Monitoring and controling froth status in froth flotation with 3D computer vision techniques

Created by Yongheng Zhao from CSU, China.

The froth status, including Scale, Flow Velocity, Burst Rate are usually considered as the key observations in froth floatation. The normal methods based on 2D images could cause inaccurate scale estimation because the distance between the camera to the froth surface is dynamic. We are the first to apply 3D computer vision techniques to deal with froth status estimation.

RGB-D data collected from on site kinect:

3D point cloud of the froth and mapping result:

Watershed segmentation :

The froth flow velocity:

You could find more detials here: https://youtu.be/8N4nFz-fPsk