基于声音采集与分类技术的报警系统设计

Alarm System Design Based on Sound Acquisition and Classification Technology

环境声音包含丰富的信息,通过声音识别可以快速判断异常信息,从而为异常确认、快速处理争取时间。给出了一种基于FPGA+ADC的声音报警系统,模拟采集由语音芯片XX73311完成,数据传输由太网PHY电路XX83848实现。系统可高质量完成模拟量到数字量的转换,并以短时能量端点检测、特征参数提取为基础训练样本,以声音识别技术完成异常声音判断,实现异常报警。试验结果表明,使用声音识别具有良好的实际效果,报警系统能够完成目标任务。

Environmental sound contains a wealth of information.Through sound recognition,abnormal information can be quickly judged,so as to buy time for abnormal confirmation and rapid processing.A sound alarm system based on FPGA+ADC is presented.The analog acquisition is completed by voice chip XX73311,and the data transmission is realized by Ethernet PHY circuit XX83848.The system can complete the conversion from analog quantity to digital quantity with high quality.It trains samples based on short-term energy endpoint detection and feature parameter extraction,and uses sound recognition technology to judge the abnormal sound to realize abnormal alarm.The test results show that the use of sound recognition has good practical results,and the alarm system can complete the target task.