基于特征角点的目标跟踪和快速识别算法研究

提出了一种基于特征角点的目标跟踪、识别方法,其运算效率较高,且角点不易丢失。从对基于灰度的角点提取方法和基于边缘的角点提取方法的比较入手,提出建立新特征模型的必要性。随后给出了一种既能提高运算效率又能简化跟踪模型的特征角点法。选取了飞行速度为300m/s的某战机序列共11帧连续图像作为处理对象,通过在主要配置为Pentium 4、80G内存计算机的、Matlab2006a软件的环境中进行仿真,算法的运算速度可达0.7s,与其他跟踪算法相比跟踪速度较快,表明该方法是一种简洁有效的目标跟踪识别方法。

Fast Target Recognition and Tracking Method Based on Characteristic Corner

A fast target recognition and tracking method based on characteristic corner is introduced,which is efficient in target tracking and holds corners very well during tracking.The contrast of the cornerpick-up methods based on gray histogram and based on edge indicates the necessity of a new characteristic model,and a simple tracking model based on characteristic corners is advanced,with improved processing efficiency and simplified tracking target.Throughout the simulation on 11 frames of the video sequence of a battleplane,whose speed is 300 m/s in the environment of computer with Pentium 4,80 G random access memory,software Matlab2006a,the processing speed reaches 0.7 s,and the result indicates that this method is effective and simple compared with other tracking methods.