Generating Simulated Infrared Sensor Images Using Deep Learning Neural Networks

This study presents a method for generating infrared images using deep learning techniques, specifically Generative Adversarial Networks (GANs), trained on data from real flight experiments. Implemented in Python, the approach leverages a U-Net adapted for gray-scale IR image generation. After training on a multi-core computer, the method achieved accurate infrared image generation, applicable in flight simulation and autonomous navigation.