blue banner

Graduate Projects - Details

Computer Science Program

Project ID: 428
Author: Xinyi Wang
Project Title: A Hybrid Data Mining Approach for Hyperspectral Images Analysis
Semester: 3 2014
Committe Chair: Dr. Longzhuang Li
Committee Member 1: Dr. Scott A. King
Committee Member 2: -
Project Description: Hyperspectral images are a series of image frames in different wavelength bands which form a cube of data with both spatial and spectral information. Data mining is to mine the useful data from the huge datasets. Since the pixels of hyperspectral images are all in very high dimensional space, data mining for hyperspectral images, or also called hyperspectral image classification, is quite challenging. A different approach with more process steps is introduced for better accuracy. The system contains hyperspectral data understanding, data preparation, vegetation detection algorithm, the k-means algorithm, and support vector machine. Data understanding is the first step to let people know about the datasets. Data preparation is the basic step for input of data mining algorithms because some of the algorithms need to have a specified format. Vegetation detection algorithm is to mine the useful data for the k-means algorithms. The k-means algorithm can generate some feature for support vector machine (SVM). SVM is to apply the features from the previous step and to mine the data, and finally some accuracies are generated according to those processes of hyperspectral imaging dataset, which helps us test the model.
Project URL:   428.pdf
© Texas A&M University-Corpus Christi • 6300 Ocean Drive, Corpus Christi, Texas 78412 • 361-825-5700