基于声音的跟踪摄像的初步研究

仿照生物的趋声性,可以实现摄像头追踪拍摄一个不在其视场范围内的声源;依据图像处理技术,可以精确定位在摄像头视场内的声源。将这两种技术结合起来,建立了一套实验系统。从利用声源发出的声音信号传输到两个分开一定距离的声音探测器上的时间差初步定位声源,再利用声源的运动规律使用图像处理技术识别并精确定位声源,通过系统与PC机间的实时通信和回转部件带动声音探测器和摄像头一起做精确灵活的转动,摄像头最终对准了声源。实现了基于声音的跟踪摄像过程。

Primary research for voice-activated camera tracking

Camera can track a sound source by imitating the creature's phonotaxis when the sound source stands outside its field of view.Also,camera can track the sound source exactly using the image processing techniques when the sound source stands in its field of view.An experiment system integrating these two technologies was set up and kept communication with the PC.When sound signals phonated by a sound source,they were received at two sound detectors which arranged apart.The PC identified the position of the sound source primary by computing the interaural time differences(ITD) of the two signals,and located the position by rotate the system.After taking several pictures,the PC identifies and locates the sound source exactly using the image processing techniques.The camera rotates and aims at the sound source finally.The voice-activated camera tracking works.