GY-63_MS5611/libraries/Kelvin2RGB

Kelvin2RGB

Arduino library for converting temperature to RGB values

Credentials

This library is based upon an article of Tanner Helland and a related story by Neil Bartlett http://www.tannerhelland.com/4435/convert-temperature-rgb-algorithm-code/ http://www.zombieprototypes.com/?p=210
https://en.wikipedia.org/wiki/Color_temperature#Categorizing_different_lighting

There are more approximation formulas, some claim to be better, however these are not investigated.

Description

The library converts a temperature in Kelvin and a brightness (0..100%) to 3 numbers red, green and blue. These numbers are weights can be used to correct a colorimage for virtual white temperature.

There are 2 convert functions where the convert_NB() is claimed to be the more accurate one.

With the numbers R,G,B calculated one can convert images so they will look more like taken with candle light, sunrise or sunset etc.

pseudo code

Kelvin2RGB KRGB;

KRGB.convert(1850, 100);  // sunrise light factors

for each pixel in image
{
  red   *= KRGB.red();
  green *= KRGB.green();
  blue  *= KRGB.blue();
  drawPixel();
}

The numbers can also be used to reduce the blue channel so it has less effect on "getting sleepy".

The library uses floats for the R,G and B weights to keep values as accurate as possible. Especially with images with more than 8 bits per channel this is preferred. That said it is also possible to use this on a 565 image or to adjust color lookup tables.

Interface

The interface is pretty straightforward.

• Kelvin2RGB() constructor
• begin() for now an empty function.
• convert_TH(float temperature, float brightness = 100)
  temperature = 0..65500 temperature below 1000 is not well defined.
  brightness = 0..100%,
• convert_NB(float temperature, float brightness = 100)
  temperature = 0..65500 temperature below 1000 is not well defined.
  brightness = 0..100%, Is a bit more accurate and slightly slower (few %). Read link above for more information.
• float temperature() returns temperature, to check the value used.
• float brightness() returns brightness, to check the value used.
• float red() returns red channel weight 0.0 .. 1.0
  note this is different from Helland / Bartlett who both use an integer value 0 .. 255
• float green() returns green channel weight 0.0 .. 1.0
• float blue() returns blue channel weight 0.0 .. 1.0
• uint32_t RGB() retuns a 24 bit RGB value,
  more efficient than 3 floats for communication.

future

• separate brightness per color channel to mimic "artificial illumination"
• define constants like candleLight as parameter.

Operations

See examples