CAJ | 학술논문

为提高遗传算法全局寻优的性能,将标准遗传算法改进并运用到摄像机标定,提出一种新的双目摄像机标定方法。通过改进编码方法,自适应调整变量搜索区间;引入微分进化进行交叉和变异,使群体呈多元化发展。该方法与现有的非线性标定方法相比,标定精度更高,可有效解决传统优化方法难以解决的多参数、非线性的复杂函数优化问题。实验结果表明,该方法具有较强的泛化能力,鲁棒性好。
위제고유전산법전국심우적성능,장표준유전산법개진병운용도섭상궤표정,제출일충신적쌍목섭상궤표정방법。통과개진편마방법,자괄응조정변량수색구간;인입미분진화진행교차화변이,사군체정다원화발전。해방법여현유적비선성표정방법상비,표정정도경고,가유효해결전통우화방법난이해결적다삼수、비선성적복잡함수우화문제。실험결과표명,해방법구유교강적범화능력,로봉성호。
In order to improve the global optimization performance of genetic algorithm,the standard genetic algorithm is improved and used to camera calibration,so that a new kind of binocular camera calibration method is proposed.By improving the encoding methods,the variable searching interval can be adjusted adaptively.By introducing the differential evolution into crossover and mutation,the group can be led into diversified development.Compared with some existing non-linear calibration methods,the proposed one holds a higher accuracy calibration.It＇s an effective solution to the multi-parameter and nonlinear complex function optimization problems that can not be solved by the traditional optimization methods.Experimental results show that,the proposed method is of better generalization ability and good robustness.