Infrared Image Segmentation Method Based On Ant Colony Algorithm

In recent years, the field of infrared imaging applications more widely. However, the current research of image segmentation and identification, has been unable to meet the needs of practical engineering applications. With the rise of bionics, the infrared image segmentation have a new development direction. Based on this, the paper studies on the combine of genetic algorithm and colony algorithm used in infrared image processing. The paper main contents include:1) According to the characteristics of infrared image processing, genetic algorithm is improved. So that the image segmentation effect is more obvious and speed up the processing speed of the algorithm. First of all, quick sort order is used to instead of the sort, and then introduced hereditary algebra into the crossover and mutation probability. Secondly,two-dimensional OTSU algorithm is used as fitness function to adjust neighborhood gray value. According to the distance between of neighborhood pixel and center pixel's different,proportionality coefficients is introduced to adjust weighting factor. Again, noise reduction is determined by the thermal infrared image noise points. Finally, the introduction of intra-class variance enhance the accuracy of the algorithm. Improved genetic algorithm reduces the genetic algorithm prematurity, while improving arithmetic operation and post-convergence.2) In order to better meet the needs of the infrared image processing, to further enhance the adaptive capacity of ant colony algorithm. On the basis of the original ant colony algorithm, the algorithm is modified as follows. First, the pheromone factor, inspired prime factor and the evaporation coefficient adaptive adjustment based on changes in the number of iterations; Secondly, the way to update pheromone changes are on the optimal path and the worst path pheromone update settings, improving the performance of the algorithm. New algorithm further enhance the ability to pre-optimization and post-convergence.3)In the genetic algorithm and ant colony algorithm combining process, the Second-best Solutions of genetic operation is converted into a weighting factor, and acts on the state transition probability, increasing the ability of optimization algorithms. Genetic Algorithm-Ant Colony Optimization Algorithm full use strong global search capability ofGenetic algorithm and Positive feedback fast convergence capability of ant colony algorithm,to meet the infrared image processing speed and accuracy.The experimental result shows that, the combination algorithm can achieve the effect of learn from each other. Meanwhile, the fusion algorithm improved the efficiency and speed in segmentation thermal infrared images. The research results have great practical engineering application.