Object detection algorithm using deep learning model

The Artificial Intelligence Remote Sensing Laboratory focuses on research in semantic segmentation using satellite imagery. Traditional satellite image analysis methods have relied on complex preprocessing steps and rule-based approaches. However, by utilizing deep learning models, we can achieve higher accuracy in segmentation through automated feature learning. This research compares and analyzes key semantic segmentation models, such as U-Net, DeepLabV3+, and PSPNet, using satellite image datasets, and explores methods for optimizing the performance of each model. Additionally, the study covers applications in cloud detection, urban change detection, building and road detection, and other domains. Based on these efforts, the laboratory aims to contribute to various fields, including urban and agricultural change detection, and the detection of natural disasters such as floods and wildfires.