Classification of Images Using Support Vector Machines (original) (raw)
Related papers
Support Vector Machines for Land Cover Mapping from Remote Sensor Imagery
Monitoring and Modeling of Global Changes: A Geomatics Perspective, 2015
Landscape mapping is an important activity leading to the generation of various geospatial products in support of sustainable land use planning and environmental management. A variety of pattern recognition techniques have been identified to automate this process from remote sensor imagery. Support vector machines (SVM) as a group of relatively novel statistical learning algorithms have demonstrated their robustness in mapping homogeneous and heterogeneous landscape types from remote sensor imagery. In this paper, we review the status and potential challenges in the implementation of SVM for landscape mapping. The paper is organized into two major parts. The first part reviews the research status of using SVM for landscape mapping from remote sensor imagery, focusing on some comparative studies that demonstrated the effectiveness of SVM over other conventional classifiers. We identify several areas for additional research, which are mostly related to appropriate treatments of some parametric and non-parametric factors in order to achieve improved mapping accuracies particularly for heterogeneous landscape types and for large datasets. Then, we implement the support vector machine technique to map various land cover types from a remote sensor image covering an urban area, and demonstrate the robustness of this geospatial technique for mapping heterogeneous landscapes.
Different land cover classification methods exist. This paper shows that Support vector machines (SVMs) have the potential to outperform Maximum Likelihood and Minimum Distance Classification in rugged area covered with rainforest and in undulating areas dominated by fields, grassland and coniferous forest. However, SVM does not guarentee better classification results. Kernel and generalization parameters as well as the input variables significantly influence the performance of SVM.
Image Classification Using SVMs: One-against-One Vs One-against-All
Computing Research Repository, 2007
Support Vector Machines (SVMs) are a relatively new supervised classification technique to the land cover mapping community. They have their roots in Statistical Learning Theory and have gained prominence because they are robust, accurate and are effective even when using a small training sample. By their nature SVMs are essentially binary classifiers, however, they can be adopted to handle the multiple classification tasks common in remote sensing studies. The two approaches commonly used are the One-Against-One (1A1) and One-Against-All (1AA) techniques. In this paper, these approaches are evaluated in as far as their impact and implication for land cover mapping. The main finding from this research is that whereas the 1AA technique is more predisposed to yielding unclassified and mixed pixels, the resulting classification accuracy is not significantly different from 1A1 approach. It is the authors conclusion therefore that ultimately the choice of technique adopted boils down to personal preference and the uniqueness of the dataset at hand.
This paper studies the impact of several learning issues in an image classification task with SVMs, such as rich featurebased representations, optimization and sensitivity to novelty in the test data sets. The employed imagery refers to the city of Rome, Italy and is acquired in different years and seasons by the European Remote Sensing Satellites ERS-1 and ERS-1/2 tandem mission. A comprehensive evaluation according to varying training conditions is reported, showing that SVMs provide robust and largely applicable tools.
Image Classification and Land Cover Mapping Using Sentinel-2 Imagery: Optimization of SVM Parameters
Land
Land use/cover (LU/LC) classification provides proxies of the natural and social processes related to urban development, providing stakeholders with crucial information. Remotely sensed images combined with supervised classification are common to define land use, but high-performance classifiers remain difficult to achieve, due to the presence of model hyperparameters. Conventional approaches rely on manual adjustment, which is time consuming and often unsatisfying. Therefore, the goal of this study has been to optimize the parameters of the support vector machine (SVM) algorithm for the generation of land use/cover maps from Sentinel-2 satellite imagery in selected humid and arid (three study sites each) climatic regions of Iran. For supervised SVM classification, we optimized two important parameters (gamma in kernel function and penalty parameter) of the LU/LC classification. Using the radial basis function (RBF) of the SVM classification method, we examined seven values for both...
2011
Land cover classification is a key research field in remote sensing and land change science as thematic maps derived from remotely sensed data have become the basis for analyzing many socio-ecological issues. However, land cover classification remains a difficult task and it is especially challenging in heterogeneous tropical landscapes where nonetheless such maps are of great importance. The present study aims to establish an efficient classification approach to accurately map all broad land cover classes in a large, heterogeneous tropical area of Bolivia, as a basis for further studies (e.g., land cover-land use change). Specifically, we compare the performance of parametric (maximum likelihood), non-parametric (k-nearest neighbour and four different support vector machines-SVM), and hybrid classifiers, using both hard and soft (fuzzy) accuracy assessments. In addition, we test whether the inclusion of a textural index (homogeneity) in the classifications improves their performance. We classified Landsat imagery for two dates corresponding to dry and wet seasons and found that non-parametric, and particularly SVM classifiers, outperformed both parametric and hybrid classifiers. We also found that the use of the homogeneity index along with reflectance bands significantly increased the overall accuracy of all the classifications, but particularly of SVM algorithms. We observed that improvements in producer's and user's accuracies through the inclusion of the homogeneity index were different depending on land cover classes. Earlygrowth/degraded forests, pastures, grasslands and savanna were the classes most improved, especially with the SVM radial basis function and SVM sigmoid classifiers, though with both classifiers all land cover classes were mapped with producer's and user's accuracies of around 90%. Our approach seems very well suited to accurately map land cover in tropical regions, thus having the potential to contribute to conservation initiatives, climate change mitigation schemes such as REDD+, and rural development policies.
Sensors
In previous classification studies, three non-parametric classifiers, Random Forest (RF), k-Nearest Neighbor (kNN), and Support Vector Machine (SVM), were reported as the foremost classifiers at producing high accuracies. However, only a few studies have compared the performances of these classifiers with different training sample sizes for the same remote sensing images, particularly the Sentinel-2 Multispectral Imager (MSI). In this study, we examined and compared the performances of the RF, kNN, and SVM classifiers for land use/cover classification using Sentinel-2 image data. An area of 30 × 30 km 2 within the Red River Delta of Vietnam with six land use/cover types was classified using 14 different training sample sizes, including balanced and imbalanced, from 50 to over 1250 pixels/class. All classification results showed a high overall accuracy (OA) ranging from 90% to 95%. Among the three classifiers and 14 sub-datasets, SVM produced the highest OA with the least sensitivity to the training sample sizes, followed consecutively by RF and kNN. In relation to the sample size, all three classifiers showed a similar and high OA (over 93.85%) when the training sample size was large enough, i.e., greater than 750 pixels/class or representing an area of approximately 0.25% of the total study area. The high accuracy was achieved with both imbalanced and balanced datasets.
Support Vector Machine for Remote Sensing image classification
2013
The presented work deals with remote sensing data classification. The major goal is to provide the la nd characterization for multispectral image observations. Channel images contain data acquired from different wavelen gth within the frequency spectrum. Due to the multiple radianc e reflection, the land characterization in the observation space became complex and inefficient. The goal of this work is to perform a feature space for observations. Then a statically le arning classifier using the Support Vector Machine is deve loped for a reliable land characterization. Keywords— Fusion, Segmentation, Classification, Support Vector Machine, Feature extraction
An svm multiclassifier approach to land cover mapping
Computing Research Repository, 2010
From the advent of the application of satellite imagery to land cover mapping, one of the growing areas of research interest has been in the area of image classification. Image classifiers are algorithms used to extract land cover information from satellite imagery. Most of the initial research has focussed on the development and application of algorithms to better existing and emerging classifiers. In this paper, a paradigm shift is proposed whereby a committee of classifiers is used to determine the final classification output. Two of the key components of an ensemble system are that there should be diversity among the classifiers and that there should be a mechanism through which the results are combined. In this paper, the members of the ensemble system include: Linear SVM, Gaussian SVM and Quadratic SVM. The final output was determined through a simple majority vote of the individual classifiers. From the results obtained it was observed that the final derived map generated by an ensemble system can potentially improve on the results derived from the individual classifiers making up the ensemble system. The ensemble system classification accuracy was, in this case, better than the linear and quadratic SVM result. It was however less than that of the RBF SVM. Areas for further research could focus on improving the diversity of the ensemble system used in this research.