一种受人脑三阶段记忆机制启发的鲁棒运动目标跟踪方法

发布时间：2019-04-02 15:39

【摘要】：目前大多数运动目标跟踪算法在目标姿态发生突变或目标被严重遮挡后会产生目标丢失现象,为解决此类问题,受人脑三阶段记忆信息处理机制的启发,本文提出一种三层旋转圆记忆模型(Spinning Tri-Layer-Circle Memory Modeling,STLC-MM)用于在目标跟踪过程中对目标模板进行更新.模型定义了三个旋转圆记忆空间用于存储和处理跟踪过程中用过的目标模板.三个圆记忆空间旋转时,记忆空间中的模板也随之以相同的速度旋转,同时模型通过模拟一些人类的认知行为,如,记忆、提取、遗忘等对模板进行更新.最后,为了验证所提方法有效性,本文将STLC-MM嵌入到粒子滤波目标跟踪框架进行目标跟踪实验.实验结果表明所提方法在目标姿态突变以及严重遮挡等方面具有较强的鲁棒性.
[Abstract]:At present, most moving target tracking algorithms are inspired by the three-stage memory information processing mechanism of the human brain in order to solve this problem, which will result in the loss of the target after the sudden change in the attitude of the target or the serious occlusion of the target. In this paper, a three-layer rotating circular memory model (Spinning Tri-Layer-Circle Memory Modeling,STLC-MM) is proposed to update the target template during target tracking. The model defines three rotating circular memory spaces for storing and processing target templates used in the tracking process. When the three circular memory spaces are rotated, the templates in the memory space are rotated at the same speed. At the same time, the model updates the template by simulating some human cognitive behaviors, such as memory, extraction, forgetting and so on. Finally, in order to verify the effectiveness of the proposed method, this paper embeds STLC-MM into the particle filter target tracking framework for target tracking experiments. The experimental results show that the proposed method has strong robustness in the aspects of sudden change of target attitude and severe occlusion.
【作者单位】：中国石油大学(华东)信息与控制工程学院;
【基金】：国家自然科学基金项目(No.61271407,No.61402535,No.61671480) 山东省自然科学基金项目(No.ZR2013FQ015)
【分类号】：TP391.41

【相似文献】