LinkSprite facial recognition welcome system consists of the following: a face capture camera, a pcDuino9s, a router and a cloud reporting system (gatekeeper.linksprite.com). Please note that LinkSprite also have percepcam system. It uses same face capture camera as the one used in this welcome system. But they are different systems.
The following highlights the main steps:
- Configure pcDuino9s Android to be English if it defaults to Chinese. A video can be found at here.
- Install the welcome apk on pcDuino9s. This apk will work as a FTP server to accept face capture push from face capture camera and do local face recognition on pcDuino9s (this is very different from the percepcam system, where the face recognition is done on the cloud). The welcome system does the facial recognition on pcDuino9s locally. So if there is Internet disruption, the welcome system still works. When the Internet resumes, it will sync the log with the cloud (gatekeeper.linksprite.com).
- Configure a fixed IP of pcDuino9s in the router. A video can be found at here.
- Create an account at gatekeeper.linksprite.com. If you need approval faster, please email email@example.com.
- Add the pcDuino9s to your gatekeeper account by MAC address of pcDuino9s. A video can be found at here.
- Configure the built-in FTP client of the face capture camera, ie. point the FTP client of the face capture camera to the pcDuino9s FTP server. When the face capture camera detects a face, it will push the captured face to the FTP server of the pcDuino9s, and pcDuino9s will do facial recognition locally. (how to access face capture camera J1’s configure page, video)
- Manage the subject on the cloud portal, and it will sync with the pcDuino9s when there is Internet connection to the pcDuino9s.
Finally, it works!!!
Android pcDuino9s has limited processing power. One pcDuino9s can be used to process faces from one face capture camera. If you need to process more cameras, please consider percepcam system. Or you can use more pcDuino9s box.