home / centre for automotive safety research / Publications / List / Details
|What computer vision can tell us about road safety?
|Howe M, Mackenzie JRR, Reid I
|Aims to reduce fatalities on city roads by 2030 and within Australia by 2050 have been set but current methods to evaluate and validate safety through crash data do not allow for fast development cycles, often taking years to collect statistically representative data. If these aims are to be met, new methods of safety evaluation must be utilised to decrease the evaluation and validation cycle. Computer vision (CV) has recently become a viable technology for use in this area with its ability to extract a range of road user characteristics. This research explores previous safety research using CV, the current limitations of CV systems, and the next advancements of CV that could give new, exciting tools for evaluating safety. Also presented is a CV system capable of extracting road user characteristics for road safety studies and to build upon with new CV techniques.
|2021 Australasian Road Safety Conference
|28-30 September 2021
|Available online via ARSC website
|Howe M, Mackenzie JRR, Reid I (2021). What computer vision can tell us about road safety?. 2021 Australasian Road Safety Conference, Melbourne, Australia, 28-30 September 2021. [PRESENTED ABSTRACT]