Prediction And Application Of Laser Washing Process Parameters For Denim

With the development of industrial technology,the innovation of jeans fashion continues to increase,and people are increasingly open to the pursuit of new aesthetics and new styles.The different styles of denim clothing come from different washing processes.Due to its advantages of green environmental protection,energy conservation and efficiency,as well as personalized customization,denim laser washing technology stands out among various washing processes and has a broader market prospect.However,the current laser water washing process also has certain difficulties.Only by relying on their experience and optimizing the process parameters of laser water washing for many times can a water washer achieve satisfactory water washing results.Therefore,the existing laser water washing process has relatively low efficiency and wastes manpower and material resources.Therefore,this paper proposes to use a laser process parameter prediction model trained by convolutional neural networks to replace artificial experience in predicting approximate process parameters for denim laser washing,in order to improve the timeliness and accuracy of the denim laser washing process.The research content of this article mainly includes three aspects:(1)Using pure cotton indigo denim fabric for laser washing experiments,a network model dataset was constructed.Collect laser washing images of denim,and use the laser processing parameters that generated the images as the label values of the images to form the original dataset for this study.According to this dataset,an improved convolutional neural network prediction model is proposed based on the Le Net model,and the prediction effects of the improved model are compared with the artificial neural network model and the convolutional neural network Alex Net model.The results show that the improved model can predict approximate laser processing parameters through laser water wash images,and each processing parameter fits well,with better prediction performance than other prediction models.In addition,the predicted parameters and actual parameters are used for laser water washing experiments.The faded denim fabric was characterized by tensile testing and color K/S testing to further verify the reliability of the predicted effect.(2)To improve the prediction accuracy of the model,this paper proposes a hybrid network model based on attention mechanism.This model is constructed by combining the Residual Network(Res Net)module and the Squeeze Excitation Network(SENet)module.This model embeds the SELayer structure into the Res Net base block,thereby constructing a Res Net SENet hybrid model with higher prediction accuracy.Use this hybrid model to predict laser washed water images.Compared with existing models,the proposed hybrid model has a higher degree of fit and a smaller error between the predicted process parameters and the actual parameters.(3)An application software based on C/S architecture for predicting the parameters of laser washing process for jeans was developed.Develop based on user needs to achieve functions such as uploading water washing images to predict laser processing parameters,viewing historical prediction records,and saving prediction images and process parameter data.Software testing shows that the system meets the design requirements and can meet the functional requirements of each module.