Color Analysis and Classification Based on Machine Learning Technique Using RGB Camera Industrial Practice and Experience Paper

Automation plays an important role in production industry for improving the quality and throughput of the end product. In most of the textile industries, manual inspection of the manufactured product is done for identification of quality parameters such as dimensions and features. Manual assessment is costly, time-consuming and sometimes inaccurate for color analysis and matching. As a solution to this problem, we can use image processing techniques for color analysis and matching by using Microsoft Visual Studio, OpenCV, Python and C#. The GigE camera is used for acquiring the image of the fabric and analyzed for color matching using the above software platforms. The objective of this project is to perform color matching in fabrics for quality measures and to increase the efficiency of the process. Euclidian distance method and k-means color clustering method are performed for getting the required output and increasing throughput. It is found after experimentation that k-means color clustering method gives better results as compared to the normal Euclidian distance method. This is due to machine learning involved in k-means color clustering method, which helps in the identification of different color clusters in a fabric piece and setting standard values for color match of a fabric, thereby increasing the chance of getting a perfect color match.