New image-processing algorithm for measurement of bubble size distribution from flotation froth images

Research and experience have demonstrated that the operating conditions of the flotation process are reflected in the froth's appearance. Despite recent advances in image analysis and several algorithms developed for froth flotation, there still is not a comprehensive algorithm to accurately determine bubble features from actual froth images. Being able to accurately and automatically measure bubble size distribution is an important requirement for optimization and control of the flotation process. Segmentation is one of the most effective approaches for accurately determining the bubble size distribution. In the present study, a new bubble segmentation algorithm, which utilizes an adaptive marker-based watershed transform, is proposed to measure the bubble size distribution from froth images. The developed algorithm is validated using some industrial scale froth images from flotation cells at different duties.