Image classification model for pulping froth state detection: An industrial application

Introduction:

Supervision and control of flotation plants include automated control (e.g., regulatory control of pulp levels and air rates, supervisory control of mass pull) and decision making by operators and metallurgists (e.g., specification of automated control set points and limits, manual interventions during abnormal conditions).

The condition of the froth is an important indication of froth cell performance. Froth sensors that track froth properties (e.g., bubble size, bubble velocity, bubble burst rate) are widespread in industry. Real-time froth measurements can be incorporated in automated control, as well as being available to operators and metallurgists for their decision making. Froth sensors also make real-time images of froths available in the control room, further supporting decision making.

 -> Click to read more