Yarn breakage location for yarn joining robot based on machine vision

In order to identify and locate the yarn breakage in the spinning process for the yarn joining robot through visual method and to simplify the mechanical structure, a recognition and positioning algorithm for yarn characteristics is proposed according to the image characteristics. An industrial camera was used to collect the image of the yarn being sucked into the suction nozzle, and the contrast between yarn features and background was enhanced through an improved gray enhancement method, using Canny operator for yarn edge detection. The image features of the yarn were obtained by dividing the interest regions and optimized using Hough line detection method, and the positioning algorithm was used to extract the required location information. The experimental results show that the position information extracted by the proposed algorithm has high accuracy, the error of coordinate points is 1.42 pixels, and the error of angle α is 0.60°. Compared with the use of the traditional location detecting algorithm, the running time of the program is reduced, and the average recognition time is in the order of 10-1 s, with good real-time performance. The research results can be applied to the development of yarn joining robot products. © 2022, Periodical Agency of Journal of Textile Research. All right reserved.