Yarn tube sorting is an important function of fine winding equipment. Since the yarn tube unloaded from the winding machine may not be completely unloaded, it is necessary to detect the remaining yarn amount and sort the tubes based on the results. . At present, automatic winding machines equipped with fine coupling devices generally use mechanical detection devices to determine the remaining yarn amount of the bobbin by touching the bobbin with various forms of beaters. The disadvantage of this type of solution is that when the beater quickly falls and hits the yarn on the surface of the bobbin, it may have adverse effects on the bobbin formation and yarn quality.
In recent years, non-contact residual yarn amount detection methods have emerged, which mainly detect the color of yarn tubes through color sensors or cameras to achieve data collection. The processing methods include comparing empty tubes of the same type as preset values, or using support vector machines or The neural network divides the colors to extract yarn regions. These solutions rely more on the color difference between the bobbin and the yarn. When the two colors are similar, it is easy to misdetect, and the detection effect of the residual bobbin wrapped with a very small amount of yarn is not ideal.
In order to improve the detection accuracy, a researcher from a university proposed a non-contact detection algorithm for the remaining yarn amount of the bobbin based on Gabor filter. Gabor filters have bandpass and directionality and have been widely used in edge detection, texture segmentation, defect detection and other fields. Taking into account the color difference and surface texture difference between the bobbin tube and yarn, and using the odd part inverse symmetry of the Gabor filter, a parameter-optimized multi-directional and multi-scale filter bank is designed to filter the two-dimensional image signal and extract the wrapped yarn. The edge of the part is used to calculate the amount of remaining yarn on the bobbin. Through in-depth research on the residual yarn amount detection algorithm based on Gabor filter, the detection system separates the yarn tube image acquisition and processing. The two are performed simultaneously without interfering with each other, ensuring the real-time and stability of the system.
</p
