Improved hyperspectral image classification by active learning using pre-designed mixed pixels
作者：
Highlights：
• New Sample design protocol for active learning was investigated.
• Three hyperspectral images were used to evaluate the performance of the method.
• The values of mixed pixels, both for labeled and unlabeled samples, were confirmed.
摘要
•New Sample design protocol for active learning was investigated.•Three hyperspectral images were used to evaluate the performance of the method.•The values of mixed pixels, both for labeled and unlabeled samples, were confirmed.
论文关键词：Sample design,Low-cost,Active learning,Pixel purity index,Support vector machine,Hyperspectral image,Classification
论文评审过程：Received 10 January 2015, Revised 11 August 2015, Accepted 22 August 2015, Available online 25 September 2015, Version of Record 27 November 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.08.019
