Automatic inspection (locating and quality assessment) of centering marks engraved on ophthalmic lenses presents special difficulties. The scene is noisy and blurred, and the engraved marks are weak, low-contrast and partially occluded. To locate the marks, we take advantage of their binary-shaped characteristic to design a model that greatly simplifies the numerical computation of correlation without the need to binarize the scene. Three figures of merit are defined to ensure the robustness of the whole procedure: the contrast of the target against the background, the uniqueness of the correlation peak, and the similarity between target and model. By comparing these figures with suitable threshold levels, decisions can be reached based on existence, multiplicity, and completeness of the marks. The computational simplification enables a low-cost industrial implantation based on a CCD camera, a frame grabber, and a 486/33 PC. A precision of 0.1 mm in positioning is reached with a processing time of less than 1 s.