An Adaptive Tracking for Moving Target using Active Camera

使用主动相机的运动目标自适应跟踪

基本信息

批准号：
07650509
负责人：
OZAWA Shinji
金额：
$ 1.28万
依托单位：
Keio University
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (C)
财政年份：
1995
资助国家：
日本
起止时间：
1995 至 1996
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-07650509/
关键词：
Active Camera Moving Target Tracking Moving Picture Processing

项目摘要

1. Construction of the Active CameraIn 1995, we construct the zoom control algorithms on the work station which we have had, and fundamental experiments based on personal computer system is done. At the active camera system which we construct for fundamental experiments, the camera zooming control algorithms and processing of the image of CCD camera are executed by personal computer, then it was not sufficient to track the target which moves rapidly.In 1996, we construct new active camera, in which we used image processor, so that the speed of tracking process is increased and all of the image processing are completed in video frame rate (33msec). Moreover image processing and camera control are done parallel to speed up for process.2. Application for estimation of 3 dimensionla moving of target.In 1996, we construct the tracking algorithm for two or three moving targets. From the experiments, it is confirmed that the active camera system is applicable for the scene in which two persons are walking.In 1997, as the application of the automated tracking system, we use the active camera system for motion stereo method to estimate the 3 dimensional moving of target. At the 3 dimensional measurement by the traditional stero method, it has the problem that if the distance between camera and target becomes longer the disparity value becomes smaller, so that the accuracy of measurement is low.So we set active camera system on the robot arm. And according to the distance between target and camera which is calculated by motion stereo principle, the system changes the distance and focal length adaptive. The results of experiments show that active camera which we constructed, works as if the distance between target and camera becomes longer, the focus length of the lens is changed adaptive, then it is confirmed that the accuracy of measurement does not become low.

1。主动摄影师1995的构建，我们在工作站上构建了缩放控制算法，并完成了基于个人计算机系统的基本实验。在我们为基本实验构建的主动摄像机系统上，摄像机变焦控制算法和CCD摄像机图像的处理是由个人计算机执行的，那么它不足以跟踪迅速移动的目标。在1996年，我们构建了新的Active Camper，在该摄像机中，我们使用了新的Active Camera，在该摄像机中，我们使用了图像处理器，因此跟踪过程的速度增加了图像处理的速度和所有图像处理率（33333333帧均已完成（33框架）（33框架）（33框架）（33框架）（33框架）（33框架）（33框架）（33框架）（33框架）（33框架）（33帧）。此外，图像处理和摄像机控制是平行于加速进行处理的2。估计目标3二维移动的应用。在1996年，我们为两个或三个移动目标构建了跟踪算法。从实验中可以证实，活动相机系统适用于两个人走路的场景。在1997年，作为自动跟踪系统的应用，我们使用活动摄像机系统进行运动立体声方法来估算目标的3维移动。在传统的Stero方法的3维测量中，它的问题是，如果相机和目标之间的距离变长，则视差值会变小，因此测量的准确性较低。因此，我们将活动相机系统设置在机器人臂上。根据目标立体原理计算的目标和摄像机之间的距离，系统会改变距离和焦距自适应。实验的结果表明，我们构建的主动摄像机工作，就像目标和相机之间的距离变长一样，镜头的焦点长度改变了自适应，然后确认测量的准确性不会较低。