Comparing YOLO-Based Parking Space Detection Performance Between GPU, Non-GPU, and Raspberry Pi Systems
Document Types
Paper Presentation
Research Theme (for Paper Presentation and Poster Presentation submissions only)
Computer and Software Technology, and Robotics (CSR)
School Name
De La Salle University, Manila
Track or Strand
Science, Technology, Engineering, and Mathematics (STEM)
Research Advisor (Last Name, First Name, Middle Initial)
Cempron, Jonathan Paul, C.
Start Date
25-6-2026 10:30 AM
End Date
25-6-2026 12:00 PM
Zoom Link/ Room Assignment
Online - https://zoom.us/j/91936856247?pwd=oCMfMsh44I2wb0dYsEgoInDJy59bOq.1 Meeting ID: 919 3685 6247 | Passcode: research
Abstract/Executive Summary
With the development of smart parking management systems, automatic parking space detection using object detection has become an effective approach for managing parking lots. Existing technologies, such as ground sensors and ultrasonic sensors, face challenges related to scalability, cost, and environmental sensitivity, creating the need for more reliable and efficient alternatives. While previous computer vision based studies have focused on improving model architectures, this work focuses on the performance differences of object detection across different hardware systems. This study evaluates YOLO for parking slot detection using three hardware platforms: a Raspberry Pi 5, a non-GPU system, and a GPU system. The model was developed and trained using the publicly available PKLot dataset and assessed based on training time, inference time, IOU, recall, precision, and F1 score. Results showed that the GPU system achieved the best overall performance, particularly in training and inference speed, demonstrating the effectiveness of hardware acceleration in object-detection.
Keywords
computer vision; object detection; vehicle detection; YOLO; NVIDIA CUDA GPU
Initial Consent for Publication
yes
Statement of Originality
yes
Comparing YOLO-Based Parking Space Detection Performance Between GPU, Non-GPU, and Raspberry Pi Systems
With the development of smart parking management systems, automatic parking space detection using object detection has become an effective approach for managing parking lots. Existing technologies, such as ground sensors and ultrasonic sensors, face challenges related to scalability, cost, and environmental sensitivity, creating the need for more reliable and efficient alternatives. While previous computer vision based studies have focused on improving model architectures, this work focuses on the performance differences of object detection across different hardware systems. This study evaluates YOLO for parking slot detection using three hardware platforms: a Raspberry Pi 5, a non-GPU system, and a GPU system. The model was developed and trained using the publicly available PKLot dataset and assessed based on training time, inference time, IOU, recall, precision, and F1 score. Results showed that the GPU system achieved the best overall performance, particularly in training and inference speed, demonstrating the effectiveness of hardware acceleration in object-detection.
https://animorepository.dlsu.edu.ph/conf_shsrescon/2026/BoA_CSR/7