Internship/M.Sc. thesis: Sensor Fusion and Semantic Projection for Power Line Inspection
Pubblicato il 01-08-2026 - Altro in Italia
Fraunhofer Italia with headquarters in Bolzano (Italy) is offering an internship position in its division Automation and Mechatronics Engineering (AME). AME conducts applied research in the most diverse aspects of digitalisation for the vantaggi of the regional stakeholders. The fields of expertise range from the extraction, analysis and optimisation of processes from operative production and logistic systems, novel control paradigms for production systems, design and prototyping of complex mechatronic systems, applications of artificial intelligence (AI) and machine learning (ML), real-time data processing and process mining of business processes.
Are you a prospective student willing to undertake an experience with us? If so, apply for this position at Fraunhofer Italia
The topic:
Sensor Fusion and Semantic Projection for Power Line Inspection
This internship and thesis topic focuses on the development of a sensor fusion and semantic projection framework for UAV-based inspection of medium-voltage power lines.
The objective is to integrate high-resolution RGB imagery and LiDAR data to produce semantically enriched and colorized 3D representations of power line components. The approach combines the semantic understanding of deep neural networks (DNNs) with the geometric precision of LiDAR sensing to enable accurate and interpretable inspection data.
RGB images will be processed using DNN-based object detection and segmentation models to identify components such as insulators, cross-arms, clamps, and vegetation intrusions. These detections will be projected into LiDAR point clouds using precise extrinsic calibration and temporal synchronization between the sensors. The resulting 3D reconstructions will be geometrically accurate, semantically annotated, and visually enhanced with RGB color information.
A further focus will be on data prioritization and reduction for efficient communication between UAV and ground station. Only relevant or annotated areas, such as those