Once the images have been acquired, the next step is for Split-Online to transform them by applying image processing routines to extract useful information.

The first step is basic pre-processing algorithms that will prepare the image for the correct application of the information extraction algorithms.

After pre-processing the image, industry-specific modules will extract the important information desired.

Fragmentation Measurement: The fragmentation module transforms a grayscale image of rocks into an image showing three labels, particles, boundaries and patches of fines. The figure at the left shows a grayscale image before processing and the figure below it shows the output labeled image. There are a considerable number of algorithm parameters that must be accurately determined at commissioning, but once the range of these have been determined, they are dynamically calculated so that a wide range of sampling conditions can be accurately delineated and labeled. Some of the algorithms include automated dynamic delineation parameters and automatic fines identification. The accuracy of delineation is the key to obtaining accurate size results. The software allows an observer to monitor the delineations.

Pellet Measurement : Individual pellets are identified in the image and a sphere delineation is drawn around the boundary of the pellet. There are different modules available depending on the characteristics of the pellets to be measured.