We specialize in
developing "image understanding" software
so that a robot or operator assist system can perceive
its environment. Algorithms recognize "features"
in the robot's environment such as the target task,
positions of other vehicles and humans, and obstacles.
We have developed algorithms for applications as diverse
as detecting dumptruck locations for loading applications
to classifying animal behaviors to detecting conveyor
belt splice failures. Our recent focus is developing
sophisticated classifiers for very complex, outdoor
terrain.
| •
Color segmentation
• Terrain /
object classification
• Thermal image processing
|
•
Medical image analysis and registration
• Dynamic object
classification
• Fusion with range sensor data
|
Cut line for alfalfa harvesting
|
 |
Infrared image
|
|
Vegetation detection
|
|
