摘要
本文提出了由于噪声影响实时电视跟踪系统的目标位置点的误差分析算法。从系统的跟踪算法中导出目标跟踪系统误差分析模型。 本文导出目标质心位置的统计模型。第一种模型,即目标跟踪误差简化模型,通过假设目标的误分类概率等于背景的误分类概率推导出。第二种模型,即目标跟踪误差扩展模型,通过假设目标的误分类概率不等于背景的误分类概率推导出来。本文描述这些模型及其数学含义。
This paper presents the target point error anslysis algorithm of the real-time TV tracking system affected by noise. The target tracking system errors analysis models are derived from the tracker algorithm.Two statistic models of the target centroid tracking precision are derived. The first model, a simplified target tracking error model, is derived from the assumption that the probability of incorrect classification of the target pixel is equal to that of the background pixel. The second model, an extended target tracking error model, is derived from the assumption that the probability of incorrect classification of the target pixel differs from that of the background pixel. These models are described and their mathematical implications are discussed.
出处
《系统工程与电子技术》
EI
CSCD
1990年第3期51-56,共6页
Systems Engineering and Electronics
