Intelligente Erfassung der Parkraumbelegung mit einem Multikamerasystem und CNNs
Student Assistant (m/f/d)
Intelligent Parking Space Detection Using a Multi-Camera System and CNNs
Description
The efficient detection of parking occupancy is a key component of smart traffic management systems. This project aims to utilize a multi-camera system to monitor a parking deck in Erlangen, employing Convolutional Neural Networks (CNNs) to accurately analyze and process occupancy data.
A CNN is a specialized artificial neural network optimized for processing and recognizing patterns in image data. It will be used to automatically detect vehicles in the camera feed and determine their positions within the parking deck. The system operates within a 5G campus network, enabling real-time data processing and transmission.
Also available as: Research Internship or Thesis (Master’s/Bachelor’s)
Research Questions
- How reliably can CNNs be used for parking space detection?
- What challenges arise from various environmental conditions (e.g., rain, shadows)?
- How can a multi-camera system be optimally configured for data fusion?
Objectives
The goal of this project is to maintain and monitor the experimental setup, collect and annotate vehicle data, and train and optimize a CNN for robust vehicle detection. The system will be tested under real-world conditions and continuously improved.
Tasks
- Maintaining and monitoring the experimental setup (multi-camera system, sensors)
- Processing collected data and annotating vehicle images
- Training CNNs for vehicle detection
- Conducting literature research on existing methods and optimization strategies
Your Skills
- Interest in computer vision and AI
- Basic knowledge of deep learning and neural networks (CNNs)
- Experience with Python (e.g., TensorFlow, PyTorch, OpenCV)
- Familiarity with data annotation and image processing is a plus
If you are interested in this exciting research opportunity, please send your application with a short resume to