Based Regularized Learning for Time-Invariant Detection of Paddy Growth Stages from MODIS Data

Most current studies have been applying high temporal resolution satellite data for determining paddy crop phenology, that derive into a certain vegetation indices, by using some filtering and smoothing techniques combined with threshold methods. In this paper, we introduce a time invariant detection of paddy growth stages using single temporal resolution satellite data instead of high temporal resolution with complex cropping pattern. Our system is a kernel-based regularized learner that predicts paddy growth stages from six-bands spectral of Moderate Resolution Image Spectroradiometer (MODIS) satellite data. It evaluates three Kernel-based Regularized (KR) classification methods, i.e. Principal Component Regression (KR-PCR), Extreme Learning Machine (KR-ELM), and Support Vector Machine with radial basis function (RBF-SVM). All data samples are divided into training (25%) and testing (75%) sampling, and all models are trained and tested through 10-rounds random bootstrap re-sampling method to obtain more variety on hypothesis models during learning. The best model for each classifier method is defined as the one which has the highest kappa coefficient during testing. The experimental results show that the classification accuracy of each classifiers on testing are high competitive, i.e. 84.08%, 84.04%, and 84.95% respectively.

[1]  Danny Lo Seen,et al.  A Comparative Study on Satellite- and Model-Based Crop Phenology in West Africa , 2014, Remote. Sens..

[2]  Chee Kheong Siew,et al.  Extreme learning machine: Theory and applications , 2006, Neurocomputing.

[3]  Changsheng Li,et al.  Mapping paddy rice agriculture in South and Southeast Asia using multi-temporal MODIS images , 2006 .

[4]  Bingfang Wu,et al.  Monitoring Crop Phenology with MERIS Data | A Case Study of Winter Wheat in North China Plain , 2009 .

[5]  J. Mustard,et al.  Wavelet analysis of MODIS time series to detect expansion and intensification of row-crop agriculture in Brazil , 2008 .

[6]  T. Sakamoto,et al.  A crop phenology detection method using time-series MODIS data , 2005 .

[7]  Changsheng Li,et al.  Mapping paddy rice agriculture in southern China using multi-temporal MODIS images , 2005 .

[8]  Akara Prayote,et al.  Rice Phenology Monitoring Using PIA Time Series MODIS Imagery , 2013, 2013 10th International Conference Computer Graphics, Imaging and Visualization.

[9]  Sidik Mulyono,et al.  Genetic algorithm based new sequence of principal component regression (GA-NSPCR) for feature selection and yield prediction using hyperspectral remote sensing data , 2012, 2012 IEEE International Geoscience and Remote Sensing Symposium.

[10]  Dailiang Peng,et al.  Detection and estimation of mixed paddy rice cropping patterns with MODIS data , 2011, Int. J. Appl. Earth Obs. Geoinformation.

[11]  A. Huete,et al.  Mapping paddy rice with multi-date moderate-resolution imaging spectroradiometer (MODIS) data in China , 2009 .

[12]  T. Basaruddin,et al.  A paddy growth stages classification using MODIS remote sensing images with balanced branches support vector machines , 2012, 2012 International Conference on Advanced Computer Science and Information Systems (ICACSIS).

[13]  Rick Archibald,et al.  Feature Selection and Classification of Hyperspectral Images With Support Vector Machines , 2007, IEEE Geoscience and Remote Sensing Letters.

[14]  Lin Wang,et al.  Multi-Temporal Detection of Rice Phenological Stages Using Canopy Spectrum , 2014 .

[15]  Seungtaek Jeong,et al.  Development of Variable Threshold Models for detection of irrigated paddy rice fields and irrigation timing in heterogeneous land cover , 2012 .

[16]  Lorenzo Bruzzone,et al.  Kernel-based methods for hyperspectral image classification , 2005, IEEE Transactions on Geoscience and Remote Sensing.

[17]  A. B. Harto,et al.  Detecting Rice Phenology in Paddy Fields with Complex Cropping Pattern Using Time Series MODIS Data , 2010 .