Service Solution
About Project : Color Shade Sorting System with Artificial Intelligence
This project develops an automated color sorting system using artificial intelligence (K-means Clustering) to enhance efficiency and accuracy in sorting products by color. The system aims to reduce labor costs and production time, increase production accuracy, minimize errors, and generate insights for analysis and quality improvement.
Scope of Work
Machine Vision: Employs high-resolution cameras with optimized lighting to capture product images, develops algorithms for image analysis and color detection, and distinguishes product types (e.g., color, texture) while identifying defects like stains or flaws.
Database: Gathers data on product color variations and quality, processes information for in-depth analysis, including color consistency and defect frequency.
K-means Clustering: Leverages K-means clustering algorithms to group color data, classifying products based on color and ensuring color uniformity across batches.
Retrospective Data Verification System: This system ensures the accuracy and comprehensiveness of data used for training the algorithms by verifying previously collected data.
User-Friendly Color Addition System: This system allows users to easily add new color variations to the system without coding expertise.
Project Achievements
This project aims to develop an automatic product color shade sorting system that operates quickly and accurately, reducing the error rate in product sorting. It also aims to create a product color shade database for analyzing and improving the production process. This system will enable employees to easily access and analyze sorting results, leading to reduced labor costs and sorting time, increased production efficiency, reduced errors, improved product quality, and the generation of insights for future product analysis and development.
The computer program for checking color shades was copyrighted in 2019 under the name "Automatic Shade Inspection Program (IKS2 Shading)," and was submitted to the Department of Intellectual Property with notification number 362179.