提出了一种机场场面监视雷达图像的机场目标特征提取方法。首先,为了解决雷达图像的亮点噪声问题,采用类形态学的图像预处理方法;然后,采用邻域链路法标记目标的轮廓像素;最后,单元面积作为权值,估计极坐标图像下机场目标的特征,并对权值的计算方法进行了讨论。实验表明,所提方法能够消除亮点噪声,得到比较准确的特征信息,为目标分类识别和多目标跟踪提供正确的特征。

A feature extraction method of airport targets in surface movement radar (SMR) image is proposed. First, aiming at solving the problem of spot noise on the SMR image, a semi-morphology means is employed; second, the neighborhood link method is used to mark the pixels of target contour; third, as the unit area being the weight, the features of airport targets are estimated in polar coordinate image, and the calculation method of weights is discussed. Experiments show that the proposed method can eliminate well the bright noise and achieve accurate feature information, providing correct features for target recognition and tracking multiple targets.

TN957