Automated visual inspection of dry carbon-fibre reinforced composite preforms

A vision system is described which performs real-time inspection of dry car bon-fibre preforms during lay-up, the first stage in resin transfer mouldin g (RTM). The position of ply edges on the preform is determined in a two-st age process. Firstly, an optimized texture analysis method is used to estim ate the approximate ply edge position. Secondly, boundary refinement is car ried out using the texture estimate as a guiding template. Each potential e dge point is evaluated using a merit function of edge magnitude, orientatio n and distance from the texture boundary estimate. The parameters of the me rit function must be obtained by training on sample images. Once trained, t he system has been shown to be accurate to better than fl pixel when used i n conjunction with boundary models. Processing time is less than I s per im age using commercially available convolution hardware. The system has been demonstrated in a prototype automated lay-up cell and used in a large numbe r of manufacturing trials.