基于模糊聚类的光电经纬仪多子弹弹道测量

Opto-electronic theodolite multiple ballistic test based on fuzzy c-means clustering

本文对密集多子弹在空间坐标系中的左右相机弹点分布识别提出了一种新的方法。在各帧拍摄所得多子弹成像点中,提取子弹的形心坐标,根据子弹运动特点建立状态方程和测量方程,利用Kalman算法进行滤波。利用模糊c-均值聚类算法确定测量与弹丸轨迹的关联程度,从而实现弹道识别。仿真实验结果说明,该方法能有效地进行多子弹的弹道识别跟踪,并且算法简单、计算量小、易于工程实现。

A fuzzy logic data correlation approach for multiple bullets tracking was proposed. The approach was constructed based on fuzzy clustering means algorithm. The state equation and measurement equation were established according to the character of bullets in each frame image data. The membership degree of each measurement was calculated by fuzzy c-mean （FCM） algorithm. The right assignment is made according to the maximum degree of membership. Computer simulations indicate that this scheme achieve multiple bullets tracking perfectly with low computational complexity, high precision and easy realization.