画像処理を用いた検品支援ツールの精度改善に関する研究

抄録

Many quality inspection tasks have been mechanized and automated in the quality control departments of manufacturing companies. On the other hand, manual inspection is essential for products that cannot be automated. For example, the tips of rotary tools used in dental treatment are covered with fine diamond particles by electrodeposition, so no one has the same shape. We have researched inspection support tools for rotary tools using machine learning. However, it has become clear that it is difficult to discriminate samples that are difficult to judge visually with high accuracy, even with machine learning. In this study, we aim to improve the discrimination system by applying filter-based preprocessing to image data.