Використання методів комп’ютерного зору для виявлення небезпечних ділянок велосипедних маршрутів

Abstract

This article considers the problem of cycling safety in urban environments. The relevance of the topic is related to the growing popularity of cycling as a means of transportation and the insufficient level of safety for cyclists in cities. Many dangerous situations during cycling are not reflected in official road accident statistics. The paper proposes an approach for analyzing video data collected during cyclists’ trips. Methods of computer vision and object detection models are used to process video streams. Based on this analysis, the system can identify risk factors and potentially dangerous road sections. The results can help detect unsafe areas for cyclists. They may also be used to create safety maps of cycling routes and support improvements in urban cycling infrastructure