Human behavior plays a major role in the causation of road traffic crashes and the severity of their outcomes. The accurate and detailed registration of road user behavior is therefore vital to identify variables that can decrease the number of road crashes. Naturalistic observation of road user behavior in the field presents a main pillar of injury prevention research. Since observation is highly time-consuming when carried out by human observers, video-based naturalistic observation has been proposed as an efficient alternative.
With this poster we want to present attendees with a low-cost camera system that can be built from of the shelf components by researchers in low- and middle-income countries (LMIC) to conduct video-based naturalistic observation. Furthermore, we want to share our experience when coding road user behavior with open-source software, using a study conducted in Myanmar as an example.
For a study focused on the helmet use of motorcycle riders, two video cameras were built to observe traffic in seven cities of Myanmar. The cameras were built from a Raspberry Pi mini-computer, a compatible camera, and a power-bank for an approximate cost of 120 US$ per camera. The open-source software BORIS (Behavioral Observation Research Interactive Software) was used to code road user behavior in the recordings.
The presented video-based observation system allowed researchers to collect data at a variety of locations simultaneously as well as at different times. The analysis of the recorded data can be repeated for different research questions. Collection and analysis of data can be conducted independently, reducing travel cost and facilitating international collaboration. Since multiple raters can analyze the same data, recordings allow a retroactive assessment of inter-rater reliability.
Low-cost video-based observations systems present an efficient and flexible method to collect naturalistic road user behavior data in LMIC.