船舶目标检测与跟踪技术的特征信息提取研究

Research on feature information extraction of ship target detection and tracking technology

为提升船舶目标检测和跟踪的技术水平,研究船舶目标检测与跟踪技术的特征信息提取方法。首先利用高斯混合背景模型得到船舶前景图像,然后分别提取船舶前景图像的亮度、颜色、方向特征和船舶轮廓特征信息,最后利用支持向量机和Chan-Vese模型利用特征实现船舶目标检测与跟踪。实验结果表明,该方法提取船舶前景图像时可有效提取其微小细节,提取的船舶图像亮度与颜色与原始值吻合度极高,可有效提取船舶轮廓特征,船舶类型识别和跟踪误差小,应用效果较为显著。

In order to improve the technical level of ship target detection and tracking,the feature information extraction method of ship target detection and tracking technology is studied.The ship foreground image is obtained by using Gaussian mixture background model.the brightness,color,direction features and ship contour feature information of the ship foreground image are extracted respectively.The ship target detection and tracking are realized by using support vector machine and Chan Vese model.The experimental results show that this method can effectively extract the small details of the ship foreground image,the brightness and color of the extracted ship image are highly consistent with the original values,and the ship contour features can be effectively extracted.The error of ship type recognition and tracking is small,and the application effect is relatively significant.