CAJ | 학술논문

提出了一种将改进的遗传算法和Levenberg Marquardt(LM)算法相混合优化的摄像机自标定方法.首先将Hartley定义的简化Kruppa方程转换为优化代价函数,然后利用改进的遗传算法和LM算法相混合的优化算法求优化代价函数的最小值,进而求得摄像机的内参数.实验结果表明,与单一的优化方法相比,该方法的标定精度得到了较大的提高.
제출료일충장개진적유전산법화Levenberg Marquardt(LM)산법상혼합우화적섭상궤자표정방법.수선장Hartley정의적간화Kruppa방정전환위우화대개함수,연후이용개진적유전산법화LM산법상혼합적우화산법구우화대개함수적최소치,진이구득섭상궤적내삼수.실험결과표명,여단일적우화방법상비,해방법적표정정도득도료교대적제고.
This paper presented a new camera self-calibration method by combining an improved genetic algorithm (GA) with Levenberg-Marquardt (LM) algorithm. Firstly, translated the simplified Kruppa equation defined by Hartley into the optimized cost function. Secondly, calculated the minimum value of the optimized cost function by a hybrid optimization algorithm which combined the improved GA with LM algorithm. Finally, obtained the intrinsic parameters of the camera. The experimental results show that the accuracy of the proposed method is obviously improved compared with single optimization methods.