The camera detection logic currently uses a lot of CPU calling mxcam functions. On the beaglebone, whcih only has one core, this causes a ton of lag in the system as geo-video-server starts using 35-50% of the cpu trying to detect cameras. This is also not ideal for the RPI, but it at least has multiple cores which stops it from bogging down the rest of the system.
Temporarily, a limitation on the camera detection logic has been made where it will skip detection if one camera is already detected. When we want to use multiple cams on the RPI, we will need to revisit this.
The camera detection logic currently uses a lot of CPU calling mxcam functions. On the beaglebone, whcih only has one core, this causes a ton of lag in the system as geo-video-server starts using 35-50% of the cpu trying to detect cameras. This is also not ideal for the RPI, but it at least has multiple cores which stops it from bogging down the rest of the system.