The Research Of ACA In Flame Image Segmentation

Rotary kiln belongs to large thermal equipment with complicated structure，nonlinear，multi-interferences，large time delay and strong coupling，which is widelyapplied to sinter process to productive alumina. image processing has been applied tothe kiln temperature measuring system，which is based on image segment technology.However, it is difficult to extract flame image feature with coal-bearing dust’sinterference. Through the study of some methods used flame image segmentation, thispaper tries to present ant colony algorithm based on flame image segmentation.Theconcrete content as follows.Firstly， the ant colony algorithm is thorough studied， including the basicprinciple，mathematic model，some advantages and disadvantages. Some improved antcolony algorithms are also introduced in detail. The application of ant colonyalgorithm in TSP implemented by programming lays the foundation for thefollowing work.Secondly，some image segmentation methods are explained deeply and summedup，such as threshold method，edge detection method，method based on image regionand method based on specific theories. Experiments indicate that some commonimage segmentation methods are not too good to extract the black handle regionalfeatures.Finally，the image segmentation based on ant colony algorithm is studied. Theimage segmentation model based on ant colony algorithm is established. However，some problems still exist. For example，huge workload，time consuming，etc. So a newflame image segmentation algorithm based on polymorphic ant colony algorithm isproposed in this paper. This method makes wavelet transform to an image，splits thelowest frequency coefficient region， adjusts ant colony clustering center updatemechanism using Otsu method and K-means clustering algorithm,and then multipleant colonies form local perception to the entire perception of the image complete thetask of image segmentation cooperatively. Experimental results proves that theproposed method not only can enhance image segmentation accuracy，but also reducerunning time.