Inspection systems with a machine vision camera have been extensively applied in factory automation detecting unclear defects. In this automatic inspection, optical engineering technology needs to emphasize the contrast of defects in order to quickly find them in a camera image and improve image recognition accuracy. Using the phase-shift illumination method with striped structured illumination, we develop optical engineering and image processing technology to enhance defects in transparent materials. Our challenge has been to distinguish actual defects from dark fringes due to artificial three-dimensional texture on a target sample. We proposed theoretically that an innovative method providing an illumination pattern with finite-width dark regions would make the gentler slope of artificial texture less visible than actual defects. We extended our theoretical model in the case of a square wave as the typical illumination pattern and established an inspection method distinguishing defects from artificial texture. We confirmed with simulations and experiments that the square-wave illumination enables us to distinguish between defects and artificial texture.

带有机器视觉相机的检测系统已广泛应用于工厂自动化检测不明缺陷。在这种自动检测中，光学工程技术需要强调缺陷的对比度，以便在相机图像中快速找到缺陷并​​提高图像识别精度。我们利用相移照明方法和条纹结构照明，开发光学工程和图像处理技术来增强透明材料的缺陷。我们面临的挑战是区分实际缺陷和目标样品上人造三维纹理造成的暗条纹。我们从理论上提出，一种提供具有有限宽度暗区的照明图案的创新方法将使人造纹理的平缓坡度比实际缺陷更不明显。我们在方波作为典型照明模式的情况下扩展了我们的理论模型，并建立了一种区分缺陷和人造纹理的检查方法。我们通过模拟和实验证实，方波照明使我们能够区分缺陷和人造纹理。