Hyperspectral image (HSI) is celebrated for its detailed spectral information but faces significant challenges in dimensionality reduction stemming from excessive spectral dimensions. Band selection (BS) is a critical technique in dimension reduction, aiming to identify an optimal subset of spectral bands with minimal redundancy and maximal feature separability. Swarm intelligence methods are renowned for their flexibility and high performance in optimization problems. These methods have been extensively introduced by scholars to address BS tasks in hyperspectral imaging. Among these, the remora optimization algorithm (ROA) stands out for its exceptional optimization proficiency. However, its conventional evolutionary operators are susceptible to local optimum stagnation. Therefore, a novel BS method based on an improved ROA, termed IROA-BS, is proposed for HSI classification. First, an evaluation function is designed to estimate the class separability and redundancy of selected band subsets. Second, the hybrid evolutionary operators are intended to diversify potential solutions, while a multistage mutation module is implemented to circumvent local optimum stagnation. Furthermore, a guided learning strategy is utilized to fine-tune the equilibrium between exploration and exploitation processes. The effectiveness of the proposed IROA-BS method is rigorously validated across three widely recognized hyperspectral remote sensing image datasets. Comparative analysis with the other advanced BS methods and swarm intelligence techniques validates the superiority and efficacy of the IROA-BS method in HSI BS applications.

Existing data acquisition technologies desire further improvement to meet the increasing need for big, accurate, and high-quality data collection. Most of the collected data have redundant information such as noise. To improve the classification accuracy, the dimensionality reduction technique, which is also known as the feature selection, is a necessity for data processing. In this paper, the slime mould algorithm (SMA) is optimized by introducing the composite mutation strategy (CMS) and restart strategy (RS). The improved SMA is named CMSRSSMA, which stands for the CMS, RS, and the improved SMA. The CMS is utilized to increase the population diversity, and the RS is used to avoid the local optimum. In this paper, the CEC2017 benchmark function is used to test the effectiveness of the proposed CMSRSSMA. Then, the CMSRSSMA-SVM model is proposed for feature selection and parameter optimization simultaneously. The performance of the model is tested by 14 data sets from UCI data repository. Experimental results show that the proposed method is superior to other algorithms in terms of classification accuracy, number of features and fitness value on most selected data sets.