Industrial furnace is a kind of high temperature equipment that uses the heat of fuel combustion to process materials.If it is not used properly,it is very easy to explode.By observing the flame image of the industrial furnace,the combustion of fuel can be accurately un-derstood to avoid accidents.But the flame image of industrial furnace is easily affected by combustion conditions and other factors,it is diffi-cult to select the segmentation threshold and cannot accurately segment the flame image.Therefore,a threshold segmentation algorithm of in-dustrial furnace flame image based on improved ISCA algorithm is proposed.The ISCA algorithm is improved by means of chaotic mapping and normalization,and the opposite learning mechanism is set up according to greedy selection to increase population diversity and effectively avoid falling into local optimal solution.Based on the median filtering algorithm,the noise of the flame image is removed,and the threshold segmen-tation function model of the flame image is established by using threshold segmentation and gray histogram.The model is solved by combining Kapur entropy maximization and cumulative distribution function to complete the threshold segmentation of the industrial furnace flame image.The experimental results show that the error rate of the proposed algorithm is always lower than 2％,the threshold segmentation time is less than 0.3 s,and the convergence can be completed by the 17th iteration.The algorithm has fast operation speed and low error rate of threshold segmentation in the process of industrial furnace flame image threshold segmentation,which can effectively improve the accuracy of industrial furnace flame image threshold segmentation.
关键词
工业炉火焰图像/改进ISCA算法/粒子空间位置/混沌映射/对立学习机制/最优阈值/阈值分割
Key words
flame image of industrial furnace/improved ISCA algorithm/particle space position/chaotic mapping/opposite learning mecha-nism/optimal threshold/threshold segmentation
维普
万方数据