High-speed Image Capture and Cone Detection Using a Raspberry Pi and Camera Module: A Design Science Approach

Abstract

The purpose of this paper is to provide a highspeed image capturing and cone detection algorithm using lowcost devices, such as a Raspberry Pi and Pi NoIR V2 camera module. Little research focuses on high-speed, low-cost image processing. Most existing literature aimed at achieving a high FPS, use expensive tools, such as LiDARs or high-speed cameras. Thus creating a gap in knowledge. Design science was used to develop and evaluate our artifact. For evaluation, controlled experiments were used to gather data. The collected data shows that our artifact is able to capture and detect cones at 30-35FPS with varying, 42-83%, cone detection rate and cone detection accuracy ranging from 9% to 69%. The results show that lowcost, real-time image capture and processing are achievable. The study could be used to provide a cost effective solution for cone detection. Additionally, the solution’s low-cost provides more opportunities to develop and innovate in this area.