Let's talk DSO, page-190

Tomra Mining have a sensor-based sorting test with a new feature called OBTAIN.

Here's a breakdown of the key points:

Particle size distribution: The test uses a wide range of particle sizes, from 8 to 35-40 millimeters. This is outside the ideal range (6-100 mm) for XRT sorting mentioned earlier.
Spiked sample: The material being sorted has trace amounts of something added, likely to make the sorting process easier to visualize.
Sorting goals: The goal is to eject 20% of the material at a high capacity of 70 tons per hour. This suggests prioritizing high throughput over perfect separation.
Obtain function: This is a new feature that uses machine learning to identify individual rocks within clusters.Traditional sorting might struggle with clustered particles, potentially ejecting good material with bad.OBTAIN allows for precise analysis of each individual rock, regardless of how fast the processing is.

Benefits of OBTAIN:
Improved sorting accuracy: By separating clusters, OBTAIN enables more precise ejection of unwanted particles.
Increased recovery: No more good material being lost along with bad due to overlapping particles in clusters.
Higher capacity: Sorting can be done at high speeds without compromising accuracy thanks to OBTAIN.

Overall, OBTAIN allows for achieving both high product quality and high recovery rates, overcoming a traditional trade-off.In essence, this test demonstrates how OBTAIN can significantly improve sensor-based sorting for challenging scenarios with large particle size variations and clustered materials. It achieves this by leveraging machine learning to analyze individual rocks within clusters, leading to more accurate sorting and higher throughput.