中文摘要：

为了更加快速准确地检测出混纺纱线的混纺比，采用神经网络方法对混纺比检测进行了尝试性研究．针对基本BP网络算法的局部最小和不收敛问题，通过对学习速率进行优化调整以改进BP算法，并将其应用于纱线混纺比的检测．仿真结果表明：本文方法对于局部最小值和可能的不收敛问题有很大改进，可以为预测纱线的混纺比提供一种更加精确快速的方法．

英文摘要：

In order to detect the yarn blending ratio more rapidly and accurately,the method of neural network is used toconduct a research for blending ratio test. Because of the local minimum and the non-convergent problems of basicBP algorithm,through the optimization of learning rate,a new BP algorithm is used for different yarn test. Thesimulation results show that the method can improve the convergence and local minimum of BP algorithm,so thata more accurate and quick way of predicting the blending ratio is provided.