Virtual SAR
收藏资源简介:
Virtual SAR是由尼赫鲁大学信息技术与工程学院创建的合成数据集,旨在训练基于深度学习的斑点噪声减少算法。该数据集包含31500张图像,源自NWPU-RESISC45数据集,通过添加不同程度的斑点噪声生成。创建过程中,研究者确保每种噪声级别在训练集中具有足够的多样性,以增强神经网络的鲁棒性并避免过拟合。该数据集主要应用于SAR图像的斑点噪声去除,以提高图像处理和计算机视觉任务的性能和效率。
Virtual SAR is a synthetic dataset created by the School of Information Technology and Engineering, Jawaharlal Nehru University, designed for training deep learning-based speckle noise reduction algorithms. This dataset contains 31,500 images derived from the NWPU-RESISC45 dataset, generated by adding speckle noise at varying intensity levels. During the dataset creation process, researchers ensured that each noise level had sufficient diversity in the training set to enhance the robustness of neural networks and avoid overfitting. This dataset is mainly applied to speckle noise removal for SAR images, to improve the performance and efficiency of image processing and computer vision tasks.

- 1Virtual SAR: A Synthetic Dataset for Deep Learning based Speckle Noise Reduction Algorithms信息技术与工程学院,尼赫鲁大学,艾哈迈达巴德,印度 · 2020年



