No abstract

No abstract

.Hyperspectral images (HSIs) have recently been exploited in several aspects as HSIs contain many contiguous and narrow discriminative spectral bands. The problem of dimensionality is a significant dilemma for HSIs due to there being plenty of irrelevant and redundant spectral bands and highly correlated bands that lead to Hughes phenomenon. To this end, we present an approach to selecting the most informative and relevant spectral bands for HSI dimensionality reduction using the Krill Herd (KH) algorithm. Moreover, KH is a heuristic search method that seeks to reach the optimum global solution within the search space and effectively evade falling into the local optima. Then an edge-preserving filter was employed to extract the spatial features while reducing noise and obtaining a suitable smoothing that improves the classification performance. Finally, the support vector machine classifier was performed at the pixel level for HSI classification. Furthermore, the proposed work was compared with the harmony search, genetic algorithm, bat algorithm, particle swarm optimization, and firefly algorithm. The experimental results demonstrated outstanding performance with an overall accuracy equal to 96.54%, 98.93%, 99.78%, and 98.66% on four hyperspectral datasets: Indian Pines scene, Pavia University scene, Salinas scene, and Botswana scene, respectively.

Hyperspectral images (HSI) have recently been exploited in several aspects as they contain many contiguous narrow spectral informative-rich bands. The curse of dimensionality in hyperspectral images is an essential challenge as it possesses plenty of redundant bands that lead to the Hughes phenomenon. However, many feature selection or band selection techniques have been performed for dimensionality reduction of HSI. In this manuscript, firstly, a novel approach for spectral bands selection process is presented for hyperspectral images dimensionality reduction using Krill Herd (KH) Algorithm. However, KH is a heuristic search method that seeks to reach the optimum global solution within the search space and evade falling into the local optima. KH relies on simulating the herding behavior of krill in the sea in order to determine the most informative and relevant bands. Secondly, an Edge-preserving filter (EPF) was utilized to extract the spatial characteristics while reducing noise and obtaining a suitable smoothing that improves the performance of the classification process. Finally, the support vector machine (SVM) classifier at pixel level was performed for the classification of HSI. Moreover, the proposed work was compared to the Harmony Search (HS), Genetic Algorithm (GA), Bat Algorithm (BA), Particle Swarm Optimization (PSO), and Firefly Algorithm (FA). In addition, the classification results for overall accuracy (OA) on four popular publicly datasets were 96.54%, 98.93%, 99.78%, and 98.66% for the Indian Pines scene, the Pavia University scene, the Salinas scene, and the Botswana scene, respectively.

No abstract