<h1>De Waaiburg App</h1>
<p>The De Waaiburg App project aimed to expand an existing app. Upon evaluating the original app, we identified suboptimal code and vulnerabilities, leading us to start from scratch. The project involved developing an app with an open section containing information about De Waaiburg, an organization for Special Youth Care. Additionally, we planned to create a closed section for tracking a client's journey within De Waaiburg.</p>
<h2>App Development</h2>
<p>With limited prior experience in Laravel and Flutter, we embarked on this project as an opportunity to learn and grow. We chose Flutter as the framework for the main app, ensuring a seamless and user-friendly experience across different platforms. For the web app, we explored Laravel 9, developing the necessary skills to build an interactive and comprehensive online platform.</p>
<h2>Open Section: Information about De Waaiburg</h2>
<p>The open section of the app provides users with valuable information about De Waaiburg, an organization dedicated to Special Youth Care. This includes details about the organization's mission, vision, services offered, and contact information. Users can easily navigate through the app to explore various aspects of De Waaiburg and gain a comprehensive understanding of its operations.</p>
<h2>Closed Section: Work in Progress</h2>
<p>The closed section of the app, dedicated to tracking a client's journey within De Waaiburg, is currently under development and scheduled for implementation in the next semester. This section will feature a unique component called "Leefboom" (Tree of Life), where clients can engage in self-reflection by answering open-ended questions. As they progress and resolve these questions, a virtual tree will grow, symbolizing their personal growth and achievements. Once the tree is fully grown, clients will have the opportunity to review and reflect on their trajectory within De Waaiburg, fostering valuable insights and experiences.</p>
<h2>Conclusion</h2>
<p>The De Waaiburg App project provided us with an invaluable learning experience. Despite having limited prior knowledge of Laravel and Flutter, we embraced the challenge and acquired new skills throughout the development process. While the closed section of the app is still a work in progress, we are excited about its potential to enhance the client experience within De Waaiburg.</p>


Enhanced app for tracking client journeys and providing information about Special Youth Care services.

De Waaiburg App

<h1>Race To The Future</h1>
<p>Race To The Future is a project where we transformed an RC car to enable it to drive autonomously and participate in a race. We were provided with an unmodified RC car and a set of components that we could install on the vehicle. The race was hosted by Thomas More.</p>
<h2>Hardware Optimization</h2>
<p>To achieve autonomous driving, we performed hardware tweaks on the car. We added a Raspberry Pi to control the entire vehicle, a camera to keep the car on the track, and a PWM board to control the motors accurately.</p>
<h2>OpenCV Lane Detection</h2>
<p>To ensure that the car stays on the track, we utilized OpenCV Lane Detection. This advanced computer vision technology ran on the Raspberry Pi and allowed us to detect the lane markings and adjust the steering input accordingly.</p>
<h2>OpenCV Topics</h2>
<p>During the project, we focused on four major topics within OpenCV:</p>
<ul>
<li>
<p><strong>Filter (HSV):</strong> We applied a filter to the car's camera image using HSV (Hue, Saturation, Value). This enabled us to extract specific color segments and optimize the image processing.</p>
</li>
<li>
<p><strong>Canny (Edge Detection, Contrast):</strong> We utilized the Canny algorithm for edge detection and contrast enhancement in the image. This was crucial for accurately detecting the lane markings.</p>
</li>
<li>
<p><strong>Region of Interest (ROI):</strong> By defining a specific region of interest (ROI) in the camera image, we could concentrate on the relevant parts of the track. This improved the performance and efficiency of the lane detection.</p>
</li>
<li>
<p><strong>Hough Lines (Connecting Points):</strong> Using the Hough Lines algorithm, we connected points on the lane markings. This allowed us to determine a smooth and consistent path for the car.</p>
</li>
</ul>
<h2>Challenges and Optimizations</h2>
<p>During the practical implementation of the project, we encountered various challenges. Working with different Python libraries that didn't always cooperate seamlessly required us to fine-tune each library individually. We adjusted specific formulas and parameters to achieve the desired functionality. We also implemented optimizations such as using a Rolling Average method with weights to prevent excessive steering. Throughout the process, we observed occasional spikes in processing load, so we made some adjustments to enhance performance.</p>
<h2>Testing</h2>
<p>On the first test day, we took the car outside for testing. We adjusted the contrast to improve the visibility of the lane markings. Additionally, we made other modifications, including introducing a delay for the camera to stabilize the image during startup and enlarging the ROI for better overall visibility.</p>
<h2>Results</h2>
<p>During the race, we faced challenges, particularly due to sunny weather causing strong shadows and glare. Nevertheless, we achieved an outstanding second place. This is a remarkable result considering the competition involved two teams from the bachelor's program (including our team of second-year students) and two teams from the master's program.</p>
<h2>Collaboration and Pride</h2>
<p>We completed this project in collaboration with Sens and Nick, and we take great pride in our accomplishments. It provided us with an opportunity to further develop our skills in autonomous driving, hardware optimization, and image processing. We eagerly look forward to new challenges and projects that will allow us to expand our expertise. We chose to use OpenCV for this project because we had limited experience with AI, and OpenCV provided powerful computer vision capabilities that suited our needs.</p>


Autonomously driving RC car project incorporating Raspberry Pi, OpenCV Lane Detection, and PWM control for motors.

Race To The Future

<h1>PeopleCounter</h1>
<p>The PeopleCounter project is a comprehensive solution designed to address the challenge of monitoring and managing classroom occupancy in the DI (Digital Innovation) department. The project revolves around utilizing advanced technologies to provide accurate and real-time data on the number of people present in the classroom. This data serves as valuable evidence to support decision-making processes and address issues related to classroom congestion.</p>
<h2>Project Overview</h2>
<p>In response to the observed high levels of classroom congestion, a DI coach requested a solution that could accurately track the number of individuals present in the classroom and provide data-backed evidence for requesting additional classrooms during peak periods. To fulfill this requirement, we developed the PeopleCounter project, which incorporates the following key components:</p>
<ul>
<li>
<p>Raspberry Pi with camera: The Raspberry Pi is equipped with a camera and runs YOLOv5, a state-of-the-art computer vision model capable of person detection.</p>
</li>
<li>
<p>FastAPI: We developed a custom API using FastAPI, a high-performance web framework for building APIs with Python. The FastAPI handles the communication between the Raspberry Pi and the server.</p>
</li>
<li>
<p>Proxmox VM server: The Proxmox VM server hosts the FastAPI server, which receives the person count data from the Raspberry Pi and writes it into a dedicated database.</p>
</li>
<li>
<p>Discord bot integration: To ensure easy access to the occupancy data, we implemented a Discord bot that interacts with the API. This allows DI students and staff to view the real-time person count directly on our Discord server.</p>
</li>
<li>
<p>UBEAC integration: The PeopleCounter project also integrates with UBEAC, a powerful platform that provides a visually appealing interface and generates insightful graphs showcasing the occupancy trends over time. This integration enhances the ability to present data-backed evidence for decision-making and resource allocation.</p>
</li>
</ul>
<h2>Achieving Data-Backed Insights</h2>
<p>The PeopleCounter project not only fulfills the specific feature requested by the DI coach but offers a comprehensive solution to monitor classroom occupancy. By leveraging computer vision, API development, and data visualization techniques, we provide accurate and real-time data on the number of individuals present in the classroom. This data enables the DI department to identify peak periods of congestion, evaluate classroom capacity, and request additional resources based on concrete evidence.</p>
<h2>Conclusion</h2>
<p>The PeopleCounter project has been a valuable learning experience for our team. We had the opportunity to work with cutting-edge technologies such as YOLOv5 for person detection, FastAPI for API development, and integration with platforms like Discord and UBEAC. Through this project, we gained practical knowledge and skills in computer vision, API development, and data visualization. By providing accurate and real-time data on classroom occupancy, PeopleCounter enables the DI department to make informed decisions and address issues related to classroom congestion. We are proud of the practical solution we have developed and look forward to applying our newfound expertise to future projects in the field of digital innovation.</p>


Real-time person detection solution for monitoring classroom occupancy and addressing congestion in the DI department.

Recent Projects