Computer vision in granular processes by real-time physics
| Reference number | |
| Coordinator | Umeå universitet - Institutionen för matematik och matematisk statistik |
| Funding from Vinnova | SEK 498 600 |
| Project duration | October 2016 - April 2017 |
| Status | Completed |
| Venture | Strategic innovation programme for process industrial IT and automation – PiiA |
| Call | Strategic innovation program PiiA - summer 2016 |
Important results from the project
The ability to track and determine the state of material in granular processes by combining sensor data with physics based simulation has been investigated. The aim is to increase the degree of teleoperation and automation, and better understand the causal relationship between different subprocesses to optimize the full process. A number of solutions for mining and mineral processes have been outlined and tested in simulation.
Expected long term effects
The feasibility study has resulted in three solution proposals. 1) Machine vision for rock breaker automation. 2) Bulk material property determination on conveyors. 3) Tracking of material in ore stock piles. Each solution proposal is accompanied with a demonstrator, an algorithm, literature study and needs analysis. The solutions are expected to eventually lead to teleoperated semi-autonomous rock breakers and more efficient crushing and milling of ore thanks to increased control over property of the incoming material.
Approach and implementation
The project involved selecting and focusing on specific sub-processes, problem description & needs analysis of these, and test and demonstration of the proposed solution using simulations/simulators. This approach created a sustained interest and motivation of all parties to contribute with their knowledge and ideas, provide background data and create meetings with key persons. Project meetings were held weekly and by remote.