This study uses remote sensing technology that can provide information aboutthe condition of the earth's surface area, fast, and spatially. The study areawas in Karawang District, lying in the Northern part of West Java-Indonesia. Weaddress a paddy growth stages classification using LANDSAT 8 image dataobtained from multi-sensor remote sensing image taken in October 2015 to August2016. This study pursues a fast and accurate classification of paddy growthstages by employing multiple regularizations learning on some deep learningmethods such as DNN (Deep Neural Networks) and 1-D CNN (1-D ConvolutionalNeural Networks). The used regularizations are Fast Dropout, Dropout, and BatchNormalization. To evaluate the effectiveness, we also compared our method withother machine learning methods such as (Logistic Regression, SVM, RandomForest, and XGBoost). The data used are seven bands of LANDSAT-8 spectral datasamples that correspond to paddy growth stages data obtained from i-Sky (eye inthe sky) Innovation system. The growth stages are determined based on paddycrop phenology profile from time series of LANDSAT-8 images. The classificationresults show that MLP using multiple regularization Dropout and BatchNormalization achieves the highest accuracy for this dataset.
展开▼
