Transformer with Reinforcement and Actor-Critic for Edge-vectorization (TRACE)

Current Automated Feature Extraction (AFE) methodologies, primarily based on Convolutional Neural Network (CNN) interpretation and segmentation of satellite imagery, often yield incomplete, inaccurate, and non-routable road networks, particularly in scenarios with visual obstructions like tree-lined roads. To address this problem, the collaboration between In-Depth Engineering and Miami University of Ohio proposes the “Transformer with Reinforcement and Actor-Critic for Edge-vectorization” (TRACE) solution – a refined approach to advance road extraction methodologies by incorporating advanced Reinforcement Learning (RL) techniques using a transformer in a previously proven method of road extraction. The RL framework is particularly well-suited to learn the decision-making process involved in extending road networks, and TRACE addresses the challenge of balancing long-term rewards against immediate cues, a critical aspect in managing occlusions and accurately delineating complex road intersections and junctions.