Color Vision Mechanism

Color Vision and Camera

How we see color 

This post is about the color vision and camera. If you read my earlier post you would know that the photoreceptors in the retina convert the light rays (electromagnetic radiation) to biologically usable electro-chemical signals within neurons. There are two types of photoreceptors involved in vision. They are

  1. The rods and
  2. The cones

The rods and cones share the basic anatomical structure but they vary in their ability to process light. The rods are very sensitive to light and can even record one photon (light particle). There are about 120 million rod cells in a human eye. The rods are involved in scotopic vision (low light vision) and lack the ability to process color. Therefore, when the light level is very low, we have difficulty distinguishing colors. The rods contain a protein called opsin which when bound by retinal forms a pigment molecule called rhodopsin. Loss of rods or defects in rods leads to night blindness.

The cones are involved in color vision. The cones are less sensitive than rods to photons and require brighter light levels to be activated (photopic vision). The cones are essential for visual acuity and resolution. The cones are densely situated at a point directly behind the lens on the retina. This area is called the fovea or fovea centralis. The concentration of cones is highest at the fovea centralis. There are around 6 million cones in the retina. The cones contain different types of opsins that combine with retinal to form three different types of the pigment photopsin. Loss of cones or defect in cone cells impairs color vision and leads to legal blindness.

eye and color vision

“Cone-fundamentals-with-srgb-spectrum” by BenRG – Own work. Licensed under Public Domain via Wikimedia Commons –

Color vision involves three different cone cells.The three types of cone cells differ in their sensitivity to short, medium and long wavelengths of light and hence are classified as S-cone, M-cone and L-cone. The S-cone cells are highly sensitive to light at 420nm (blue color); the M-cone cells are highly sensitive to light at around 534nm(green color); and the L-cone cells are highly sensitive to light at around 564nm(red color). It should be noted that the photoreceptor cells themselves do NOT see the different colors, but only the different wavelengths and intensities of light. It should also be noted that the Red, Green and Blue colors do not correspond accurately to the wavelengths indicated above; the Red-Green-Blue color model is mainly for convenience in describing the color processing by the three types of cone cells.

Color vision

Distribution of Cone cells in the fovea. Mark Fairchild [CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0) via Wikimedia Commons

The S-, M- and L- cones absorb light from a range of wavelengths. Therefore, there can be an overlap in the wavelengths that are absorbed by the three types of cones. The S-cone is sensitive to a range of wavelength from 400-500nm, but peak sensitivity is at 420 nm. Likewise, the range for M-cone is from 450-630nm (peak sensitivity at 534nm) and the range for L-cones is from 500-700nm (peak sensitivity at 564nm).

The S-, M- and L-cone cells record not only the different wavelengths of light, but also the number of photons (intensity of light) that hit them. The light energy is then converted in these cells to electro-chemical stimuli that pass on the information to bipolar cells, which are connected to the cone cells. The bipolar cells convey the information to the retinal ganglion cells that will process and organize the information before relaying it to the brain through the optic nerve. Further, processing of signals occur in the visual cortex and other parts of the cerebral cortex where the brain compares the signals that originated from each cone cell to arrive at the color image. The electro-chemical impulse arising from each cone cell and its neighbors is compared by the brain during the processing of color to ultimately visualize the color image.

The digital sensor (how the camera sees colors)

Similar to our eyes, the light particle (photon) reflected from an object enters the lens where it is bent due to refraction and focuses on a recording surface at the back of the camera. This recording surface is a digital camera is called a sensor.

digital sensor

The Bayer Filter; image courtesy of Wikimedia Commons

The digital sensor contains millions of tiny light-gathering slots called pixels (picture elements) that record the photons that strike them. The pixels have photodiodes that record the small electric current produced when the photons hit the sensor. The pixels can only measure the amount or intensity of light (the number of photons) impinging on them. Therefore, the image will only be monochromatic (no color will be recorded).

To obtain a color image, the digital cameras use a filter called a Bayer filter that is placed in front of the sensor. The Bayer filter (image above) contains an array squares consisting of the three primary colors, red, green and blue. Therefore, when reflected light passes through the Bayer filter and hits the sensor, each pixel in the sensor will absorb red, green or blue color. Consider for example, a red rose. When the reflected light from the red rose hits the sensor, only the red wavelength is absorbed and all other wavelengths of light are filtered by the Bayer filter in front of the sensor.

On the Bayer array, there are twice as many green filters as red or blue. This is done to simulate the human color vision, as the eye is most sensitive to green color. The on-board computer in the camera will then employ a process called demosaicing, which is the use of sophisticated algorithms, to interpolate colors of neighboring pixels to form a color image. To avoid color artifacts arising from this interpolation of RGB values, the digital sensors also employ an anti-aliasing filter to improve color accuracy.

There is another type of digital sensor called the Foveon X3 sensor, which is used by the Sigma company. The Foveon sensor is a silicon wafer in which three photo-diodes are stacked in a vertical arrangement. The different wavelengths of light penetrate the silicon wafer at different levels. The photo-diode for blue (short wavelength) is at the top, followed by the photo-diodes for green and red respectively. Therefore, each site is capable of recording the three primary colors. Sigma claims that the use of this sensor results in increased resolution in the images. This is because each photo-site is capable of capturing the three primary colors and hence there is no interpolation and anti-aliasing involved with the Foveon sensor.

Further reading

https://faculty.washington.edu/chudler/eyecol.html

http://webvision.med.utah.edu/book/part-vii-color-vision/color-vision/

Leave a Reply

Your email address will not be published.

Time limit is exhausted. Please reload CAPTCHA.