支持向量机的海上移动目标位置跟踪  

Support vector machine for tracking the position of moving targets at sea

作　　者：彭晓雷 江田辉 吕月惠 PENG Xiaolei;JIANG Tianhui;LU Yuehui(Xiamen Ocean Vocational College,Xiamen 361100,China)

机构地区：[1]厦门海洋职业技术学院,福建厦门361100

出　　处：《舰船科学技术》2024年第22期178-181,共4页Ship Science and Technology

摘　　要：为获取海上移动目标的位置信息,精准掌握其移动轨迹,研究支持向量机的海上移动目标位置跟踪方法。采集海上移动目标的连续帧图像,构建2个样本数据集组成训练集,从训练集中提取海上移动目标图像的边缘特征和局部灰度最大值特征,构建特征向量集训练排序支持向量机,获取学习排序函数,输入海上移动目标的图像帧及原始位置,通过排序支持向量机计算排序函数分值,确定目标位置,并不断重复该过程以实现连续的位置跟踪。实验结果显示,所提方法可以有效提取用于表征海上移动目标的边缘特征以及局部灰度最大值特征,实现移动目标位置的准确跟踪,并生成其位置移动轨迹。To obtain the position information of moving targets at sea and accurately grasp their movement trajectory,a method for tracking the position of moving targets at sea using support vector machines is studied.Collect continuous frame images of moving targets at sea,construct two sample datasets to form a training set,extract edge features and local grayscale maximum features of moving target images from the training set,construct a feature vector set to train a sorting support vector machine,obtain a learning sorting function,input the image frames and original positions of moving targets at sea,calculate the sorting function score through the sorting support vector machine,determine the target position,and repeat the process continuously to achieve continuous position tracking.The experimental results show that the proposed method can effectively extract edge features and local grayscale maximum features used to characterize moving targets at sea,achieve accurate tracking of the position of the moving target,and generate its position movement trajectory.

关 键 词：支持向量机 海上移动目标 位置跟踪 特征提取 排序函数 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]