Vision position using monocular slam such as svo,ptam

Hi guys
I want to use monocular slam for vision position I read this tutorial , to understand how to use lpe . does it need an extra program to scale output vision position from such as ptam to metric the position or not and I can send position that comes from monocular slam directly to px4 and don’t scale the position?

Some suggest that you should use multisensor fusion framework to fuse position corrections from different sensors, as well as scale estimation in SVO or PTAM or other visual odometry or slam framework.