[Object] To find, with high sensitivity, an unevenness defect or the like that has occurred on the surface of an inspection object having a surface roughness comparable to wavelengths of visible light and is comparable to several times the surface roughness, and accurately distinguish between dirt and an unevenness flaw present on the surface of the inspection object, and also enable a reduction in the size of an apparatus.

[Solution] An inspection object imaging apparatus according to the present invention includes: a light source configured to produce a light beam belonging to an infrared wavelength band and having a predetermined spread half-angle on a surface of an inspection object; a projection optical system configured to project the light beam on the surface of the inspection object at a predetermined projection angle; and an imaging unit configured to image reflected light from the surface of the inspection object. The imaging unit includes an imaging optical system including at least one convex lens, configured to condense reflected light and branch the reflected light to two different directions, and a first image sensor and a second image sensor each configured to image the reflected light that has passed through the imaging optical system. The first image sensor is positioned on the inspection object side with respect to a position of the imaging optical system that is conjugate with the surface of the inspection object, along an optical axis of the reflected light. The second image sensor is positioned on the reflected-light travel direction side with respect to the conjugate position.

A learned model generation method includes: using a teacher image including a defect map that is an image indicating a distribution of a defect portion of a surface of steel and having an equal image size, and presence/absence of periodic defects assigned in advance to the defect map; and generating a learned model by machine learning, the learned model for which: an input value is a defect map that is an image indicating a distribution of a defect portion of a surface of steel and having an image size of the equal image size; and an output value is a value concerning presence/absence of periodic defects in the defect map.

A method of detecting defects in a rolling element is provided herein. The method includes collecting visual data or information via a microscope assembly. This visual data is then processed and algorithms, filters, or other analytical tools or engines are applied to detect if any defects are present on the rolling element. The method includes automatically flagging or identifying any defects, and this step is then checked or verified by a user or different entity. Depending on input from the user, the defects can either be confirmed or can be identified as a false detection. Information from the user's decision-making process is then fed back into the system, processors, algorithms or other analytic aspects of the system. This information is then used to improve the accuracy of the detection algorithms. This disclosure provides an automated system and process for more efficiently and reliably identifying defects on rolling elements.

Methods and systems are disclosed for analyzing one or more images of a textile to determine a presence or absence of defects. In one example, an image of at least a portion of a textile may be obtained and compared to a reference image of a reference textile. Based on the comparison, one or more areas indicative of a height variation between the textile and the reference textile may be determined. An action may be performed based on the one or more areas indicative of the height variation.

Methods and systems are disclosed for analyzing one or more images of a textile to determine a presence or absence of defects. In one example, an image of at least a portion of a textile may be obtained and compared to a reference image of a reference textile. Based on the comparison, one or more areas indicative of a height variation between the textile and the reference textile may be determined. An action may be performed based on the one or more areas indicative of the height variation.