Data technologies in robotic sealing: a systematic review on inspection and quality control

The application of sealants and adhesives is a critical step in modern manufacturing, essential for the structural and functional integrity of products in sectors such as automotive and aerospace, demanding rigorous precision and quality control. This Systematic Literature Review (SLR) analyzes the role of data analysis and digital technologies related to intelligent robotic sealing process. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, 65 studies from the Scopus, Web of Science, and Engineering Village databases were selected and analyzed. The survey showed that data-based techniques in robotic sealing evolved from 2D computer vision methods to Deep Learning (DL) architectures (YOLO family and CNNs), achieving a Mean Average Precision (mAP50) of up to 95.2% and speeds of 189 FPS in detecting complex defects. In trajectory planning, the integration of 3D sensors and registration algorithms (such as ICP and TEAR) enabled adaptive fixtureless systems, reducing the mean width error to 1.5%, compared to 59.2% with traditional methods. The frontier of knowledge lies in the use of Digital Twins to mitigate the shortage of industrial data through synthetic data. A critical gap has been identified in the integration of closed-loop systems capable of making real-time parameter adjustments.