Research On The Accurate Identification Of Tobacco Leaf Maturity And Quality Based On SSA-SVM

Accurate tobacco maturity and quality identification is an important tool to improve cigarette quality inspection and grading.However,existing deep learning-based tobacco maturity and quality recognition methods require large amounts of training data and powerful computational resources,which are not suitable for deployment in resourceconstrained embedded devices.In addition,traditional machine learning-based tobacco maturity and quality recognition methods suffer from low recognition accuracy when the number of features is large and the sample size is small.To address the above problems,this paper designs a lightweight method for accurate tobacco maturity and quality recognition using sparrow search algorithm and support vector machine,which mainly includes:(1)To address the problem of low maturity recognition accuracy caused by the large number of features and small sample size,this paper proposes a fresh tobacco leaf maturity recognition method combining principal component analysis,sparrow search algorithm and support vector machine.Firstly,the original fresh tobacco images are pre-processed to construct the feature matrix of color and texture information;then the principal component analysis is used to remove the redundant information,and the reduced-dimensional feature matrix is used as the input of the support vector machine;finally,the sparrow search algorithm is used to optimize the parameters of the support vector machine to complete the training of the model.The experimental results show that the recognition rate of the proposed method is better than other methods,and its recognition rate in the test set is 93.55%.(2)To address the problem of low recognition accuracy of roasted tobacco quality due to the large number of features and small sample size,this paper proposes a roasted tobacco quality recognition method combining partial least squares,sparrow search algorithm and support vector machine.First,the color feature matrix is constructed by using the reflection map and perspective view of roasted tobacco;then,the best principal component is selected as the input of the support vector machine by using partial least squares;finally,the parameters of the support vector machine are optimized by using the sparrow search algorithm to complete the training of the model.The experimental results show that the prediction performance of the proposed method is better than other methods,with a coefficient of determination of 0.8656 and a root mean square error of 1.6301 in the test set.