Carbon fibres have become the natural choice as reinforcements for polymer composite materials (PMCs). The non-destructive inspection of dry carbon fibre preforms has the potential to increase the reproducibility and reduce the cost of PMC manufacturing, by identifying defects in dry multilayer preforms prior to resin injection. However, use of optical excitation thermography for inspecting dry carbon fibre preforms that constitute the structural reinforcement precursor in the manufacturing of PMCs is poorly documented in the open literature. In this work, optical excitation thermography was used for inspecting six dry multilayer carbon fibre preforms featuring different textile structures, thicknesses and defects, for the first time. Advanced image processing techniques were used in processing the thermographic data for comparative purposes. In particular, partial least square thermography, as a recently proposed technique, was studied in detail. Finally, the performance of different thermography techniques was analysed in terms of: (1) summarizing the capabilities of image diagnosis/processing techniques by signal-to-noise ratio analysis, and (2) identifying the monitoring modalities most suitable to industrial manufacturing.

碳纤维已成为聚合物复合材料（PMC）增强材料的自然选择。通过在树脂注入之前确定干燥的多层预成型件中的缺陷，对干燥的碳纤维预成型件进行无损检查具有提高重现性并降低PMC制造成本的潜力。然而，在公开文献中很少记录使用光学激发热成像法检查在制造PMC中构成结构增强前体的干碳纤维预成型坯。在这项工作中，首次使用光学激发热成像技术检查了六个具有不同织物结构，厚度和缺陷的干式多层碳纤维预成型坯。出于比较目的，先进的图像处理技术被用于处理热成像数据。特别地，作为最近提出的技术，对局部最小二乘热成像技术进行了详细的研究。最后，从以下方面分析了不同热成像技术的性能：（1）通过信噪比分析总结图像诊断/处理技术的能力，以及（2）确定最适合工业制造的监控方式。