基于蚁群遗传算法的BP神经网络摄像机标定  

BP Neural Network Camera Calibration Based on Ant Colony Genetic Algorithm

作　　者：江祥奎[1] 汪友明[1] 

机构地区：[1]西安邮电大学自动化学院,陕西西安710121

出　　处：《机械与电子》2013年第12期60-62,共3页Machinery & Electronics

基　　金：西安邮电大学青年教师科研基金资助项目(ZL2012-14)

摘　　要：摄像机标定是从二维图像提取三维空间信息的关键步骤。为了有效解决传统摄像机标定算法中的多参数和计算费时费力等问题,提高摄像机标定的精度和速度,首次将蚁群遗传算法应用于摄像机标定中。方法初期采用遗传算法过程生成信息素分布,后期利用蚁群算法正反馈求精确解,最后用优化后的BP神经网络来进行摄像机标定,充分发挥遗传算法的全局搜索能力和蚁群算法的正反馈收敛优势。Camera calibration is the key procedure for extracting three dimensional information from two dimensional image. In order to solve the problem of multi- parameters and calculation wasting time and energy, and promote the accuracy and speed of camera calibration, the paper firstly applies ant colony genetic algorithm to camera cali- bration. This method uses the genetic algorithm to process the early generation information grain dis- tribution,later using positive feedback of ant colo- ny algorithm for the exact solution, finally using the optimized BP neural network for camera cali- bration, which give full play to the global search ability of genetic algorithm and the positive feed- back convergence advantage of ant colony algorithm.

关 键 词：摄像机标定 神经网络 蚁群遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]