STOUT TRUE VISION™
Stout has developed a proprietary vision system combined with cutting edge software called Stout True Vision™.
Stout True Vision™ cameras capture high resolution color images, passing them through a convolutional neural network for analysis. Post processing algorithms produce unmatched precision for identification and automation implementations.
Stout True Vision™ Agriculture – Deep Learning Advantage
- True plant identification
- Deep learning-based neural network differentiates crops from weeds
- Precision seed line tracking improves coverage by minimizing blade opening
- Neural network trained with Deep Reinforcement Learning
- Vision system does not need to be trained for specific fields
- System detects actual plant centers, allowing for accurate blade control
- Adapts automatically to ambient lighting conditions, soil type, and plant maturity
- Not dependent on plant size or spacing, allowing use on transplants or direct seed
Stout True Vision™ – Mechanical Design
- Robust stainless-steel housing
- Sealed, long life LED strobe lights
- Fixed-focus rugged industrial camera and lens
- Rugged, waterproof, and shockproof industrial cables and wire harness
- Epoxy-potted interface electronics
Future of Stout True Vision™
This revolutionary vision system is at the heart of what we do and will provide future capabilities that industries have been requesting for years. Contact us for more details.
HOW IT WORKS
AGRICULTURE
- Proprietary vision software and firmware developed in-house at Stout
- Camera images go through a convolutional neural network (CNN) which detects the location, type of plants, and weeds in an image
- Post-processing algorithms convert the CNN output to timing information for the machine’s tools
- Controller activates tools as plants pass by the correct location
