Main Article Content
This paper presents an algorithm with improved genetic algorithm design template in cellular neural networks parameters and realizes the human body edge detection in infrared image. The algorithm uses population Cross generational elitist selection and the subgroups of parallel population that overcome the shortcomings of the simple genetic algorithm in solving for the optimal template in cellular neural networks, which is premature convergence. The improved genetic algorithm can converge quickly to the stable and optimal value. The simulation results show that this algorithm is more effective than traditional algorithms based on Particle Swarm Optimization and on a simple genetic algorithm. At the same time, compared with the traditional Canny algorithm, this algorithm can detect the human body more clearly and accurately in infrared images, with low miss-detection. This greatly improves the processing speed of the subsequent target tracking and provides a new method for body edge detection in an infrared image.