Recently, Fabio Manganiello, a foreign technology blogger who has just been promoted as a dad, used Raspberry Pi and Tensorflow to create a smart baby monitoring system in order to better take care of his children. The system not only automatically sends out notifications when a child is crying, but also analyzes the cry to infer the child's needs.
It is reported that Fabio usually stays in the studio at home, and it is often difficult to hear the children in the baby room because of his concentration on work, so he needs a smart device that can monitor the baby's situation and send out notifications automatically. But the baby monitoring devices available on the market can only serve as a simple "walkie-talkie" and are not intelligent enough in all aspects. Therefore, as a programmer, Fabio has the idea of making a self-made baby monitoring device.
Fabio first defined some required functions for this "smart baby monitor":
-
Based on low-cost Raspberry Pi (RaspberryPI) and USB microphone .
-
The device is not a mere walkie talkie. When the child starts/stops crying, it can detect the child's cry and issue a notification (ideally via a mobile phone), or track data points on the dashboard, or run corresponding tasks.
-
The device can transmit audio on speakers, smart phones, computers and other devices.
-
The system is not affected by the distance between the source and the speakers, and there is no need to move the speakers around the house.
-
Equipped with a camera , you can use the camera to monitor the child in real time. When he starts crying, I can capture pictures or short videos of the crib to check what's wrong.
Based on the above requirements, Fabio chose a combination of RaspberryPi Zero and PiCamera to create a monitoring device:
The brief step is to install the Linux operating system on the Raspberry Pi through the SD card, run the Tensorflow model, and collect audio samples of the baby's crying. Once you have enough audio samples, you can generate a data set and train the model, and then use the model to predict the baby's crying behavior. Finally, you can watch the video on your phone through the VLC application or the RPi Camera Viewer application. Detailed code implementation process