Hammer-IMS makes progress in machine vision and artificial intelligence as an additional eye in quality control of textiles, plastics and other inline produced products. Recently, Hammer-IMS engineers, in collaboration with the PXL Electronics and ICT campus in Hasselt (Toon Sterckx), developed camera-assisted defect detection based on machine learning. This feature will be a key part in the newest generation of Hammer-IMS’s Edge-Vision-4.0-CURTAIN products. The feature distinguishes defects and classifies these into multiple defect categories. The robust AI-based feature for inline quality control will soon be available, delivering state-of-the-art visual inspection.
Hammer-IMS makes progress in machine vision and artificial intelligence as an additional eye in quality control of textiles, plastics and other inline produced products. Recently, Hammer-IMS engineers, in collaboration with the PXL Electronics and ICT campus in Hasselt (Toon Sterckx), developed camera-assisted defect detection based on machine learning. This feature will be a key part in the newest generation of Hammer-IMS’s Edge-Vision-4.0-CURTAIN products. The feature distinguishes defects and classifies these into multiple defect categories. The robust AI-based feature for inline quality control will soon be available, delivering state-of-the-art visual inspection.
Based on a dataset of images, pre-classified by a customer, Hammer-IMS uses artificial intelligence to execute a learning process offline and builds a machine learning model for use during inline quality control. When inspecting textile or plastic inline, the newly developed software continuously traces product defects, foreign item on the product, and color-related variations resulting from e.g. ink stains, melt lines, die lines, and streaks. The machine vision solution can be tuned towards the specific product and application of each customer, covering fail categories as desired. The camera-based machine learning solution excels in virtually-zero false positive inspection results, providing maximum viability for the quality control industry. Furthermore, the system offers fast, effective and robust machine learning performance.
