Aiming at the problems of insufficient feature extraction and the dependence of model parameters on manual experience in the traditional method of identifying boiler combustion state by using flame images,this paper proposes a method of identifying boiler flame combustion state based on multi-feature fusion and whale algorithm optimized support vector machine.In order to characterize the flame image from different aspects,the method firstly extracts the contour features,texture features and local features of the boiler flame image using Histogram of Orientation Gradient,Local Binary Pattern and Scale Invariant Feature Transform,respectively.Then,the three extracted features are fused to achieve complementarity between different features and to improve the feature representation of the flame image.Finally,the fusion features are used as the input sample of the support vector machine classifier for flame combustion status recognition,and the whale algorithm is used to optimize the parameters in the support vector machine to improve the accuracy of the classification model.The experimental results show that the classification accuracy of multi-feature fusion is significantly improved compared with the classification results of single feature,and the final classification accuracy of the support vector machine model optimized by the whale algorithm is as high as 96.64％.
Flame images Boilercombustion state recognitionMulti-feature fusionWhale algorithmSupport vector machine
NETL
NSTL
万方数据
