Project number
27064
Organization
Elbit Systems of America
Offering
ENGR498-F2026-S2027
The TRACE project (Target Recognition via Adaptive Contrast Enhancement) challenges the student team to develop a prototype system that improves the visibility of a man‑sized target through adaptive contrast‑enhancement algorithms, man-sized assumed to be 1.5 meters square at 250 meters or about 2.7 mrad angular resolution. Using standard commercial off‑the‑shelf (COTS) hardware, the team will collect test imagery across varying illumination conditions (full sun, overcast, dusk/twilight) and diverse target–background combinations such as white clothing, camouflage, urban environments, and desert terrain. Core project objectives include defining a quantitative visibility metric, performing research into contrast‑enhancement techniques, and developing a post‑processing algorithm that measurably improves target visibility. The final system should operate in real time at 30 Hz with 100 ms processing latency.
As a stretch goal, the team can integrate live target marking with real‑time enhancements onto the scene in a wide-field-of-view lens.
Visibility metric:
This is intended to be a scalar score that is computed by the software based on the contrast of the target to its background. It can be per pixel or given to the target as a whole. Then after the post-processing enhancement algorithms, there should be able to an increase in the metric. This will require research, thought, and perhaps some experimentation by the team to land on a suitable method for calculating "visibility".
Hardware:
COTS monochromatic visible spectrum digital camera and lenses. Team can run video processing discrete embedded electronics, but running all software on a computer with accompanying display is acceptable.
Testing:
At least 3 light levels, e. g. full sun, overcast, dusk, twilight, etc.
At least 2 target contrast levels including light or reflective colored clothing and camouflage
At least 2 different environments, e. g. urban, sandy, rocky, forest, etc.
Stretch goal to capture and enhance images of Bigfoot.
***This project is conducting remote interviews. Please see remote interview sheet on BrightSpace (D2L) to schedule a time to speak to the sponsor.***
As a stretch goal, the team can integrate live target marking with real‑time enhancements onto the scene in a wide-field-of-view lens.
Visibility metric:
This is intended to be a scalar score that is computed by the software based on the contrast of the target to its background. It can be per pixel or given to the target as a whole. Then after the post-processing enhancement algorithms, there should be able to an increase in the metric. This will require research, thought, and perhaps some experimentation by the team to land on a suitable method for calculating "visibility".
Hardware:
COTS monochromatic visible spectrum digital camera and lenses. Team can run video processing discrete embedded electronics, but running all software on a computer with accompanying display is acceptable.
Testing:
At least 3 light levels, e. g. full sun, overcast, dusk, twilight, etc.
At least 2 target contrast levels including light or reflective colored clothing and camouflage
At least 2 different environments, e. g. urban, sandy, rocky, forest, etc.
Stretch goal to capture and enhance images of Bigfoot.
***This project is conducting remote interviews. Please see remote interview sheet on BrightSpace (D2L) to schedule a time to speak to the sponsor.***