The application of spatial domain in optimum initialization for clustering image data using particle swarm optimization

作者：

Highlights：

• New PSO-based initializing algorithm for satellite image clustering.

• Increasing the image clustering accuracy using spatial objective functions in PSO.

• Applying simple and efficient criteria for initial seeds.

• Desirable results compared to the traditional initializing methods.

摘要

•New PSO-based initializing algorithm for satellite image clustering.•Increasing the image clustering accuracy using spatial objective functions in PSO.•Applying simple and efficient criteria for initial seeds.•Desirable results compared to the traditional initializing methods.

论文关键词：Initialization,Clustering,spectral–spatial information,Particle swarm optimization,Satellite images

论文评审过程：Received 24 July 2019, Revised 19 September 2020, Accepted 1 November 2020, Available online 6 November 2020, Version of Record 24 January 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.114224