基于改进的差异演化算法的多视角离散数据配准

Matching unorganized points data under different viewpoints based on improved evolution algorithm

针对初始位置差异比较大的不同视角的离散数据的配准,提出了基于差异演化算法和ICP的数据配准方法。首先通过改进的差异演化算法进行点云数据的粗配准,然后利用ICP算法进行精确配准。在演化过程中,通过四元素法减少解空间个体参数个数,对选择操作进行修改;同时采用自适应的交叉和变异概率,以避免过早收敛,提高差异演化的寻优速度。通过实例验证算法有较好的配准效果和运行速度。

In order to align partly overlapped data clouds measured from different viewpoints and with greate difference in initial position,this paper proposed a detecting method based on differential evolution and ICP algorithm. Firstly,roughly registrated data clouds with differential evolution algorithm method and then employed ICP algorithm method in the accuracy registration. In differential evolution,used quaternion method to decrease the individual numbers of revolution space and accordingly adapted the selection operation,to avoid premature convergence and improve optimizing speed,adaptively adjusted the probabilities of crossover and mutation by means of adaptive algorithm. Some examples prove the method is effective and efficient for aligning large number of three dimension clouds data.