AntiTow

Inspiration

We got inspired to create this when one of our team member walked out of his friend's apartment and his car was not where he parked it. After the initial wave of panic, he realized that there were two options for where his car could be. It could either be stolen or towed. After spending hours figuring out it was the latter, He would then spend more time arranging for and picking up his car from the towing company. With this time without a car, he found himself researching the unregulated and sometimes predatory practices of towing. He realized when a car is towed, the only notification a tow company has to give is to the local police thus creating a situation where one is left without a mode of transportation. Stranded. These uncomfortable situations alongside the possibility of getting wrongfully towed inspired our team to find a solution to a problem that federal policies have yet to solve.

What it does

Our project aims to utilize the cameras of these vehicles, and use computer vision to analyze vehicles passing by and pinpoint the logos of towing vehicles that have been stored in a database. When a logo has been detected for a sufficient amount of time near the camera of a user’s car, our project then utilizes the Twilio API in order to automatically send the user an SMS notification, notifying and alerting the user of a potential towing event.

How we built it

We built the application using HTML/CSS/JS and the Computer Vision was done using Opencv, an open source library for multiple computer vision applications. Utilizing Opencv, our project initializes a python script that compares a sample logo to instances of that logo in the environment. This python script uses a brute force matching algorithm in Opencv in order to match key points and descriptors of the first image to the training image. Our team deduced that if enough similarities were found, this would be substantial for a text notification to be sent to the user utilizing the Twilio API.

Challenges we ran into

Although our project makes great leaps in solving a problem for many citizens around the world, we ran into issues regarding time constraints and coming up with a functional application. Our computer vision model could be improved to be more precise and provide users with a better experience. The problem, our team set out to solve is rooted in an $8 billion a year industry backed by federal policies and therefore, the computer vision implementation alongside a convenient user experience is needed in order to prevent wrongful towing. As this problem is pervasive, our project requires a database of different towing company logos that could not be implemented at this time.

Accomplishments that we're proud of

We are proud that we came up with a solution to a problem that goes widely unregulated and causes a lot of stress for the people of the world. We were able to find a use case for computer vision that solves a seemingly mundane problem in a streamlined and convenient way. We were able to explore computer vision, SMS text messaging and UI/UX design through our project in order to create an experience that helps citizens and improves autonomous/self-driving vehicles.

What we learned

We learned to create a minimalist and sleek UI/UX design that utilizes computer vision to solve a problem. This problem required innovation and research in computer vision and design. Building upon our knowledge of programming, utilizing computer vision in this project further broadens the uses of technology that are available to our team. -To create a minimalist UI/UX design -A computer vision implementation that matches logos to their real-life counterparts identifying them in a training image. -Utilization of the Twilio API to send SMS text messages to users in a streamlined and convenient manner.

What's next for Tow

We intend, in the future to make a variety of improvements to our project such as a mobile version of the application with enhanced features such as a car backup camera feed in order to assess the situation after a user gets a notification and AI to not only determine changes in the orientation of the car when it is hitched but also create a more precise computer vision model. This increase in precision will further improve users' experiences and make our project a substantially more convenient solution.