Plastic injection molded Fresnel lens is one of the important components for illumination in smart devices. To perform inspection on this type of optical component is challenging for machine vision due to the presence of groove pattern and texture. This paper discusses the limitation of classical image analysis for defect inspection and proposes a Deep Convolutional Neural Network (CNN) with Transfer Learning for defect classification. This paper also presents a Hybrid CycleGAN and geometric augmentation to expand image dataset for model training.

Smart devices components come in a variety of shapes, sizes and texture. To perform a high throughput inspection on all the six sides of the surface of the component is challenging due to tolerances of product dimension and the multiple planes on its surface. The image acquisition system for surface inspection comprises of a CMOS camera, telecentric lens, LED light source and motorized actuator. HALCON vision library was used for image processing. One of the most important requirements of an effective Automated Vision Inspection (AVI) is consistent part presentation. Thus, to perform the positioning correction we propose two position recognition method using 2D-metrology and Smallest Rectangle Segmentation. The result of the position recognition was fed to XY rotary table for offset correction before picked up by turret fingers. Actuators for camera focusing were added to increase the flexibility of the system to inspect various component sizes. The Exposure End event of a modern CMOS camera was used for parallelization to optimize the machine vision cycle time.

Characterize driving behaviors in people with migraine from the ObserVational survey of the Epidemiology, tReatment and Care Of MigrainE (OVERCOME) study.