Kitti Koonsanit

Band Selection for Dimension Reduction in Hyper Spectral Image Using Integrated Information Gain and Principal Components Analysis Technique

Nowadays, hyper spectral image software be- comes widely used. Although hyper spectral images provide abundant information about bands, their high dimensionality also substantially increases the computational burden. An important task in hyper spectral data processing is to reduce the redundancy of the spectral and spatial information without losing any valuable details. In this paper, we present band selection technical using principal components analysis (PCA) and information gain (IG) for hyper spectral image such as small multi-mission satellite (SMMS). Band selection method in our research not only serves as the first step of hyper spec- tral data processing that leads to a significant reduction of computational complexity, but also a invaluable research tool to identify optimal spectral for different satellite applications. In this paper, an integrated PCA and IG method is proposed for hyper spectral band selection. Based on tests in a SMMS hyper spectral image, this new method achieves good result in terms of robust clustering.

Determination of the Initialization Number of Clusters in K-means Clustering Application Using Co-Occurrence Statistics Techniques for Multispectral Satellite Imagery

 Abstract—Nowadays, clustering is a popular tool for explo- ratory data analysis, such as K-means and Fuzzy C-mean. Automatic determination of the initialization number of clus- ters in K-means clustering application is often needed in ad- vance as an input parameter to the algorithm. In this paper, a method has been developed to determine the initialization number of clusters in satellite image clustering application using a data mining algorithm based on the co-occurrence matrix technique. The proposed method was tested using data from unknown number of clusters with multispectral satellite image in Thailand. The results from the tests confirm the effectiveness of the proposed method in finding the initializa- tion number of clusters and compared with isodata algorithm. Clustering is a popular tool for data mining and explora- tory data analysis. One of the major problems in cluster analysis is the determination of the number of clusters in unlabeled data, which is a basic input for most clustering algorithms. In this paper, we propose a new easy method for automatically estimating the number of clusters in unlabeled data set. Pixel clustering technique in a color image is a process of unsupervised classification of hundreds thou- sands or millions pixels on the basis of their colors. One of the most popular and fastest clustering techniques is the k- means technique. The results of the k-means technique depend on different factors such as a method of determina- tion of initial cluster centers as shown in Fig. 1. Such sensi- tivity to initialization is an important disadvantage of the k- means technique. In this paper, a method has been devel- oped to determine the initialization number of clusters in satellite image clustering application using a data mining algorithm based on the co-occurrence matrix technique. Therefore, automatic determination of the initialization number of clusters can greatly help with the unsupervised classification of satellite Image.

Parameter-Free K-Means Clustering Algorithm for Satellite Imagery Application

Unsupervised classification is a popular tool for unlabeled datasets in data mining and exploratory data analysis, such as K-means and Fuzzy C-mean. Although these unsupervised techniques have demonstrated substantial success for satellite imagery, they have some limitations. The initialization number of clusters in K-means clustering application is often needed in advance as an input parameter to the algorithm. Our previous paper regarding the initialization number of clusters in K-means clustering application with a co-occurrence matrix technique has been published. Although our previous approach regarding the number of cluster was discovered, but it was limited to count a number of peaks in occurrence matrix as the number of clusters by manual counting. The best of our previous approach need to automatically find and count a number of peaks in occurrence matrix. In this research, we assume that the satellite imagery is given and we have no knowledge beforehand for segmentation. Hence, this paper presents a simple, parameter-free K-means method for K-means in satellite imagery clustering application to determine the initialization number of clusters with image processing algorithms based on the co-occurrence matrix technique. A maxima technique is proposed for automatic counting a number of peaks in occurrence matrix as the number of clusters. The parameter-free method was tested with hyperspectral imagery and multispectral imagery. The results from the tests confirm the effectiveness of the proposed method in K-means method and compared with isodata algorithm.

Band selection for hyperspectral image using principal components analysis and maxima-minima functional

Nowadays, hyperspectral image software becomes widely used. Although hyperspectral images provide abundant information about bands, their high dimensionality also substantially increases the computational burden. An important task in hyperspectral data processing is to reduce the redundancy of the spectral and spatial information without losing any valuable details. In this paper, we present band selection technical using principal components analysis (PCA) and maxima-minima functional for hyperspectral image such as small multi-mission satellite (SMMS). Band selection method in our research not only serves as the first step of hyperspectral data processing that leads to a significant reduction of computational complexity, but also a invaluable research tool to identify optimal spectral for different satellite applications. In this paper, an integrated PCA and maxima-minima functional method is proposed for hyperspectral band selection. Based on tests in a SMMS hyperspectral image, this new method achieves good result in terms of robust clustering.

A simple estimation the number of classes in satellite imagery

Clustering is a popular tool for exploratory data analysis, such as K-means and Fuzzy C-mean. A simple estimation the number of classes for segmented areas (K) in satellite imagery application is often needed in advance as an input parameter to the K-means algorithm. In this paper, a method has been developed to estimate the number of classes for segmented areas in satellite imagery clustering application using an image processing technique based on the co-occurrence matrix technique. The proposed method was tested using data from known the number of classes with satellite imagery. The results from the tests confirm the effectiveness of the proposed method in finding the estimation the number of classes and compared with ground truth data.

A Simple Anomaly Detection for Spectral Imagery Using Co-occurrence Statistics Techniques

Anomaly detection has always been a hot research field of data mining. Anomaly detection is important in many fields. Automatic determination of the anomaly cluster is often needed to eliminate that anomaly cluster. In this paper, a method has been developed to determine the anomaly regions in satellite image using a data mining algorithm based on the co-occurrence matrix technique in order to determinate that anomaly. Our method consists of four stages, the first stage estimate a number of cluster by co-occurrence matrix, the second stage cluster dataset by automatic clustering algorithm, the third stage detect anomalous clusters by threshold value and the final stage defines clusters, which are lower than threshold value, to be anomalous clusters. The proposed method was tested using data from unknown number of clusters with multispectral satellite image in Thailand. The results from the tests confirm the effectiveness of the proposed method in finding the anomaly regions.

A Simple Automatic Outlier Regions Detection

Automatic determination of the outlier regions is often needed to eliminate that outlier region. In this paper, a method has been developed to determine the outlier regions in satellite image using a data mining algorithm based on the co-occurrence matrix technique in order to determinate that outlier. Our method consists of four stages, the first stage estimate a number of region by co-occurrence matrix, the second stage cluster dataset by automatic clustering algorithm, the third stage detect outlier regions by automatic threshold value and the final stage defines outlier regions, which are lower than threshold value, to be outlier regions. The proposed method was tested using data from unknown number of regions with multispectral satellite image in Thailand. The results from the tests confirm the effectiveness of the proposed method in finding the outlier regions.