A Nonoverlapping Sampling Approach With Peak Data Utilization for Hyperspectral Classification
Yanzi Shi,Yaping Yin,Xiangyu Song,Hua Cui,Dan Zhu,Huansheng Song
DOI: https://doi.org/10.1109/lgrs.2024.3359609
IF: 5.343
2024-02-14
IEEE Geoscience and Remote Sensing Letters
Abstract:Data-driven hyperspectral image classification has gained significant attention across various applications, leading to the development of numerous novel neural networks for effective classification. However, existing spatial–spectral works often overlook the critical aspect of how training and testing sets are split, resulting in data leakage between these sets and suboptimal performance in real-world scenarios. To address this issue, this letter proposes a nonoverlapping sampling approach named PDUnS, drawing inspiration from the theory of extremal polyominoes to achieve peak data utilization, that is, to maximize the number of testing patches. Specifically, PDUnS begins by determining the number of training patches and a random seed point in each connected component of each class. Subsequently, among all constructed polyominoes that cover the seed point and intersect at least pixels with , our PDUnS method searches for those with the minimum perimeter . Finally, we remove pixels in the polyominoes that are not in , and then delete pixels with a degree of 2 and farthest from the boundary of until pixels are retained to construct training patches. Experimental results and in-depth analysis on the indian pines (IP) dataset reveal that PDUnS demonstrates an average 12.38% increase in the number of testing samples to the comparing method. This affirms the superior effectiveness of PDUnS in maximizing data utilization. The code will be available at https://github.com/Yanzi-S/PDUnS.
imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geochemistry & geophysics