Improved Quantum Firefly Algorithm And Its Application In Image Threshold Segmentation

Accurate image segmentation plays a key role in image processing.However,due to the influence of many factors such as geometric distortion,image noise,color texture change and so on,the traditional segmentation methods present a large amount of calculation,and the segmentation characteristic of low accuracy,and they are difficult to improve the segmentation precision of image processing.A few years ago,the minimum cross entropy is widely used in multilevel threshold extraction,therefore,many kinds of intelligent optimization algorithms have been proposed to find the optimal combination of the different thresholds to solve this problem effectively,within a certain amount of time,which makes the cross entropy minimum.Because the performance of firefly algorithm is superior,it has few parameters and calculation,in addition,quantum computing has the characteristics of quantum parallel,and it can make the classical algorithm to speed up the speed.Basic quantum firefly algorithm?QFA?is easy to fall into local optimum,and have the defect of premature,so this paper puts forward the quantum firefly algorithm based on three kinds of combination strategy,the minimum cross entropy as the objective function,and applies it in the helan mountain rock paintings and classical image threshold segmentation issue,the better results were obtained.The relevant work is as follows:?1?the firefly algorithm,quantum computing and image segmentation theory are summarized,and the firefly algorithm is applied to a typical multi-peak function for performance test,and the advantages and disadvantages of the firefly algorithm are analyzed.?2?three strategies for improving the QFA are studied,which are adaptive step adjustment strategy,elite reverse learning strategies and border control strategy,respectively.The three strategies in composite are applied to the basic quantum firefly algorithm,three kinds of improved QFA are named after AQFA,AQFA₁,AQFA₂,respectively.AQFA₂ is adaptive quantum firefly algorithm.And we test the improved algorithm in 10 benchmark test functions,under certain conditions the results show that the adaptive quantum firefly algorithm has better optimization performance,and has strong advantage in solving multi-dimensional function,and can effectively improve the calculation speed of high dimensional function and save computing time.?3?improved QFA was exploited to classic image and the helan mountain rock art image segmentation research.In this paper,taking the minimum cross entropy as the fitness function,the results show that the improved QFA,quantum adaptive firefly algorithm,has improved computing speed for threshold segmentation algorithm and shorten the time of segmentation.Among other things,in either MS SIM or PSNR,we got significant improvement.Therefore,AQFA₂ algorithm proposed in this paper provides a novel and efficient method for image segmentation.