Scholarlink

Hyperspectral images are affected by many factors in the imaging process will produce non-stationary noise pollution, which will interfere with the spectral curve of the target itself in the field of view affecting the spectral feature extraction for hyperspectral image classification. To this end, a method based on automatically determining the FrFT of the optimal fractional domain is proposed to preprocess the image to obtain the target spectral curve in the intermediate domain, and to reconstruct the spectral curve by SG smoothing filter to avoid feature redundancy phenomenon after the data are transformed by FrFT. The experimental results show that the method can effectively improve the target spectral curve features and increase the classification accuracy of the classifier. ...

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

Spectral Aerial Hyperspectral Image Classification Based on FrFT Target Spectral Band Reconstruction

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Xiqing Li,
  • Pengge Ma,
  • Daijun Liu,
  • Jiankang Zhang,
  • Junling Sun,
  • Jinwang Qian,
  • Qiuchun Jin,
  • Ran Tao

摘要

Hyperspectral images are affected by many factors in the imaging process will produce non-stationary noise pollution, which will interfere with the spectral curve of the target itself in the field of view affecting the spectral feature extraction for hyperspectral image classification. To this end, a method based on automatically determining the FrFT of the optimal fractional domain is proposed to preprocess the image to obtain the target spectral curve in the intermediate domain, and to reconstruct the spectral curve by SG smoothing filter to avoid feature redundancy phenomenon after the data are transformed by FrFT. The experimental results show that the method can effectively improve the target spectral curve features and increase the classification accuracy of the classifier. ...