Feature extraction techniques for ground-based cloud type classification

作者：

Highlights：

• Artificial neural network is used for seven cloud types classification.

• We propose a novel feature extraction method called k-FFTPX.

• Two algorithms are given in the paper.

• We conduct five experiments to evaluate five feature extraction techniques.

• Our second algorithm combined with texture features performs at 90.40% accuracy.

摘要

•Artificial neural network is used for seven cloud types classification.•We propose a novel feature extraction method called k-FFTPX.•Two algorithms are given in the paper.•We conduct five experiments to evaluate five feature extraction techniques.•Our second algorithm combined with texture features performs at 90.40% accuracy.

论文关键词：Image processing,Cloud classification,Ground-based images

论文评审过程：Available online 22 May 2015, Version of Record 28 July 2015.

论文官网地址：https://doi.org/10.1016/j.eswa.2015.05.016