Per-pixel classification of clouds from whole sky HDR images

作者：

Highlights：

• We propose an automatic weakly-supervised method for extracting labels of cloud types from images of the sky.

• The labelling is assigned per-pixel, allowing for more precise details of clouds to be extracted.

• The method uses High Dynamic Range imagery as an input; leading to a richer set of features.

摘要

•We propose an automatic weakly-supervised method for extracting labels of cloud types from images of the sky.•The labelling is assigned per-pixel, allowing for more precise details of clouds to be extracted.•The method uses High Dynamic Range imagery as an input; leading to a richer set of features.

论文关键词：Cloud-type classification,Per-pixel classification

论文评审过程：Received 11 October 2018, Revised 11 July 2020, Accepted 16 July 2020, Available online 23 July 2020, Version of Record 3 August 2020.

论文官网地址：https://doi.org/10.1016/j.image.2020.115950