Level of Education of Students Involved

Undergraduate

Faculty Sponsor

Sami Khorbotly

College

Engineering

Discipline(s)

Computer Vision, Machine Learning

ORCID Identifier(s)

Charles Smith, 0009-0000-9036-5182

Presentation Type

Poster Presentation

Symposium Date

Spring 4-25-2024

Abstract

Parking assessments are frequently performed to analyze parking lot usage patterns, including peak utilization periods and overall lot distribution throughout the day. Usually these studies rely on tedious manual labor, requiring researchers to physically go through each parking lot and individually count cars in parking spaces multiple times a day. This puts a limit on how much data can be collected and how many lots can be studied. However, recent advancements in technology, particularly the availability of camera-equipped drones and progress in computer vision and deep learning techniques, have facilitated a transition towards high accuracy, cost-effective, automated approaches. Our study introduces a comprehensive computer vision-based solution developed and evaluated for parking utilization studies on the Valparaiso University campus. Our approach involves programming a drone to follow a designated flight path over campus, capturing photos of each parking lot throughout the day. These images are then input into our system which automatically processes them and counts the number of vehicles in each parking lot designation (staff, commuter, resident, etc). To accomplish this, a Python script pre-processes the image before running a pre-trained deep learning model to find the locations of vehicles in each image. Then, the Python script uses these locations to filter vehicle counting based on parking designation. It then generates reports detailing the occupancy of different sections within each parking lot across various timeframes. The results show that the system was able to successfully report the number of vehicles with a 100% accuracy rate.

Biographical Information about Author(s)

I am Charles Smith and I am a junior Computer Engineering major at Valpo. I have always been very interested in programming, but recently have also become intrigued by artificial intelligence. After taking a class on machine learning, I developed a strong interest in this area and this project was a great opportunity to practice. In the future I would like to work with this technology and continue using it to solve complex problems.

Download

Share

COinS