Key Takeaways
- AI fragmentation analysis with Orica FRAGTrack uses stereoscopic 2D/3D imaging to size post-blast rock automatically, with no manual sieving or scaled photos.
- FRAGTrack ships in several forms: Conveyor, Crusher, GeoSpatial and a newer excavator-mounted variant. Together they cover everything from the dig face to the crusher.
- WipWare WipFrag and Split-Desktop are the long-standing image-analysis alternatives if you want a lower-cost or desktop option.
- Binocular (stereoscopic) capture lets these systems handle variable lighting and rock colour without a fixed reference scale in every shot.
TL;DR
Point cameras at broken rock, on the muck pile, the shovel, the truck or the conveyor, and let computer vision measure the particle size distribution. Orica’s FRAGTrack automates this continuously, while WipFrag and Split-Desktop suit ad-hoc or single-site use. Then compare results against each blast’s design to learn what works.
What Is Fragmentation Analysis?
A blast breaks solid rock into pieces of many sizes. Fragmentation analysis measures that size range, usually as a particle size distribution (PSD) curve and values such as P50 and P80. It tells you how well the blast did its job.
Fragmentation matters because it drives everything downstream. Oversize boulders slow loading and may need secondary breaking. Coarse feed also cuts crusher and mill throughput. Too many fines can mean wasted explosive energy, or ore losses in some deposits. Historically, sites measured fragmentation by screening samples or by taking photos with a scale object in frame. Both methods were slow and disrupted operations, so few blasts were measured.
How Do I Monitor Fragmentation With AI?
Today the measurement work is done by computer vision. The main choice is how far you want to automate it.
Continuous fragmentation analysis with FRAGTrack
Orica’s FRAGTrack uses binocular stereoscopic cameras and AI image analysis. It detects scale automatically and estimates PSD from each image, without a reference object or a pause in the load-and-haul cycle. The Conveyor version triggers on passing material and feeds comminution and throughput tracking. At the crusher, another version analyses fragmentation truck by truck as material dumps. The GeoSpatial variant georeferences each sample. Finally, the excavator-mounted version pushes measurement back to the dig face.
Lighter options: WipFrag and Split-Desktop
If you are not ready for a fixed sensor network, WipWare’s WipFrag is a good entry point. It is the original commercial optical granulometry tool, in development since 1986. You capture images with a phone, tablet, drone or camera. It then estimates PSD and rock-quality designation without scales, screens or sieves, and it detects scale and slope from the image itself. Split-Desktop is the other well-established tool in this space.
Linking results back to the blast
Measurement alone does not improve blasts. The value comes from linking each PSD result to the blast that produced it. GeoSpatial-style georeferencing makes this possible. You can then compare fragmentation against powder factor, hole spacing and rock type. That feedback loop is what AI-optimized blast design platforms learn from.
Where Fragmentation Analysis Fits in Blast Evaluation
Across all three tools, the core work is the same: detect particle boundaries, estimate scale and compute a size distribution. The products differ mainly in deployment automation and integration depth, not in the AI technique. Therefore, choose based on how often you need data and where it must go.
Fragmentation is one part of a wider blasting evaluation, alongside ore movement and wall damage. It also connects forward to the plant, because conveyor measurements help explain changes in crusher performance.
Upstream, better autonomous drilling accuracy tends to give more even fragmentation. Keep in mind that image methods see only the surface of a pile. Fines are often hidden or blurred, so treat the fine end of any curve with care.
FAQs
How accurate is image-based fragmentation analysis?
It is accurate enough to track trends and compare blasts, especially for coarse and mid-size fragments. Accuracy drops for fines, because small particles hide behind larger ones or fall below camera resolution. Many sites calibrate against occasional physical screening.
What is the difference between FRAGTrack and WipFrag?
FRAGTrack is built for continuous, automated measurement at fixed points such as conveyors, crushers and excavators. WipFrag works from images you capture yourself with a phone, drone or camera. FRAGTrack suits integrated digital pipelines, while WipFrag suits ad-hoc or single-site work.
Why does fragmentation analysis matter for the mill?
Fragment size affects how fast shovels load and how much energy the crusher and mill need. Finer, more consistent feed usually raises throughput. Fragmentation analysis gives you the data to link blast design to plant performance.
Can AI fragmentation analysis replace sieving?
For routine monitoring, largely yes, because it measures far more material than any sieving program could. However, sieving is still useful for calibration and for checking the fine fraction. Most sites use a small amount of screening to keep image results honest.
Try It With Geocluster
Fragmentation data is most valuable when you reconcile it against blast design parameters and your geotechnical model, not when you view it in isolation. Geocluster helps you bring data like this into a single research harness alongside the rest of your geological and mining data. Check it out on GitHub, or read how the Geocluster Research Harness works.