Scholarlink

To effectively avoid the impact of spectral variability in hyperspectral images on endmember extraction, this paper proposes a method for endmember bundle extraction utilizing a multimodal and multiobjective quantum particle swarm optimizer and relative spectral angle distance (MMQPSORSAD). Firstly, the particles are encoded based on rows and columns, and their positions are updated through quantum particle swarm optimization. A relative spectral angle distance method was proposed to calculate the crowding distance based on the consistency of spectral curves of objects on the same site and then combined with the target space for comprehensive sorting, to achieve the extraction of endmember bundles. When the number of particles and the iterations are 20 and 300, respectively. The root mean square error (RMSE) of the algorithm on the Samson dataset and MUUFL dataset are 0.0058 and 0.0091, respectively. This provides an effective method for extracting endmember bundles using intelligent optimization algorithms in multimodal and multiobjective situations.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Endmember Bundle Extraction Based on Multimodal and Multiobjective Quantum Particle Swarm Optimizer and Relative Spectral Angle Distance

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jiewen Lin,
  • Jian Chen,
  • Xing Mao,
  • Ni Ren

摘要

To effectively avoid the impact of spectral variability in hyperspectral images on endmember extraction, this paper proposes a method for endmember bundle extraction utilizing a multimodal and multiobjective quantum particle swarm optimizer and relative spectral angle distance (MMQPSORSAD). Firstly, the particles are encoded based on rows and columns, and their positions are updated through quantum particle swarm optimization. A relative spectral angle distance method was proposed to calculate the crowding distance based on the consistency of spectral curves of objects on the same site and then combined with the target space for comprehensive sorting, to achieve the extraction of endmember bundles. When the number of particles and the iterations are 20 and 300, respectively. The root mean square error (RMSE) of the algorithm on the Samson dataset and MUUFL dataset are 0.0058 and 0.0091, respectively. This provides an effective method for extracting endmember bundles using intelligent optimization algorithms in multimodal and multiobjective situations.