The Algorithm Research Of Color-difference Detection For Dyeing Products Based On Computer Vision

In modern textile industry,textile and dyeing enterprises must strictly control the important color quality indicator which can't be ignored to improve the market competitiveness of the finished piece goods after dyeing.Traditional the color difference classification and illumination correction mainly relies on the artificial to complete.The strong subjectivity and low efficiency exist in the detection process.So it is important and necessary to consider the factor.It is a good solution to introduce the computer vision into the intelligent evaluation of color difference between printing and dyeing products to replace human visual inspection.In the color difference detection system of printing and dyeing products,we need to solve the key technical problems such as the acquisition of color difference image,the establishment of color difference evaluation model and the illumination correction of images.The research content of this paper is mainly mainly focuses on the color difference detection and illumination estimation of the printing and dyeing products based on computer vision,and this paper is committed to building have good stability of color classification model and illumination correction model.The main work and research results of this paper are summarized as follows:(1)The basic concepts and research status of color difference classification and illumination estimation of printing and dyeing products are briefly introduced.The characteristics,advantages and disadvantages of various color difference classification and illumination estimation algorithms are analyzed and compared.Aiming at the color difference classification of textiles,the support vector machine algorithm with good classification performance is mainly studied.The theoretical basis and existing problems of support vector machine are analyzed and discussed,which provides a solid theoretical foundation for the follow-up work in this paper.(2)Different illuminations adversely affect color difference evaluation of textile images in dyed fabrics.To address the problem,we propose a Rotation Forest-based ensemble particle swarm optimization and sparse least squares support vector regression(RF-PSO-SLSSVR)for building an accurate illumination correction model.In our algorithm,since the standard LSSVR cannot yield a sparse solution,we develop sparse LSSVR(SLSSVR)by calc ulating the maximal independent subset in the extracted feature space.Then,SLSSVR is embedded into Rotation Forest(RF)by substituting for the regression tree which is the base learner in the original RF,and the PSO technique is employed to obtain the optimal regularization parameter ? and kernel parameter ?.The final model is obtained by fusing the predictions of the different trees through a weighted average method and RF-PSO-SLSSVR is constructed to learn the textile illumination estimation model.To verify the effectiveness of our algorithm,we carry out the experiments on the real dyed fabric images by comparison to several related methods and the performance is measured by the different criterions,including the chromaticity error,the angle error and the Wilcoxon signed-rank test.The results show that the proposed illumination correction method can achieve a higher accuracy and be more robust than SVR and ELM.(3)In order to solve the problem that the traditional illumination correction algorithm is not accurate and the training time is slow,this paper proposes an ensemble DE-OSELM textile illumination correction model based on the Rotation Forest framework.Firstly,the Grey-Edge framework is used to extract the low dimensional and efficient image features as online sequential extreme learning machine(OSELM)input vectors to improve the training and learning speed of OSELM.Because the input weight and hidden layer bias of OSELM are randomly obtained,the OSELM algorithm has poor prediction acc uracy,and low robustness.In order to overcome this shortcoming,differential evolution(DE)algorithm that has the advantages of good global search ability and robustness is used to optimize the input weight and hidden layer bias of OSELM,namely DE-OSELM.In order to further improve the generalization ability and robustness of the illumination correction model,the Rotating Forest(RF)algorithm is used as the ensemble framework,and the DE-OSELM is used as the base learner to replace the regression tree algorithm in the original Rotation Forest.Then,the obtained multiple different DE-OSELM learners are aggregated to establish the prediction model.The experimental results show that compared with the textile color correction algorithm based on SVR and ELM algorithm,the ensemble illuminat ion correction method achieves high prediction accuracy,strong robustness and good generalization ability.(4)In order to establish the color difference classification model of printing and dyeing products,a grey Wolf algorithm optimization support vector machine based on differential evolution model is proposed in this paper.First of all,the performance of the SVM model is mainly affected by the penalty parameter C and RBF gamma two kernel width parameters,using the Grey Wolf optimization algorithm good global search capability for the best combination of parameters for the iterative optimization of support vector machine;at the same time,because of the initial population Wolf algorithm has great influence on the quality and speed of algorithm using difference evolutionary algorithm for the initial group Grey Wolf algorithm to generate more appropriate,the ability of solving some populations better.Finally,through the optimization of the penalty factor and the kernel width parameter,a support vector machine classification model with strong generalization ability is constructed.The experimental results show that compared with the Grey Wolf optimization algorithm,the color difference classification method of the original support vector machine is optimized,and the proposed method achieves high classification accuracy,and has good stability and generalization ability.