Gray-value Image Matching Method Based on Feature Points for Trademark Defect Detection of Spirit Level

Abstract

In order to achieve trademark defect detection on metal background of a spirit level, a gray-value image matching method based on feature points is introduced to improve the robustness of image matching and avoid image linear changing by camera rotation or placement deviation of spirit level on a conveyor belt which leads to an inaccurate result of the subsequent algorithm in image matching. Before the image matching, FAST and SUFT feature point detection is introduced to preliminarily correct linear image to overcome the impact of the linear transformation and to improve the accuracy and robustness of image matching.