含噪背景下的目标跟踪算法与仿真分析  

Target tracking algorithm and simulation analysis in noise background

作　　者：汪剑[1] 谭宝成[1] 

机构地区：[1]西安工业大学电子信息工程学院,陕西西安710032

出　　处：《电子设计工程》2010年第3期71-73,共3页Electronic Design Engineering

摘　　要：为了提高跟踪的精确度和实时性,提出一种基于多分段阈值分割算法的快速识别方法,此方法通过对模板直方图多段阈值提取,保留当前帧图像中属于阈值分段以内的像素,使后续的算法的进行更具针对性,从仿真结果可看出,小波去噪效果很明显,并成功保护了边缘信息,目标的定位算法准确,快捷,计算速度是传统算法的1.5～2倍,满足跟踪的精确性和实时性。In order to improve the tracking accuracy and real-time tracking,this paper provides a rapid identifying method based on multi-stage thresholding algorithm.This method extracts multi-stage thresholding of template,retained current frame pixels belong to muhi-stage thresholding .Make follow-up algorithm more targeted.From result of simulation we can know that wavelet denoising is effective,in the same time it can protect edge information, targeting algorithm is accurate and fast, calculating speed is 1.5-2 times faster than traditional algorithm,it can meet real-time tracking.

关 键 词：全自主机器人 小波去噪 图像分割 轮廓提取 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]