网络图像集合多目标元素信息跟踪识别仿真  

Multi Target Element Information Tracking and Recognition Simulation of Network Image Set

作　　者：何颖 HE Ying(Experimental Teaching Department,Guizhou University of Finance and Economics,Guiyang Guizhou 550025,China)

机构地区：[1]贵州财经大学实验教学部

出　　处：《计算机仿真》2019年第6期350-353,372,共5页Computer Simulation

摘　　要：对网络图像多目标元素信息的跟踪识别,能有效提高网络图像传输质量。网络图像集合多目标元素信息跟踪识别,需要将测得目标元素位置进行修正,再判别目标元素图像轮廓形状间距误差,完成网络图像集合多目标元素信息跟踪识别。传统方法用形态学处理方法对多目标元素初始化曲线中噪声干扰进行滤波,再对多目标元素轮廓特征进行估计,但忽略了对误跟踪识别进行判定导致跟踪识别误差较大。提出基于能量和轮廓形状间距的多目标元素跟踪识别方法,将多目标元素显著性特征作为粒子滤波状态向量。检测出视觉显著图多个目标元素,将粒子群滤波得到的最优估值与测得目标元素位置进行关联与修正,并用网络图像集合中目标元素图像能量和轮廓形状间距对导致误跟踪情况进行判别,实现多目标元素信息跟踪识别。结果证明,所提方法能有效提高多目标元素信息轨迹跟踪观测值与真实值的拟合度。The tracking and recognition for multi-objective element information in network image can effectively improve the quality of network image transmission.The traditional method uses the morphological method to filter the noise interference in the initialization curve of multi-target element,but ignores the judgment of error tracking recognition,leading to the large error of tracking recognition.Therefore,a method to track and recognize multi-objective element based on the spacing between energy and contour shape was presented.At first,this research used the salient characteristic of multi-objective element as the particle filter state vector,and then detected multiple target elements from saliency map.Moreover,the research associated and modified optimal estimation obtained by particle swarm filtering and the location of target element.Meanwhile,our research used the spacing between energy and contour shape of objective element image in network image set to distinguish the false tracking.Thus,the tracking and recognition of multi-objective element information was achieved.The results prove that the proposed method can effectively improve the fitting degree between the tracking value and the real value of multi-target element information trajectory tracking.

关 键 词：网络图像集合 多目标元素 信息跟踪 识别方法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]