RockMass is working hard to break into the mining industry and tunnel engineering niche markets.
The Toronto-based startup is using NVIDIA AI to develop a mapping platform to help engineers assess the stability of mines and tunnels under construction.
Currently, as a safety precaution, geologists and engineers will stand five meters away from the rock and visually evaluate the risk factor of the rock formation. But RockMass co-founder and CEO Shelby Yee believes that this is not an ideal way to ensure the accuracy of the results.
"Assessed by the method they are using now, the whole process takes nearly 90 minutes, and it only takes about 5 minutes to complete with our technology." Yee said.
RockMass is using the hands of field engineers to test its handheld device Mapper, which is used in mining, geological prospecting and civil engineering. The startup is developing an AI platform for robots, drones and handheld devices that capture geological data.
Now, the startup’s Mapper AI device provides a safer way to keep engineers away from tunnels that may collapse; it also provides a faster data collection and processing system. Robots and drones using this platform can enter areas with higher risk factors.
RockMass's customers include Brazilian mining company Nexa Resources, which is trying to use RockMass's technology to improve its automation and safety levels.
AI for geological technology
For years, engineers have been using traditional equipment to measure the angle of the rock surface, such as an optical measuring device mounted on a tripod, similar to theodolite. They need to find the so-called weak surface to determine the fracture point in the tunnel and rock formation.
Engineers measure the surface of the rock formations and collect data used to construct the so-called stereonet. Stereographic projection nets can present three-dimensional shapes (such as cobblestones) on two-dimensional displays.
In the traditional way, engineers need to bring the data obtained from the field back to the office, and then transfer the data to the computer in order to build a stereographic projection network.
The startup’s technology can provide an easier way. Its handheld device is equipped with a sensor to make such measurements. The lidar sensor and inertial measurement unit can map the direction of the weak surface in the rock formation. In addition, the device can work normally even in an underground environment without GPS, wireless communication and light.
By using the information provided by these sensors, RockMass's software can quickly identify the data available to engineers within minutes. The company is working to help field engineers capture and process field data. "We can view the data in real time," Yee said.
AI that meets high computing requirements
According to co-founder and CTO Stuart Bourne, the platform used by RockMass to collect field data requires very high computing power. The company's equipment relies on NVIDIA Jetson 's robot performance and is supported by CUDA, cuDNN and TensorRT software libraries.
"Compared to the energy it consumes, Jetson's computing power is very high." Bourne said.
The startup uses the CUDA library to process data in a cloud instance running NVIDIA GPU in real time, and then process stereographic projection networks for customers.
"No one can collect and process data like we do," Yee said. "We can process data in the cloud in real time, thanks to the computing power of the GPU."
RockMass plans to further develop its drones and robots to launch a trial version next year.