The Visual Servo
Smart, Active Enhancement
Based Upon Visual Intelligence
New Automatic Methods for Still Image and Real-time Video-FLIR Visual Optimization
Daniel Jobson, Glenn Woodell, Glenn Hines
NASA Langley Research Center
Zia-ur Rahman
College of William and Mary
Current enhancement technologies are entirely passive. Good images possess a high degree of contrast, clarity, and sharpness. By using these guidelines, the Visual Servo is suitable for processing scenes acquired in poor visibility conditions. It includes an automatic mechanism for detecting foggy and hazy imaging conditions and compensates for these conditions by intelligent processing, providing better-than-observer performance.
Based upon the criteria of "good images having a high degree of contrast, clarity and sharpness," we have developed a set of Visual Measures for automatic assessment of contrast and brightness, sharpness, and extremely low contrast that characterizes haziness or fog.
The Visual Measures are used in an enhancement feedback loop to provide better-than-observer performance.
The automatic Visual Servo has been implemented in fully operational, experimental software and tested on several thousand highly diverse color still images as well as on color and infrared video. These tests include data from clear weather and night time 757 flights, Langley gantry-based imagery in fog, rain, haze, and smoke, Piper Cherokee flights in haze and clouds, and highway-based imagery in heavy rain, fog, and road spray. Future flights of opportunity will include fog, rain, and haze.
A patent disclosure was filed by NASA/Langley and William and Mary in November of 2002.
Visual Servo Provides:
Better-than-observer visibility - Visibility is dramatically enhanced and the enhamncement visibility limit is set only by the signal-to-noise ratio in the image.
Near-Zero visibility - Capability of enhancing images acquired in near-zero visibility conditions.
Applications
- Real-time multisensor enhancement for high-end commercial and military aircraft
- Real-time video enhancement for GA low-cost systems
- Enhancement for underwater imaging applications
- Real-time video enhancement for automobiles
This site maintained by Glenn Woodell