| KEYWORD |
Developing Vehicular Traffic Datasets Using Drone images and AI Techniques
Reference persons CLAUDIO ETTORE CASETTI
Description Urban traffic management represents a challenge in cities, with significant implications for safety, efficiency, and environmental sustainability. Artificial Intelligence (AI) and Machine Learning (ML) technologies have demonstrated great potential in addressing these challenges, particularly through traffic monitoring and predictive analytics. However, the performance of AI systems heavily depends on the quality and completeness of the datasets used for the training phase. Current traffic datasets often fall short due to their limited coverage, lack of diversity in traffic patterns and conditions, and insufficient adaptability to real-time scenarios. This thesis proposes the creation of advanced vehicular traffic datasets by utilizing drones equipped with high-resolution imaging systems, combined with cutting-edge image processing techniques and post-processing for dataset organization.
Captured data will be processed through AI-powered techniques to detect and classify vehicles, pedestrians, cyclists, and other entities. The workflow will involve traffic flow analysis, scene segmentation, and trajectory tracking.
The outcomes of this research are expected to support advancements in autonomous vehicle navigation, smart city planning, and intelligent transportation systems. By making these datasets available to the research community, this work would contribute to the broader adoption of AI-driven solutions for sustainable and efficient urban development.
Required skills Familiarity with basic AI and ML techniques and related frameworks such as TensorFlow, PyTorch, or scikit-learn. Good programming skills in Python. Familiarity with object detection techniques in digital images (e.g., YOLO).
Deadline 27/11/2025
PROPONI LA TUA CANDIDATURA