声场可视化系统中声像阵列空间关系标定研究

在声场可视化系统中,摄像机与传声器阵列的空间位置及角度的关系不准确将导致声源定位出现误差。而目前对于声场可视化系统中的声像阵列空间关系的研究较少,缺少对声像阵列空间关系的简单有效的标定方法。因此,提出了一种可计算任意位置与姿态的摄像机与传声器阵列之间空间关系的标定方法,可有效消除由传声器阵列与相机的空间关系不匹配导致的声源定位误差。通过对已知声源在多个空间位置的声源定位误差进行最小化处理,利用分步迭代计算的方法,分别计算出摄像机与传声器阵列的空间位置与角度,最终得到声像阵列之间真实的空间位置与角度关系。仿真与实际实验均表明本方法可对任意位置与姿态的摄像机与传声器阵列的空间关系进行准确标定,能有效提高声场可视化系统的声源定位精度。提出的声像阵列空间关系标定方法可应用于各种声场可视化系统中,对提高声场可视化系统的声源定位准确度有着积极的意义。

Research on spatial relationship calibration technique for microphone array and camera in sound field visualization system

In sound field visualization systems, imprecise spatial position and angle relationship between the micro phone array and camera could lead to the error of the sound source localization. However, few of the previous studies have considered the spatial relationship between the camera and microphone array in sound field visualization systems. So there lacks a simple and effective spatial relationship calibration method for the spatial relationship between the camera and microphone array. In this paper, the calibration method for calculating any spatial orientation and position relationship between microphone array and camera is proposed, which can effectively eliminate the sound source localization error caused by the mismatched spatial relationship between the microphone array and camera, and then improve the precision of the sound source localization for sound field visualization system. Through minimizing the sound source localization errors of the known sound source in multiple positions, the spatial position and orientation between the camera and microphone array could be calculated respectively with stepped iteration calculation. Simulation and experiment results both show that the proposed method can accurately calculate any spatial orientation and position relationship between the camera and microphone array in a sound field visualization system, and could effectively improve the sound source localization accuracy of sound field visualization systems. The proposed spatial relationship calibration method could be used in various sound field visualization systems, and has an active significance for improving the sound source localization accuracy of sound field visualization systems.