基于SUSAN和改进的Snake模型的空中目标图像分割

The Image Segmentation of the Aerial Target Based on the SUSAN and Improved Snake Model

文中提出一种基于SUSAN和改进的Snake模型相结合的算法,用以进行空中目标在图像中的检测和分割。首先使用SUSAN算子检测飞行器的角点,通过连接已检测出的角点形成Snake模型初始轮廓,然后用改进的Snake算子进行目标轮廓拟合。实验结果表明,该方法可以有效地进行空中目标的检测和分割,并且克服传统Snake模型易漂移和过收敛等缺点。

In this paper,an algorithm is proposed based on SUSAN operator and improved Snake model.The algorithm is used for aerial target detection and segmentation.Firstly.the corner points of the aerial target are detected according to SUSAN operator.Initial contour is built by connecting the detected corner points.Secondly,the aerial target contour is fitted using Snake operator.The results of experiments show that this method is effective for target detection and segmentation and overcomes the drawbacks of traditional Snake model,such as excursion,local optimum,et al.