Comparative Analysis of Support Vector Machine, Maximum Likelihood and Neural Network Classification on Multispectral Remote Sensing Data

Abstract

Land cover classification is an essential process in many remote sensing applications. Classification based on supervised methods have been preferred by many due to its practicality, accuracy and objectivity compared to unsupervised methods. Nevertheless, the performance of different supervised methods particularly for classifying land covers in Tropical regions such as Malaysia has not been evaluated thoroughly. The study reported in this paper aims to detect land cover changes using multispectral remote sensing data. The data come from Landsat satellite covering part of Klang District, located in Selangor, Malaysia. Landsat bands 1, 2, 3, 4, 5 and 7 are used as the input for three supervised classification methods namely support vector machines (SVM), maximum likelihood (ML) and neural network (NN). The accuracy of the generated classifications is then assessed by means of classification accuracy. Land cover change analysis is also carried out to identify the most reliable method to detect land changes in which showing SVM gives a more stable and realistic outcomes compared to ML and NN.

Noise and Restoration of UAV Remote Sensing Images

“…8. It is obvious that these regions possess different brightness conditions due to the different spectral properties of the materials [24]. Graphs of MSE versus noise density were then plotted to investigate the relationship between them for red, green and blue channel and for each of the regions.…”

Haze Effects on Satellite Remote Sensing Imagery and their Corrections

“…ML classification was carried out using all 6 bands to produce 11 classes, which were coastal swamp forest, dry land forest, oil palm, rubber, cleared land, sediment plumes, water, coconut, bare land, urban and industry [9], [10], [11]. To carry out ML classification on the hazy scenes, we need training pixels within the hazy scene.…”

Non-linear Multiclass SVM Classification Optimization using Large Datasets of Geometric Motif Image

“…In addition, SVM determines the hyperplane by maximizing the distance among classes (maximum margin), so that it has high generalization for the testing dataset. Thus, it is better than ANN in which it searches for a hyperplane by principally minimizing its gradient and depending on the number of parameters used [5][6][7][8]. SVM classification works using the principle of Structural Risk Minimization (SRM) to enable it to produce good generalizations with hyperplane fields which can minimize the average error in managing training dataset [7].…”

“…Several previous studies showed that classification with SVM can significantly increase the accuracy value as in the classification to recognize the texture of honey pollen images in which the SVM accuracy results were better than Multilayer Perceptron classification method, Minimum Distance Classifier, and K-Nearest Neighbor [8][9]. SVM classification was tested by two public databases of DNA micro array to classify tumors and non-tumors resulting in a classification accuracy value in which SVM was better than ANN [10].…”