Study On Ant Colony Algorithm And Application Of Image Recognition Based On ACA

Now in area of studying on intelligent automatic technique the algorithms which possessed colony and intelligence have come in for more and more attention. As a representative realization of colony intelligence, ant colony algorithm has received extensive attention from academe. It was proposed first by Italy scholar M.Dorigo in 1991. It is another heuristic search algorithm applied in combinational optimized problem followed by simulated annealing algorithm, heredity algorithm, taboo search algorithm, ANN algorithm and so on. Ant colony algorithm not only intelligent search, intelligent optimize but also has many characters such as steadiness, positive feedback, parallel calculation, easy combined with other algorithms and so forth. So it has been one of the most attractive questions for discussion. The work of this task include follow three aspects: Firstly, researching on ant colony algorithm, summarize the studying headway. Sum up the applying area of ant colony algorithm and the disadvantage, and deeply analyzing the disadvantage to improve total capability of ant colony algorithm. Secondly, study and analyze functions of texture image, combine with ant colony, and propose statistical functions which can reflect the image more generally as gist of image recognition. Thirdly, according to clustering and discreteness of ant colony algorithm, it is proposed effective image recognizing clustering ant colony algorithm to extract feature functions from texture image and coin image. Finally image recognition is realized. Experimental results show that the algorithm proposed in this thesis can improve the recognizing rate of image effectively, and the procedure is simple. Now image processing methods based on ant colony algorithm which published in nation and abroad are very few, so ant colony algorithm is applied in image recognition in this thesis, which is effective attempt.