项目名称: 基于搜索目标置信度的主动视觉跟踪方法研究及低空飞行跟踪实验研究
项目编号: No.61271432
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 无线电电子学、电信技术
项目作者: 赵晓光
作者单位: 中国科学院自动化研究所
项目金额: 83万元
中文摘要: 本项目以人类视觉跟踪的实际过程为基础,提出并研制一种高机动性仿生眼摄像装置,由此建立一个能够实际运行的仿生眼运动跟踪系统,它包括图像采集、分析处理,云台摄像机定标以及对跟踪目标行动轨迹的预测等组成部分,本质上是一个闭环智能跟踪控制系统。在该系统中,首先研究双运动模式下(摄像机运动-目标运动)的仿生视觉搜索机理和方法,然后提出目标图像置信度的概念和基于粒子滤波算法的跟踪方法 ,最后结合闭环智能控制系统的实际跟踪特性,研究自适应神经网络预测的相关理论与方法,实现仿生眼视觉系统在低空飞行条件下对地面目标的运动跟踪。
中文关键词: 主动视觉;相机标定;图像采集;仿生眼;运动跟踪
英文摘要: Based on our understanding to the actual process of human visual tracking, this project proposed and will develop a high mobility bionic eye camera device, with which one can set up actually operational bionic eye motion tracking system, which is essentially intelligent closed-loop control system included image acquisition, information processing, PTZ camera calibration and prediction of tracking object action trajectory, in this system, first of all,our study focuses on bionic visual search mechanism and method of bi-mode motion(camera motion-target motion), and then the second is to put forward the concept of confidence level of the target image and tracking method based on particle filtering algorithm, and finally combined with actual tracking features of the closed-loop intelligent control system, we study predictive correlation theory and method of the adaptive neural network in order to achieve bionic vision system's motion tracking of ground targets under the low flying altitude conditions.
英文关键词: ACTIVE VISION;CAMERA CALIBRATION;IAMGE ACQUISITION;BIONICEYE;MOTION TRACKING