改进的特征匹配雷达视频运动目标跟踪算法

Improved Feature Matching Moving Target Tracking Algorithm on Radar Video

针对密集杂波和复杂多目标情况下,特征匹配跟踪算法定位跟踪精度较低的问题,提出一种改进的特征匹配雷达视频运动目标跟踪算法。首先,从单帧雷达图像检测结果中提取目标的面积、位置、不变矩信息,在运动目标得到判定后,利用卡尔曼滤波的预测位置以及上一帧特征信息进行目标匹配;然后,根据匹配位置和预测位置给出目标估计位置,并进行特征信息更新;最后,输出目标的特征信息、位置信息和运动参数。利用某型导航雷达上采集的实测数据验证了算法的有效性。

In order to solve the problem of low location tracking accuracy of feature matching tracking algorithm in dense clutter and complex multi-objective situation, an improved feature matching moving target tracking algorithm on radar video was put forward. First, target area, location and invariant moments were extracted from single frame radar image detection results, after the moving target was determined, target was matched utilizing target location predicted by Kalman filter and the prior frame target feature information. Secondly, the estimated target location was gained according to matching location and forecast location, and feature information was updated. Finally, target feature information, target location information and motion parameters were output. The effectiveness of the algorithm was verified by the collected marine radar data.