Cosmetic puff image inspection device
"Inspection with 6 cameras" Enhancing quality and efficiency with the power of images to support accurate judgment.
"Introduction of a device that inspects the scratches, dirt, and unfoamed areas of cosmetic puffs and stacks only the good products." The puffs sent from the manufacturing process are aligned in a single row on a conveyor belt, and the thickness of the front and back of the puffs, as well as four edges, are inspected using six cameras. ■ Image Acquisition High-resolution cameras and macro lenses are used to capture detailed images of the puff's surface. Uniform diffused lighting is employed to minimize shadows and reflections, making scratches and dirt more visible. ■ Image Preprocessing Noise removal and contrast adjustment are performed to facilitate the detection of abnormal areas. ■ Feature Extraction and Anomaly Detection Machine learning models and image processing algorithms are used to identify patterns of scratches and dirt. Differences in color, texture anomalies, and surface irregularities are detected. ■ Non-contact and Non-destructive Inspection Since the inspection can be conducted without damaging the products, it contributes to quality maintenance and cost reduction. ■ Judgment and Classification Using a pre-trained model, the device automatically determines whether the product is "good" or "defective." Defective products include scratches, dirt, unfoamed areas, and foreign matter contamination. ■ Result Output and Management Inspection results are transmitted in real-time to the line control system for the elimination and recording of defective products.
- Company： 真田製作所 埼玉三郷工場
- Price：Other
