满足滞留度指标的图像波门预测技术研究  

Prediction Study of Tracking Window with Specified Residence Index

作　　者：程相权[1] 王远钢[1] 郭治[1] 

机构地区：[1]南京理工大学自动化系,南京210094

出　　处：《南京理工大学学报》2001年第4期346-349,共4页Journal of Nanjing University of Science and Technology

摘　　要：将衡量目标图像跟踪好坏的滞留度指标转换成预测误差稳态方差指标 ,利用满意估计思想设计估计增益 ,对图像波门中心进行预测 ,使得稳态预测误差方差满足给定误差方差指标、预测误差系统的极点配置到给定的圆内 ,确保目标图像跟踪过程中 ,目标穿越波门的滞留度满足预定要求 ,并保证预测过程具有满意动态特性。利用线性矩阵不等式 (LMI)方法 ,将满意估计增益的求解问题转化成LMI组的可行解问题 ,后者可借助MatlabResidence requirement valuating the quality of image tracking is characterized by the steady predictive error variance index. Based on theory of satisfactory estimation, the center of the tracking window is predicted so that the steady predictive error variance might meet the given variance index and poles of predictive error system assigns in a given circular region. This guarantees the residence index of target passing tracking window to satisfy the given requirement and the predictive process to have satisfactory dynamic quality. Via linear matrix inequality (LMI) approach, the problem of finding satisfactory prediction is changed into a feasible problem of some certain LMIs, and the latter may be solved with effect by means of computing software, Matlab LMI toolbox.

关 键 词：目标跟踪 预测技术 LMI方法 图像处理 滞留度 图像波门中心 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]