关键词: 毛皮, 染整, 配色, 麻雀搜索算法, Logistic混沌映射, 径向基函数神经网络

Abstract: Due to the special structure of animal fur, its dyeing formula mainly relies on the long-term practical experience of technicians, which leads to the problems such as low work efficiency and poor consistency of dyeing results. Based on the machine learning theory, this paper proposed a new method for quickly forming a dyeing formula. According to the characteristics of fur dyeing process engineering, this method constructed a fur coloring model based on the chaotic sparrow algorithm, and used the radial basis function neural network and color difference formula as the fitness function of the algorithm, eliminating the dependence on prior knowledge and the subjective influence of the staff in the manual coloring process. The fur samples were used for model training and testing. The experimental results show that the proposed fur coloring model could converge after 40 iterations, and the mean color difference predicted by the radial basis function neural network was 0.105, and the mean difference between the predicted results and the true values was not more than 0.001 4. The dyeing experiment was carried out by using the experimental results as the formula, and the fur samples were successfully dyed in one go, verifying the accuracy of the method.

Key words: fur, dyeing and finishing, fur color matching, sparrow search algorithm, logistic chaos, radial basis function neural network