Answers are grounded in this paper and cite their sources.
In response to the issues of poor signal stability and low defect feature extraction accuracy in yarn defect detection within the textile industry, this study proposes a dynamic entropy signal processing method based on piezoelectric sensing to efficiently and accurately extract yarn defect features. By designing a piezoelectric yarn defect detection platform, the yarn vibrations are converted into high-sensitivity electrical signals, enabling high-frequency signal acquisition. A dynamic entropy feature extraction method is introduced, which uses a sliding window and dynamic binning probability distribution model to capture local features of the yarn signal. Experimental results demonstrate that this method achieves a defect recognition accuracy of over 95% for cotton knots, with an average processing time only half of that of the comparison method. A nonlinear regression model is established to map the entropy value to the cotton knot severity, with the nonlinear asymptotic regression model achieving a fitting degree of 0.996. Finally, the influence of key dynamic entropy parameters (window width, bin width) and sensor sensitivity on detection performance is discussed. This study provides an effective solution to improve the accuracy and efficiency of yarn defect detection, with potential applications in online quality control.
No takes yet. Share an insight, caveat, or question.
Qi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: