Aframe extract stream from scene with webcam and AR

I want to develop an application based on augmented reality using a-frame and I need to transmit the streaming of the scene through webrtc. I want to remotely see what the user sees locally (webcam + AR). I have been testing and I have only been able to extract augmented reality from the canvas of the scene.

//html 

    <a-scene id="escena" embedded arjs="sourceType: webcam; debugUIEnabled: false;" vr-mode-ui="enabled: false"> ...</scene>

    //javascript
    var scene = document.getElementById('scene')
    var stream =  scene.canvas.captureStream();

Imagen extraction only canvas from scene

But I need to merge webcam images and the augmented reality together from the scene. Or have simultaneously access to webcam images to be able to merge them separately.