USNR applies deep learning to automated grading systems
Advances in high-speed computing have enabled USNR to apply Deep Learning to the image processing systems that feed their automated grading systems. You benefit from faster start-ups and even more accurate grading solutions.
Precise defect detection
The foundation of highly accurate grading is highly precise defect detection. Deep Learning provides this added advantage to take your automated grading to the highest level on the market today. In development since 2012, USNR’s Deep Learning is delivering results in over 35 mills around the world.
Deep Learning has multiple features and benefits:
- Increases the speed and accuracy of defect detection
- Significantly reduces the time required for start-up and commissioning of auto-grading systems
- Enables the accurate identification of defects unique to a species, region, or grade, that would otherwise be difficult and time-consuming to identify using conventional methods
- Delivers more value and recovery through advanced optimization
- Already delivering results in over 35 mills!
It can also be used in several applications, such as:
- BioVision sawmill grading at the edger and trimmer
- LHG lineal grading in the planer mill
- THG transverse grading in the planer mill
- Upgrades can be applied to existing auto-grading and vision scanning systems