Reflection compensation and binary thresholding using the Homomorphic Filtering approach for composite manufacturing

The global aerospace manufacturing industry is witnessing a major paradigm shift in design and material choice from metals to composites such as Carbon Fiber Reinforcement Plastics (CFRP) in civil and military aircraft structure design. In prepreg-based composite fabrication, multiple sheets of prepreg material are manually laid in the desired orientation on a mold. The layout process is subjected to 100 percent ply-by-ply manual visual inspection for detecting inclusion which can arise when the protective films are missed to get removed or due to the tackiness of the prepregs layers, other materials used in the process get included between layers. Manual inspection during layup is skill-based. The black and reflective nature of the surface makes it difficult to detect the inclusions efficiently. Thus, there exists a need for an automated or semi-automated system that can aid the inspector in detecting and identifying inclusions during fabrication. This paper inculcates the preliminary work carried out towards developing a methodology to identify the inclusions using image processing algorithms. One of the challenges to using the images is the reflection from the layers. A homomorphic filtering-based approach was developed to reduce the effects of reflection and binary thresholding to detect the inclusion regions. The developed methodology was validated through experimental results. It is found that Homomorphic filtering-based reflection compensation and binary thresholding can distinguish the inclusion regions, which can be used by the inspector in the clean room environment.