This study proposes a novel approach to the classification of hyperspectral data using a one-class classifier, focusing on a single target class (mineral spectra) to address challenges such as limited training data and environmental interference. To mitigate the issue of sparse labeled samples, we employed data augmentation techniques to synthetically expand the training set. We compared the performance of augmented and non-augmented datasets, optimizing for F-Score to balance false positives and false negatives. The experimental results show that both Recall and F-Score were significantly higher for the augmented datasets. Specifically, the augmented datasets achieved higher Recall, indicating an improved ability to identify potential mineral regions, and a higher F-Score, demonstrating a better balance between precision and recall compared to non-augmented datasets. These results, validated in the Cuprite region, confirm the effectiveness of data augmentation in improving mineral area detection, enhancing the reliability of hyperspectral imaging for mineral exploration despite challenges such as limited labeled data and environmental factors.