Computer vision-based method for online measuring the moisture of iron ore green pellets in disc pelletizer

Iron ore green pellets are widely used in steel industry. In the manufacturing of green pellets in disc pelletizer, the pellet moisture (PM) should be frequently monitored to guarantee product quality. Current methods for PM measurement are off-line and time-consuming, which cannot meet the requirement for real-time measurement and control of pelletizing process. To solve this problem, a computer vision-based method is proposed to online measure the pellet moisture. Firstly, a set of experiments are conducted to capture green pellet images with different moisture. Then, the image features are extracted and sent to polynomial regression for model establishment, so that the PM can be measured by the proposed model. Test on iron ore pellets images demonstrate that the proposed method can measure the PM accurately and only cost 0.361s for each pellet image. The proposed method has advantages in rapid, low cost and non-destructive, and it is helpful to improve the automation level of production process in pelletizer.