Object Localization and Recognition in High Resolution Satellite Image Using U-Net in CNN

This study introduces an innovative approach to localize and recognize objects in high-resolution satellite images, integrating U-Net into a Convolutional Neural Network (CNN) framework. Effectively addressing challenges in detecting and classifying objects within intricate satellite datasets, our method leverages U-Net for precise object localization, capturing intricate spatial details. The CNN enhances recognition by hierarchical feature extraction. Demonstrating superior performance in accurate identification and delineation of objects, our model holds immense potential for applications in remote sensing, urban planning, and environmental monitoring. The integration of U-Net within the CNN establishes a robust solution, contributing significantly to advancements in computer vision for satellite-based applications. This research marks a noteworthy stride in developing sophisticated techniques for object localization and recognition in high-resolution satellite imagery.