where y is a constant—usually near 1.8—chosen to compensate for the nonlinear light output from the monitor.
The field-star method of white balance depends on the validity of the assumption that the average color of the brightest stars in the background is white.
Experience shows that the shortcut method works fairly well 95% of the time. Forcing the same number of bright-star pixels to become saturated usually produces images with a fair approximation of the proper color balance.
The field-star method breaks down for images containing young star clusters (because these contain large numbers of hot young blue stars), and for strongly colored objects such as bright emission nebulae, where the coloration affects the histogram endpoint pixel value.
20.4.1 Summary: White-Balance Using Field Stars
It is possible to construct "reasonably accurate" color images based on the normally justified assumption that the average color of the field stars in the image is white or nearly white. If this assumption is false, the color balance will not be accurate.
1. Take filtered image sets. You may use any exposure times for the filtered images. Take appropriate dark frames and flat fields. Calibrate and stack to produce a high-quality image of the subject through each filter.
2. If a sky background gradient is present in the filtered images, correct the gradient. One of AIP4Witfs Gradient Tools can do this for you.
3. Register the three filtered images to sub-pixel accuracy. AlP4Wiris Process RGB Tool can carry out this operation.
4. Create a color image from the three corrected color channels. You can use AIP4Witfs Join Colors Tool, with an automatic color balance endpoint that yields pleasing color results, to join the color channels into a single color image.
Color balancing with field stars is considerably easier than color balancing with solar-class G2V stars. However, although images color-balanced with field stars could in theory have color as accurate as images balanced with G2V stars, there is no guarantee that the "average color is white" assumption will be true for a particular image.
Was this article helpful?