Open AccessVideo‐based method for detecting potential risks among multiple electric bikes
Author(s) -
Xia Yingji,
Gao Yuhong,
Tian Jing,
Liu Tianze
Publication year - 2019
Publication title -
iet intelligent transport systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.579
H-Index - 45
eISSN - 1751-9578
pISSN - 1751-956X
DOI - 10.1049/iet-its.2018.5559
Subject(s) - pairwise comparison , trajectory , computer science , potential field , causality (physics) , process (computing) , artificial intelligence , data mining , physics , quantum mechanics , astronomy , geophysics , geology , operating system
The tremendous adoption and usage of electric bikes have caused a series of challenges to road safety in China. The existing researches mainly use accumulated trajectory clusters or pairwise conflict indicators to model the potential risks among electric bikes, neglecting different spatial interactions from surrounding electric bikes or the corresponding temporal causality. A novel video‐based sequential heat map method is proposed to represent the temporal influence area of a running electric bike as a teardrop‐shaped thermal potential field, instead of accumulating group conflicts from irrelevant trajectories or decomposing them into several pairwise indicators. A time decay factor along with a thermal diffusion process is introduced in order to retain both spatial and temporal motion information. Moreover, a spatial–temporal potential risk index scheme based on the sequential heat map model is proposed to detect the type of potential risk as well as the occurring time and place. The experiment results have shown its strength and efficiency of the proposed risk detecting method in seeking for the potential risk areas and electric bikes’ conflicts.