linville/gray16lib

📈 16-bit grayscale processing algorithms for use with QImage::Format_Grayscale16.

Gray16Lib + Gray16Demo

Gray16Lib is a library to provide various grayscale image processing algorithms to down-convert 16-bit grayscale images to bit-depths that are viewable on standard displays and maintain image contrast.

16-bit grayscale imagery is commonly encountered in scientific fields. Two notable examples include X-Ray and Aerial imaging. Sensors may raster data such as light intensity, or elevation data, using a 16-bit value. Typical consumer displays are capable of displaying 24-bit color (8-bits per channel: Red/Green/Blue, with some supporting pseudo 10-bit, usually 8-bit + FRC (Frame Rate Control) - basically faking the extra depth by flashing a pixel between two colors). To view the 16-bits of data on a display capable of displaying on 8-bits, it becomes necessary to appropriately "throw away" pieces of data while still maintaining important image contrast or splitting the 16-bits across multiple channels of 8-bit data.

Gray16Demo is an example application that will load a 16-bit Grayscale PNG and allow the user to interactively explore different image processing algorithms. Open your own 16-bit PNGs or load an example from the Examples menu (see the included 16-bit examples here).

Requirements

• Qt 6.2+
  • QImage::Format_Grayscale16 was added in Qt 5.13 (released June 19th, 2019). References: QTBUG-41176, Gerrit patch.

Image Processing Algorithms

1. Manual
   • Manually select Black-Point and White-Point values.
2. Offset / Gain
   • Offset pixel values then bit-shift them. No floating point operations.
   • Auto-Offset, Manual Gain
     • Offset chosen by least-exposed pixel, that isn't zero-exposed.
     • Offset (+ 255) chosen by most-exposed pixel, that isn't fully saturated.
3. False-Color (conversion to RGB32)
   • Converts a 16-bit grayscale to an RGB gradient. Supports various gradients.
   1. Black Body

Compiling Gray16Lib

git clone https://github.com/linville/gray16lib.git
cd gray16lib
mkdir build && cd build
cmake ..
make

Todo

• Add more algorithms.
• Add more examples.

Additional Reading

• Mootz and Mathews. "Displaying and Stretching 16-bit per Band Digital Imagery." USDA FSA

gray16lib's Issues

build issue: std::reduce

I am trying to build the code on ubuntu using Qt5.13.0 in combination with gcc-9 and g++-9.
All compiles well except line 119 in image_histogram_widget.cpp

I get the error: no matching function for call to ‘reduce(QVector::const_iterator, QVector::const_iterator)'

Code on line 119 | int sum = std::reduce(histogram.constBegin(), histogram.constEnd());

std::reduce takes a minimum of 3 arguments, you are just providing 2

Can you please help resolving this issue.