基于传声器阵列技术与图像识别的机动车鸣笛定位优化研究  

Research on Optimization of Whistle Source Positioning Based on Microphone Array Technology and Image Recognition

作　　者：郑福进 胡国祥 ZHENG Fujin;HU Guoxiang(School of Civil Engineering and Architecture,Wuhan Institute of Technology,Wuhan 430000,China)

机构地区：[1]武汉工程大学土木工程与建筑学院,湖北武汉430000

出　　处：《电声技术》2025年第2期113-118,共6页Audio Engineering

摘　　要：提出一种基于图像识别的机动车鸣笛声源定位优化方法,结合不同尺寸的机动车鸣笛喇叭的声学特性差异,利用图像识别技术和传声器阵列技术实现对鸣笛声源的定位优化。首先,通过传声器阵列获取机动车鸣笛声,利用梅尔倒谱系数分析其所属车辆尺寸类型。其次,通过摄像头获取鸣笛发生时的图像信息,利用YOLOv11算法获取图像内的车辆所属类型及其分布位置。最后,将鸣笛声识别结果与图像识别结果进行匹配,利用规则权重为图像中的机动车目标赋予不同的权重值,进而反映为机动车在图像坐标中的权重值,再结合麦克风阵列坐标信息,从而提高鸣笛声源的定位准确度。This paper proposes an image recognition-based method to optimize the location of vehicle whistle sound sources.Combining the acoustic characteristics of different sizes of vehicle whistle horns,the image recognition technology and microphone array technology are used to optimize the location of vehicle whistle sound sources.Firstly,the whistle sound of motor vehicles is obtained by microphone array,and the size type of the vehicle is analyzed by Mel cepstrum coefficient.Secondly,the image information of the whistle is obtained by the camera,and the type and distribution position of the vehicles in the image are obtained by using YOLOv11 algorithm.Finally,the recognition results of the whistle sound are matched with the image recognition results,and the regular weights are used to give different weights to the motor vehicle targets in the image,which are then reflected as the weight values of the motor vehicle in the image coordinates,and then combined with the coordinate information of the microphone array,so as to improve the positioning accuracy of the whistle sound source.

关 键 词：鸣笛声分类 图像识别 规则权重 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]