Report 1: 'Blind-Spot-Free Lens Defect Detection Based on Spherical Lens Surface Dark-Field Imaging Simulation'

Spherical lenses are a widely used type of optical component, highly susceptible to various surface imperfections during manufacturing and use. For the machine vision-based digital inspection of such ultra-smooth optical component surfaces, dark-field imaging is often employed due to the high contrast between defects and the background. Under the effect of diffraction broadening, detectable scratch sizes can often reach the sub-micrometer level.

In this inspection method, the spherical surface itself can easily disrupt the dark field of the illumination source. Moreover, lenses comprise two or more surfaces; as the number of surfaces increases, reflected light from the light source is more likely to interfere with defect imaging. This paper models common dark-field imaging inspection layouts and simulates multiple reflection imaging of spherical lenses, leading to a proposed blind-spot-free lens detection method. We simulated this blind-spot-free lens defect detection experiment using surface inspection vision system modeling and Bidirectional Reflectance Distribution Function (BRDF) modeling for defects.

During the conference coffee break, Professor Yang Yongying held a book signing and gifting event for the core work from the national '14th Five-Year Plan' Major Publishing Project, 'Revolutionary Optical Science and Technology Series' — 'Optical Imaging and Defect Evaluation in Machine Vision' (Tsinghua University Press).

A Collaborative Masterpiece

This specialized book, spearheaded by Professor Yang Yongying, and co-authored by key team members including Senior Engineer Cao Pin and Dr. Jiang Jiabin over three years, encapsulates the team's more than 20 years of scientific research accumulation in the field of optical defect detection:

Construction of a Theoretical System: Professor Yang Yongying led the establishment of the FDTD scattering model and the national standard for quantitative evaluation.

Breakthrough in Industrial Algorithms: Senior Engineer Cao Pin overcame the industrialization bottleneck in deep learning defect recognition.

Innovation in Simulation Technology: Dr. Jiang Jiabin developed BRDF modeling and blind-spot-free detection methods.

Ending an Era of Ambiguity

This monograph, presented as a 'three-part symphony' of 'Theoretical Foundation - Technological Breakthrough - National Standard Setting,' definitively ends the historical lack of standards for optical surface defects. It distills the FDTD scattering model, blind-spot-free detection, and deep learning algorithms into a reproducible scientific paradigm and an actionable industrial code. This lays a millimeter-precision academic groundwork for China's transition from a major optical manufacturing nation to a powerhouse in inspection technology.