Abstract:

The evenness of raw silk is one of its most valuable qualities. Seriplane inspection is a common test for raw silk evenness but is prone to human error and lacks accurate repeatability; the Uster evenness tester is another common approach but is costly and vulnerable to environmental factors. An image-based raw silk evenness detection system is proposed in this paper. The system is comprised of an image acquisition segment with a CCD image sensor, telecentric lens, light source, over feeding device, and raw silk winding device, plus an image processing segment tasked with threshold segmentation and morphology operations. Images of the raw silk are first captured with the image acquisition segment, then the images are processed by threshold segmentation and morphology processing; the diameters obtained in this segment are then used to calculate the variation coefficient (CV), which characterises the evenness of the raw silk. We conducted three experiments to test the stability, repeatability, and accuracy of the system. The results showed that the system proposed is stable, repeatable, and accurate.