The invention relates to an elevator video call system capable of voice recognition, which comprises the following steps of: firstly acquiring a monitoring video in an elevator car to obtain an image frame sequence, then establishing an activation function and a 3D convolutional neural network according to a first frame image in the image frame sequence, and then recognizing help-seeking actions of passengers based on the 3D convolutional neural network, and finally, according to the help-seeking action, starting to acquire and recognize voice data in the lift car, and sending corresponding feedback prompts to passengers. Compared with the prior art, the purpose of starting the voice monitoring function only at the required moment is achieved, and the problem of infringement of personal privacy is solved. And meanwhile, the two functional modules of video action recognition and voice recognition are matched, so that the problem of how to give an alarm for help under the condition that the passenger cannot touch an alarm key is solved, a more comprehensive and flexible emergency help mode is realized, and the safety and guarantee level of the passenger is improved.
本发明涉及一种可语音识别的电梯视频通话系统,其先获取电梯轿厢内的监控视频得到图像帧序列,然后根据图像帧序列中的第一帧图像,建立激活函数并建立3D卷积神经网络,再根据基于3D卷积神经网络识别乘客的求助动作,最后根据求助动作,开始获取并识别轿厢内的语音数据,向乘客发出相应的反馈提示。相比于现有技术,本发明达到了仅在需要的时刻开启语音监测功能的目的,解决了侵犯个人隐私的问题。同时,视频动作识别配合语音识别两个功能模块,又共同实现了乘客在无法触碰到报警键的情况下,如何进行报警求助的问题,实现了更全面和灵活的紧急求助方式,提高乘客的安全和保障水平。
Elevator video call system with voice recognition function
一种可语音识别的电梯视频通话系统
2023-12-22
Patent
Electronic Resource
Chinese
ELEVATOR VOICE CALL REGISTRATION DEVICE AND ELEVATOR VOICE CALL REGISTRATION METHOD
European Patent Office | 2016
|European Patent Office | 2017
|