For the evaluation of manufacturing skills, as represented by the National Trade Skill Test & Certification, two major challenges exist: variability in evaluation results and the complexity of establishing evaluation criteria. To address these challenges, a need exists for technology that can assess the quality of printed circuit board (PCB) assemblies automatically based on photographs of the assembled products and their corresponding work instructions. Therefore, this study proposes a technique for linking product photographs and work instructions by identifying the positions of PCB components in the images and their corresponding component numbers. The proposed method establishes a connection between the image-based information (photographs) of the manufactured product and the text-based information (work instructions) by first recognizing components in the photographs and then identifying component numbers listed in the instructions. This approach aims to reduce the manual workload associated with component identification. The effectiveness of the proposed method was evaluated using PCBs of three types. The findings confirmed that components and their respective names can be linked with an accuracy of 81%.