License Plate Scanner 5000

Inspiration

Daryl: Growing up in Indonesia, family road trips were a cherished tradition. However, these trips often involved navigating long stretches of highways with toll booths. Watching the car barely move, listening to the repetitive honking of cars, and seeing the sunlight fade, I could see my parents’ frustration growing with each passing minute as we inched forward ever so slowly.

It was then that the realisation hit me: these toll booths were a significant obstacle in a typical travel experience. I never understood how a simple payment could lead to such a frustrating, prolonged ordeal.

However, this experience provided me with the opportunity to recognize and identify small improvements that could alleviate delays and eliminate possible frustrations we may face.

Aidan: One day, my friends and I drove to a nearby grocery store. After shopping for approximately 30 minutes, we were stuck in a long, chaotic queue trying to exit the parking lot.

Although licence plate recognition technology is well-implemented in many areas, frustrating experiences like this occur quite often in various locations. This particular instance sparked the idea for a system using licence plate recognition to streamline parking lot exits, allowing for a more seamless and less stressful experience. Such a recurring problem inspired a solution to a common parking issue.

Yushan: Spotting different licence plates was a game my family and I used to play. That childhood game sparked my curiosity and drove a newfound sense of adventure.

I leveraged this early fascination to drive real-world innovation, transforming playful curiosity into technology that enhances road safety, streamlines traffic, and connects communities. Cherishing the excitement with my family in my younger years, this challenge has stimulated my keenness towards developing a licence plate detection and reader system.

Letting my childhood wonder fuel my passion has allowed me to turn my dreams into impactful technology, one plate at a time

Overall: Our stories and efforts have brought us together to idealise and innovate, with the ultimate goal of achieving a common vision and turning our ideas and imagination into reality.

What it does

Our product is designed to efficiently capture and process licence plate information. It utilises basic Python features that can be adapted to solve a wide range of modern issues. It scans for licence plates, reads the characters, and securely stores data for future use. This solution is ideal for applications in environments involving parking and vehicle management, security, and data analytics.

How we built it

Our journey began with a deep dive into the challenges we had faced. After hours of discussion, we were able to identify a critical need that is yet to be adequately addressed.

With a clear understanding of the problem, we decided to create a product specifically designed to solve the existing traffic issues. Our focus was to develop a solution that addresses our concerns and is sufficiently flexible to support a broader audience.

As we had limited experience in AI, we spent the first two days learning and discussing how to create and implement our product. We determined that we should design our own Convolutional Neural Networks (CNN) model to detect licence plates, utilise existing tools to pinpoint the corners of a licence plate, and import existing Optical Character Recognition (OCR) to read the presented text. Combining these elements enabled us to develop the desired product.

Challenges we ran into

Learning and implementing AI without prior background knowledge was definitely challenging. Moreover, all three of us had numerous assignments due during the course of this project. However, we were committed to our goal and implemented AI to the best of our abilities. Time management was crucial in overcoming these challenges. Each of us had to allocate time for our assignments, followed by working on this project. We ensured effective communication and collaboration in setting timelines, milestones, and tasks.

Accomplishments that we're proud of

Considering this was our first hackathon, we excelled in communication, engagement, and shared excitement to produce our best work. Tackling the logical and syntax errors due to our limited knowledge of AI was challenging, but we managed to resolve them successfully. Despite the difficulties, we are proud to have developed a fully functional product that can be used in many different contexts. And it’s the struggle like these that make our effort seem valuable in the eyes of our own, regardless of the result. And at the end of the day, we all have gained valuable skills that may be needed in the coming future.

What we learned

Approaching a problem with some existing knowledge provided us with a base understanding of different types of AIs and how they might behave. This included working with convolutional neural networks, leveraging external libraries, and dealing with deep learning models. Additionally, integrating and deploying the model in a productive environment was the biggest takeaway from this experience.

What's next for License Plate Scanner 5000

To further enhance our program, we plan to improve accuracy by expanding our database of licence plate images, allowing the AI to train on a more diverse set of data. Implementing post-processing will refine the system’s ability to distinguish between the actual licence plate number and additional text, such as state names, ensuring accurate data recording. We are also exploring the possibilities of optimising the algorithm for various lighting and weather conditions to ensure the system remains reliable in an ever-changing environment.