A New Remote Sensing Change Detection Data Augmentation Method Based on Mosaic Simulation and Haze Image Simulation

The quality of optical remote sensing images is largely affected by clouds and haze. In addition, the mosaicking image multiple images, due to objective factors such as acquiring time or climate conditions, will lead large spectral differences in area around seamline. aforementioned scenarios seriously affect accuracy change detection models based on deep learning. However, there still a lack methods address issues. To solve these problems, from perspective training samples, this article proposed simple but effective data augmentation method improve generalization ability model region haze cover First, characteristics itself, two simulation are designed conduct augmentation, named mosaic simulation. Then, newly augmented samples mixed with original then input into learning for training. Finally, results indicate that can effectively seamline, which has high practical value improving model's performance real-world also provides algorithm reference other intelligent interpretation tasks data.