The National Geospatial-Intelligence Agency is seeking an algorithm that can rapidly, accurately and automatically colorize large panchromatic images of up to 10 gigapixels in under 10 hours with no human intervention.
And the agency will pay $15,000 for it, according to an announcement posted Tuesday on Challenge.gov, the official listing website of challenge and prize competitions across the federal government.
“NGA is continually looking for innovative solutions to improve the delivery of GEOINT to our customers,” said Christy Monaco, director of NGA’s office of ventures and innovation. “We are also interested in innovative means to rapidly identify such solutions -- such as challenges -- and look for ways to potentially transition them into operation at NGA.”
Colored Earth images, such as those captured by satellites equipped with multispectral sensors, are comparatively grainy, less detailed and lower quality than images collected by satellites with panchromatic grayscale sensors, said Linn Wisner, NGA’s Rapid Image Colorization challenge owner. Yet, despite having the highest resolution and most detail, the lack of color in panchromatic grayscale images is sometimes a problem.
“The addition of color makes the images more useful and can assist in assessments of the information within the image,” said Wisner. “Ideally, the best images would combine high resolution with color, and that is the idea behind the challenge.”
The Challenge closes Sept. 9. Challenge solvers will not be required to transfer intellectual property or license algorithms.
For more details about the Rapid Image Colorization challenge visit: https://www.challenge.gov/challenge/rapid-image-colorization/