基于图像传感的快速被试动物追踪系统设计与实现

Design of an image sensor based fast tracking algorithm for animal exmeriments

提出了一种基于图像传感器的快速目标追踪算法及其硬件实现。本运算复杂度较低,具有良好的实时性。该算法被应用于一个全自动化智能多模态神经接口实验平台。该平台针对现今动物实验需要大量重复训练,需要记录的数据量大等问题提出了针对性解决方案,通过实验过程的自动化,实现了无人工干预的数据记录和动物训练流程控制。本系统具有以下两个主要优势:所有计算、控制和数据处理模块均在本地硬件进行,使动物实验环境不受计算机或服务器的体积及连接方式限制;所有模块与实验动物之间均采用无线通信,使实验动物不受线缆的限制,保证实验结果的准确性。

This paper proposed an image sensor based fast object tracking algorithm with its hardware implementation.The proposed algorithm features low computational complexity,enabling real-time tracking works.The proposed algorithm has been applied to an automatic neural interface platform for experiment performed on freely moving animal subjects.This platform can record data and control the process of experiment automated.The proposed platform enables local implementation for all the calculation,control and data processing,as well as wireless data/control commend communication,which resulting in a friendly implementation for experiments using freely moving animals.