A novel method for spectral-spatial classification of hyperspectral images with a high spatial resolution
- PDF / 2,317,617 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 25 Downloads / 218 Views
ORIGINAL PAPER
A novel method for spectral-spatial classification of hyperspectral images with a high spatial resolution Davood Akbari 1 Received: 3 September 2020 / Accepted: 17 November 2020 # Saudi Society for Geosciences 2020
Abstract In this research, a new method for spectral-spatial classification of hyperspectral images based on hierarchical segmentation algorithm is introduced. Among the various spectral-spatial classification algorithms, the marker-based hierarchical segmentation algorithm has so far achieved the best results in combination with the support vector machine (SVM) classification algorithm. In the proposed method adopted in this research, the dimensions of the hyperspectral image were reduced via the minimum noise fraction (MNF) algorithm. Then, five spatial/texture features of wavelet transform, Gabor transform, mean, entropy, and contrast were extracted from the obtained bands. Later on, the multi-layer perceptron (MLP) neural network and SVM classification algorithms were applied to the obtained spectral and texture features, and their results were combined. The resulting map was then used to select the markers and to combine them with the marker-based hierarchical segmentation algorithm using the majority voting rule. The proposed method was applied to three hyperspectral images, Pavia, Telops, and Washington DC Mall. The results of the experiments demonstrate the superiority of the proposed method over the initial marker-based hierarchical algorithm. Quantitatively, it was better by 8, 12, and 9% for the Pavia, Telops, and Washington DC Mall datasets regarding the Kappa coefficient parameter, respectively. Keywords Hyperspectral image . Spectral-spatial classification . Texture features . Marker-based hierarchical segmentation
Introduction Hyperspectral remote sensing technology has made significant progress over the past two decades. This advancement is very evident in the design and creation of sensors as well as in the development and implementation of data processing methods (Varshney and Arora 2004). Today, most research into hyperspectral remote sensing emphasizes the classification of these images (Chan et al. 2020; Ghamisi et al. 2018; Li et al. 2020; Pan et al. 2020; Salghuna and Pillutla 2017). The classification or transformation of images into thematic maps is faced with serious challenges due to factors such as the complexity of the study area, data selection, image processing, and the applied algorithm, and they may affect the success of the classification (Gonzalez and Woods 2002). Although the Responsible Editor: Biswajeet Pradhan * Davood Akbari [email protected] 1
Department of Geomatics Engineering, College of Engineering, University of Zabol, Zabol, Iran
ability to generate data with the above spectral, spatial, and radiometric properties leads to better analysis and successful identification of terrestrial targets, there are also problems which is viewed as a new experience in comparison with the multi-spectral data. The first problem is the relatively larg
Data Loading...
