The study explores using generative AI, specifically ChatGPT, to assist in mapping tasks by suggesting OpenStreetMap (OSM) tags based on descriptions of street-level photographs. By combining volunteered geographic information (VGI) and large language models (LLMs), the research aims to improve the efficiency of collaborative mapping efforts. The experiment involved human analysts and an artificial analyst (BLIP-2) describing street scenes from Mapillary images to prompt ChatGPT for tagging suggestions. Results show that providing detailed descriptions and additional context can significantly increase the accuracy of mapping suggestions without modifying the underlying AI models. The study highlights the potential of leveraging generative AI for enhancing map databases through innovative approaches.
เป็นภาษาอื่น
จากเนื้อหาต้นฉบับ
arxiv.org
สอบถามเพิ่มเติม