A novel Computer-vision Deep learning-based Solution Platform for Truck Parking Information System

Abstract: In this paper, we present a novel computer-vision and deep learning-based solution platform designed to address the critical issue of truck parking availability along I-80 rest area truck parking lots. Our platform integrates advanced algorithms, including YOLOv10 and SAHI, with image stitching techniques to deliver real-time, highly accurate information on parking spot availability. The system employs already available CCTV Pan Tilt Zoom (PTZ) cameras installed in the parking lots, capturing images from various angles. The processed data is then relayed to the Iowa DOT website and other third-party applications, providing truck drivers with timely information to optimize their routes. This real-time information system not only enhances operational efficiency and reduces fuel consumption but also contributes to environmental sustainability by minimizing unnecessary idling and emissions. Traditional methods for monitoring truck parking availability, such as in-ground sensors, often fall short in delivering accurate and reliable data due to the harsh weather conditions prevalent in the Midwest. These sensors are susceptible to malfunctions caused by extreme temperatures, heavy snowfall, ice, and flooding, which can interfere with their ability to function correctly. As a result, the data provided by these systems can be inconsistent and unreliable, leading to significant challenges for truck drivers who depend on accurate information to plan their rest stops and optimize their routes. A robust, weather-resistant system utilizing advanced computer vision and deep learning algorithms offers a promising alternative, ensuring that truck drivers have access to the consistent and accurate information they need, regardless of the weather conditions.
Keywords: Deep Learning; Computer Vision; Truck Parking; SAHI; Machine Learning; Parking Management