Striped Noise Removal Dataset

  This dataset is specifically designed for the removal of striping noise (also known as banding noise) in remote sensing images, serving as a valuable resource for researchers and practitioners in the field of image processing and remote sensing. It comprises two distinct components: simulated data and real-world data. The simulated data is carefully constructed to mimic various types and intensities of striping noise commonly encountered in satellite or aerial imagery, allowing users to test and validate noise removal algorithms under controlled conditions. The real-world data, on the other hand, includes actual remote sensing images captured by different sensors, which contain authentic striping noise patterns that reflect real-world scenarios. Notably, this dataset is not tailored for deep learning applications. In contrast to datasets that prioritize large-scale image collections for neural network training, this resource is explicitly designed for use with traditional signal processing and image analysis algorithms. It emphasizes compatibility with classical techniques such as Fourier transform-based filtering, wavelet denoising, statistical modeling, and adaptive filtering methods. By providing both simulated and real data, the dataset aims to facilitate the development, comparison, and optimization of conventional noise removal approaches, enabling researchers to evaluate algorithm performance across synthetic benchmarks and practical applications. Whether used for academic research, algorithm development, or industrial applications, this dataset offers a focused and versatile toolset for addressing striping noise challenges in remote sensing without relying on deep learning frameworks.