Summary Report of Pattern Constancy Performance

NASA Langley Research Center

Visual Information Processing Lab

Summary Report of Pattern Constancy Performance
of Retinex/Visual Servo (RVS) Platform

	Pattern recognition plays a crucial role in advanced, automated pilot visibility systems. Whether it be the detection of terrain features in the path of the aircraft, hazards on or around a runway, or other aircraft in flight, the automatic detection of such hazards is the logical next step in improving the safety of our airways. Here we highlight a computational foundation for resilient, generic pattern recognition for smart aviation imaging sensors. The categories above show examples of how pattern detection is improved with RVS processing for aviation hazards, especially in during poor visibility conditions. The pattern constancy demonstration addresses the issue of how well, in general, we can acquire images of the same scene over time, and extract a similar pattern despite wide-ranging changes in imaging and visibiility conditions. These results (full report here) provide strong support for RVS processing being a generic or " universal " front end for any higher level pattern recognition processes. This then is an enabling technology for automated detection of external aviation hazards.

Pattern recognition plays a crucial role in advanced, automated pilot visibility
systems. Whether it be the detection of terrain features in the path of the aircraft,
hazards on or around a runway, or other aircraft in flight, the automatic detection
of such hazards is the logical next step in improving the safety of our airways. Here we
highlight a computational foundation for resilient, generic pattern recognition for smart
aviation imaging sensors.

The categories above show examples of how pattern detection is improved with RVS processing
for aviation hazards, especially in during poor visibility conditions.

The pattern constancy demonstration addresses the issue of how well, in general, we can acquire
images of the same scene over time, and extract a similar pattern despite wide-ranging changes
in imaging and visibiility conditions.

These results (full report here) provide strong support for RVS processing being a generic
or " universal " front end for any higher level pattern recognition processes. This then
is an enabling technology for automated detection of external aviation hazards.

Responsible NASA Official: Glenn Woodell

NASA Privacy Statement, Disclaimer, and Accessibility Certification