Vision-Language Artificial Intelligence for Robotic-Based Monitoring: Concrete Defect Detection, Classification, and Localization in Two-Dimensional Maps

Abstract

AbstractThis paper introduces a novel framework that combines vision-language models (VLMs) and localization techniques to detect, classify, and localize visual structural defects using moving platforms such as robots and handheld devices, with an ...