Traditionally, once a week, operators conducted a startup of the belt press, manually adjusted the polymer dose once steady-state conditions appeared and intermittently monitored dewatering operation and cake quality throughout the day. This practice, although familiar, introduced variability: operators with different levels of experience would interpret visual cues differently, often erring on the side of higher doses to ensure dewatering targets were met, and spending considerable time walking back and forth to check sludge appearance and cake quality.

At startup, our principal scientist, Dr. Robin Giguere, noticed that the color and coverage of the sludge cake on the belt provided a clear visual indicator of dewatering effectiveness. Inspired by this observation, the team asked: what if a camera could continuously monitor these cues and deliver feedback directly to an operator or the control system? This question set the stage for the SludgeVision proof of concept.