Abstract
Hyperspectral image (HSI) data classification often faces the problem of the scarcity of labeled samples, which is considered to be one of the major challenges in the field of remote sensing. Although active deep networks have been successfully applied in semi-supervised classification tasks to address this problem, their performance inevitably meets the bottleneck due to the limitation of labeling cost. To address the aforementioned issue, this paper proposes a semi-supervised classification method for hyperspectral images that improves active deep learning. Specifically, the proposed model introduces the random multi-graph algorithm and replaces the expert mark in active learning with the anchor graph algorithm, which can label a considerable amount of unlabeled data precisely and automatically. In this way, a large number of pseudo-labeling samples would be added to the training subsets such that the model could be fine-tuned and the generalization performance could be improved without extra efforts for data manual labeling. Experiments based on three standard HSIs demonstrate that the proposed model can get better performance than other conventional methods, and they also outperform other studied algorithms in the case of a small training set.
Highlights
In the past decades, satellite remote sensing has provided advanced detection and research tools for studying the Earth’s resources, for monitoring local and regional environmental changes, and for exploring global environmental changes, given its macroscopic, comprehensive, rapid, dynamic, and accurate measurements [1]
Edge-Preserving Filtering (EPF) optimizes the classification results pixel in the local filtering framework, and the pixel-level spectral information is more by pixel in the local filtering framework, and the pixel-level spectral information is more advantageous relative to the spatial information in the spectral-spatial classification
convolutional neural networks (CNN)-AL is a combination of active learning and a convolutional neural network to achieve good classification results for small sample cases
Summary
Satellite remote sensing has provided advanced detection and research tools for studying the Earth’s resources, for monitoring local and regional environmental changes, and for exploring global environmental changes, given its macroscopic, comprehensive, rapid, dynamic, and accurate measurements [1]. To alleviate the problem of poor learning effect when samples are limited, active learning (AL) methods with high labeling efficiency [19] were proposed They have already been successfully applied to HSI classification. A new hyperspectral image data classification method with semi-supervised active learning has been proposed by previous studies [21,22], which took advantage of AL to increase training samples and improved the machine generalization performance. The new method aimed to improve the generalization ability of the machine by using pseudo-labeled samples, which include robust and informative samples, automatically and actively selected through semi-supervised learning In this way, the cost problem of manual labeling in traditional active learning is alleviated, thereby providing an efficient “self-learning process of the machine”.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
