Recognition of flooding and sinking conditions in flotation process using soft measurement of froth surface level and QTA

Accurate recognition of abnormal conditions is crucial for control and optimization of the running of flotation process. In this paper, a novel method using soft measurement of froth surface level and modified qualitative trend analysis (QTA) is proposed for flooding and sinking conditions recognition. First, the soft measurement method based on defocus depth recovery is used to derive the froth surface level from the 2D froth image. Then, a modified interval-halving QTA is developed to extract the trend information from the froth surface level. Finally, the flooding and sinking conditions can be recognized by the classification decision tree combining the froth surface level and its trend. Offline and online experiments show that the proposed recognition method works effectively and accurately even at the early stage of the abnormal conditions.