Sensing and Vision for Industry and Environment

· Image acquisition: image formation (sensor, pinhole, lens), lighting and sensor characterization (color, exposure, speed), vision in industry and environment (contactless monitoring, industrial cameras, non-ideal settings).
· Pattern recognition for vision systems: AI for image preprocessing (quality analysis, enhancement), segmentation (object detection, semantic segmentation, pixel-level annotations), 3D reconstruction (multiple-views, structured light), 2D/3D feature extraction (handcrafted, representation learning), classification and regression (nearest neighbor, neural networks, convolutional neural networks).
· Industrial monitoring: AI for vision-based monitoring of manufacturing process (detection of machinery fault, detection of tool defects, guidance of assembly lines), analysis of raw materials (volume estimation, granulometry measurement), product quality control (surface defects detection, assembly errors, predictive maintenance), virtual sensors (vision-based depth estimation, synthetic environments), human safety monitoring (person tracking, incident detection).
· Environmental monitoring: AI for processing images captured using centralized vision (detection of fire and smoke, flood and drought, landslides, structural health monitoring), monitoring using images acquired with distributed vision (crop condition analysis, wildlife monitoring, traffic monitoring, vehicle accident detection, waste and illegal drop-off detection).