Parameter tuning in the support vector machine and random forest and their performances in cross- and same-year crop classification using TerraSAR-X

Abstract

This article describes the comparison of three different classification algorithms for mapping crops in Hokkaido, Japan, using TerraSAR-X data. In the study area, beans, beets, grasslands, maize, potatoes, and winter wheat were cultivated. Although classification maps are required for both management and estimation of agricultural disaster compensation, those techniques have yet to be established. Some supervised learning models may allow accurate classification. Therefore, comparisons among the classification and regression tree (CART), the support vector machine (SVM), and random forests (RF) were performed. SVM was the optimum algorithm in this study, achieving an overall accuracy of 89.1% for the same-year classification, which is the classification using the training data in 2009 to classify the test data in 2009, and 78.0% for the cross-year classification, which is the classification using the training data in 2009 to classify the data in 2012

Mapping the Dabus Wetlands, Ethiopia, Using Random Forest Classification of Landsat, PALSAR and Topographic Data

“…Random Forest classifications performed well, confirming the widely reported effectiveness of the RF algorithm for land cover classification (e.g., [15,58,64,65]), and particularly for mapping wetlands in tropical environments (e.g., [73,89,92]). Overall, OOB accuracy was above 80% for most models, above 90% for a selection of optimized models, and as high as 99% for model M1, which integrated spectral, SAR and topographic data, and images from multiple years and seasons.…”

“…RF was selected for this study because it generally outperforms conventional classifiers such as the Gaussian maximum likelihood classifier [61,62], while performing favorably, or equally well, to other non-parametric approaches; e.g., CART [63,64], Support Vector Machines [32,65,66], Artificial Neural Networks [67], and K-Nearest Neighbor [68]. It is a powerful non-linear and non-parametric classifier that allows for fusion and aggregation of high-dimensional data from various sources (e.g., optical, SAR, and topography [30,69,70]; SAR and topography [21,58,71]; and optical and topography [72][73][74]).…”

Mapping stand age of rubber plantation using ALOS-2 polarimetric SAR data

“…Comparable to the report, Bagan et al [2012] concluded that the suitable γ value was 0.1 -0.2. It should be noted, however, that very big γ value (64) was found valuable for classification [Sonobe et al, 2014]. Similar case also applies with kernel selection.…”

“…Several supervised classification techniques have been proposed, generally categorized into parametric (including likelihood-based techniques) or non-parametric approaches, as used in decision trees and neural networks. Recent literature, however, tend to focus on the latter, especially SVM and RF [Chabrier et al, 2012;Naidoo et al 2014;Sonobe et al, 2014;Clewley et al 2015]. Both approaches have known being consistently superior than conventional methods such as maximum likelihood classification or decision trees [Rodriguez-Galiano and Chica-Rivas, 2014;He et al, 2015;Low et al 2015].…”

“…Initial SVM documentation suggested C = 1 [Vapnik, 2000], which then was also adopted in earlier SVM implementation in remote sensing [Foody and Mathur, 2004]. Recent paper indicated that lower cost setting was found optimal in the case of agricultural classification using TerraSAR-X data [Sonobe et al, 2014]. Meanwhile, Bagan et al [2012] suggested setting cost parameter between 120 and 150 for a combined multi-spectral and SAR dataset.…”