TOMRA Mining has launched CONTAIN™, a solution that, it says, represents the next breakthrough in AI-powered ore sorting.
By analysing X-ray imagery in real time, CONTAIN identifies visual patterns that traditional sorting systems miss, giving operators precise control to optimise sorting to their specific needs, the company says.
The proprietary deep learning technology is designed to enhance the recovery of inclusion-type ores that are difficult to detect using traditional sorting methods.
Developed entirely in-house by TOMRA’s software engineers and mining experts, CONTAIN uses convolutional neural networks to perform real-time analysis of X-ray imagery, visually classifying rocks based on the probability of subsurface ore mineral inclusions. These include complex mineralisations such as in tungsten, nickel and tin ores – materials that traditionally result in high misclassification or excessive product loss.
“Our system was trained on tens of thousands of ore samples and designed from the ground up for sorting inclusion-type ores,” Stefan Jürgensen, Software Team Lead at TOMRA Mining, says. “With CONTAIN, operators can dynamically adjust the grade-recovery threshold via a touchscreen interface, enabling precise control over yield and product specifications.”
TOMRA says it has been a pioneer in applying AI to the field of sensor-based sorting for decades. The company first implemented AI-driving image processing in 1993 and later expanded its capabilities with machine learning in X-ray Transmission (XRT) and Near-Infrared (NIR) sorting. In 2018, TOMRA established a dedicated deep learning team, accelerating the development of industrial-grade AI sorting platforms. This early and substantial investment culminated in the launch of OBTAIN™, which introduced single-particle precision to high throughput ore sorting. Now, CONTAIN builds upon this foundation with targeted classification of inclusion-type ores – pushing the boundaries of automated sorting decision making, it says.
Unlike traditional optical sorting systems, which often have limitations in detecting low-grade or inclusion-type ores, CONTAIN introduces a fundamentally different approach to classification. By analysing the structure of each rock using advanced deep learning algorithms, the system identifies subtle mineralogical patterns that indicate the presence of valuable metals such as tungsten, nickel or tin. Each rock is assigned a probability score based on its likelihood of containing mineralization below the surface, enabling precise, data-driven sorting decisions. This capability allows mining operations to adapt their strategies in real time – whether the goal is to maximise concentrate grade, minimise valuable material loss, or align with processing cost constraints.
CONTAIN is built for industrial-scale performance, TOMRA claims. Because it does not rely on specific throughput or spacing on the belt, the system maintains pinpoint accuracy even in dense, fast-paced input streams. This makes CONTAIN especially effective in high-volume processing plants where consistency, speed and recovery rates are critical to profitability.
CONTAIN has been engineered to handle a wide spectrum of ore grades – from high-value deposits to low-grade, inclusion-rich rocks that have historically been difficult to process efficiently. Conventional sorting systems can be configured to detect some low-grade material, but they tend to let large volumes of gangue enter the product stream, diluting output and eroding profitability. By contrast, CONTAIN uses deep learning to classify mineralisations with exceptional accuracy, enabling precise sorting thresholds that make the recovery of low-grade ores economically viable, TOMRA says.
Jürgensen added: “Existing technologies can be configured to detect low-grade material in such ores, but this results in a high quantity of waste rocks being sorted into the product stream, diluting the product beyond economic viability. CONTAIN is exceptionally accurate in evaluating the value of a rock, making sorting thresholds for such relatively low-grade ores economically viable.”
In addition to tungsten, nickel and tin, TOMRA is actively testing CONTAIN on gold and chromite, and is exploring expanded applications in iron and copper. While still in the early stages, initial results indicate promising possibilities for expanding applications across a broader range of ore types.
Field trials at Wolfram Bergbau in Mittersill, Austria, confirmed the transformative potential of CONTAIN, according to TOMRA.
Integrated alongside TOMRA’s latest COM XRT and OBTAIN technologies, the system delivered immediate performance gains. The operation rapidly increased total plant throughput by 8%, achieved a 33% reduction in ore mineral losses and recorded its lowest-ever tails grade. The visual impact of the improvements was so striking within the first minutes of operation that the team immediately requested a second installation.
The company said: “What truly set CONTAIN apart was its ability to identify tungsten-bearing inclusions that would otherwise go undetected – particularly those embedded deep within host rock. Traditional sorting systems often fail to distinguish such subtle mineralisation, resulting in either excessive gangue or compromised concentrate quality. With CONTAIN, operators were able to finetune the balance between grade and recovery in real time, producing consistent, high-spec output with a higher tails rejection volume and reduced ore mineral losses. The downstream effect was a more stable, efficient operation and a notable drop in overall production costs.”
David Comtesse, Production Manager, Wolfram Bergbau-und Hütten AG, said: “We were absolutely overwhelmed by what CONTAIN could do. It picked up mineral inclusions we didn’t think were detectable, and did it with incredible precision even at larger grain sizes up to 65 mm. It immediately changed the way we think about sorting and processing. This isn’t just an upgrade – it’s a completely new level of performance.”
CONTAIN was designed to complement and enhance TOMRA’s sensor-based sorting ecosystem, working in concert with COM XRT and OBTAIN to deliver a comprehensive, multi-layered approach to ore processing, it says. While each technology plays a distinct role – from density-based separation to particle detection and deep learning classification – they share a unified interface and operational synergy.
“This integration gives mining operations the ability to finetune performance across the entire sorting line, with data-driven control and real-time responsiveness,” the company added. “Whether the focus is on maximising recovery from complex ores or achieving tighter product specifications, the combination of TOMRA’s technologies provides unmatched flexibility and precision in inclusion-type ore sorting. It also simplifies system scalability, allowing plants to evolve their capabilities without overhauling existing infrastructure, protecting both performance and long-term investment.”
