自适应阶导向融合和多分辨率判别器的遥感图像融合方法
By constructing an integrated model of adaptive order-guided fusion and multi-resolution discriminator, the problem that deep learning models cannot simultaneously achieve spatial-spectral and spatiotemporal fusion is solved, realizing efficient multi-objective task fusion and improving the spatial resolution and spectral information fidelity of remote sensing images.
CN116958005BActive Publication Date: 2026-07-17HAINAN UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HAINAN UNIV
- Filing Date
- 2023-07-28
- Publication Date
- 2026-07-17
Smart Images
Figure CN116958005B_ABST
Abstract
本发明提供一种自适应阶导向融合和多分辨率判别器的遥感图像融合方法，主要包括步骤：构建时空融合数据集和空谱融合数据集；确定自适应阶层；双向自适应阶特征提取；自适应阶渐进式注意力导向融合；构建自适应阶分辨率U型判别器；构建复合损失函数对模型进行优化；使用批量数据同时对生成器和多分辨率U型判别器进行训练，并使用Adam优化器对集成融合模型进行优化，得到最优模型并冻结；使用最优模型对测试数据进行空谱融合和时空融合生成高分辨率图像，本发明构建的集成融合模型既能实现空谱融合又能实现时空融合，实现了多任务的高分辨率图像的生成，并提高了融合精度。
Need to check novelty before this filing date? Find Prior Art