Audio book quality inspection method and system, electronic equipment and storage medium

The invention provides an audio book quality inspection method and system, electronic equipment and a storage medium, and relates to the field of information detection. The method comprises the steps of performing background noise removal on a to-be-recognized audio based on an audio source separation method to obtain an initial to-be-recognized audio; obtaining a multi-language recognition model which is completely trained, wherein the multi-language recognition model comprises a plurality of language recognition sub-models; and inputting the initial to-be-recognized audio as a target audio into the multi-language recognition model to obtain a to-be-detected semantic text, and performing text rule detection based on the to-be-detected semantic text to obtain a detection result. 本申请提供了一种有声书质检方法、系统、电子设备及存储介质，涉及信息检测领域。方法包括：基于音频源分离法对待识别音频进行背景噪声去除得到初始待识别音频；获取训练完备的多语言识别模型，所述多语言识别模型包括多个语言识别子模型；将初始待识别音频作为目标音频输入多语言识别模型得到待检测的语义文本，基于所述待检测的语义文本进行文本规则检测得到检测结果。