What image correction methods are available for Greateyes CCD cameras

1)        Dark current correction (Dark Current Correction）

Greateyes CCDEach row of pixels on the camera has several dark pixels (dummy pixels) at its sides. These pixels are shielded and do not respond to light; they are used to measure the number of dark current electrons accumulated during exposure, and the dark current electron counts of these dummy pixels are used to correct the dark noise differences among pixel rows.

2)        Bad pixel correction (Bad Pixel Correction）

For a small number of pixels with reduced sensitivity,grateyesinterpolation is used to correct these pixels, i.e., the count values of pixels surrounding the bad pixel are multiplied by an input coefficient to serve as the count value of the bad pixel.

3)        Bias correction (Background Correction）

The count values of pixels in a bias image come from two parts: one is dark current, and the other is ambient light (light present when there is no light source to be measured). If ambient light can be completely shielded, only dark current, which is related to exposure time, will produce counts. Obviously, the bias image changes over time. Therefore,Greateyesthe image acquisition control software provides two methods for acquiring bias images:

a.         Before acquiring multiple frames of images, acquire one bias frame with the same exposure time as the multiple frames to be acquired. This bias image is used as the bias for all images and is subtracted.

b.         When acquiring multiple frames of images, acquire a bias frame before each frame, with the same exposure time as the frame to be acquired. Each frame has its own bias image.

ais suitable for situations where the environment is relatively stable (ambient temperature is very stable, and ambient light has no significant changes);bis the opposite.

What is flat field correction

1)        Ideally, when there is no light illumination,CCDthe count value (gray value) of each pixel of the camera should be0but in reality it is not0and varies from pixel to pixel. The reasons include1). the presence of bias, i.e., even when the exposure time is0there are still count values;2).when the exposure time is not0there is dark current, and the dark current of each pixel is inconsistent. For consumerCCDcameras, because they operate at room temperature, this is particularly obvious. Scientific-gradeCCDcameras generally operate at low temperatures, so the effect of dark current is relatively small.

If a bias image with the same exposure time is measured before exposure and subtracted after exposure, this constitutes bias correction.

2)        Ideally,CCDwhen uniformly illuminated by light of a certain wavelength and uniform intensity, the count values (gray values) of each pixel in the obtained image should be equal, but in reality the values of each pixel often differ significantly. This is because the response of each pixel is inconsistent, i.e., the quantum efficiency of each pixel for light of the same wavelength is inconsistent.

3)        Non-uniformity caused by other factors. For example, if microlenses are mounted on the pixels, the non-uniformity produced by the microlenses.

That is, for light of a given wavelength, when the horizontal coordinatexis the number of incident photons and the vertical coordinateyis the number of photogenerated electrons,CCDthe actual response curve of a pixel is a straight line with a positive intercept (for light of a given wavelength, the response curve is theoretically a straight line), and the slope and intercept of this straight line differ from pixel to pixel.

The so-called flat field correction is to correct these inconsistencies. Flat field correction is to make the response lines of all pixels identical by changing the slope of the response line of each pixel (i.e., the signal gainGain) and the offset (i.e., the signal offsetOffset).

The most commonly used flat field correction method is the "two-point correction method," which assumes that the pixel response is linear (a straight line).

l  First, the camera performs an exposure on a dark field (i.e., no light illumination) to obtain the offset of each pixel (Offset); next, an exposure is performed under uniform illumination to obtain a uniform field image, preferably with the count values of all pixels close to the full well capacity;

l  finally, the dark field image is subtracted from the uniform light field image, and the average pixel count value of the subtracted image is divided by the count value of each pixel to obtain the gain (Gain) correction coefficient, and the correction coefficient of each pixel is saved.

l  Image correction: when actually acquiring images, first acquire a bias image with the same exposure time, then acquire the image, subtract the bias image, and multiply the count value of each pixel of the image after bias subtraction by the correction coefficient to obtain the corrected image.


The above content was compiled and organized by Jerry Huang of our company, for the purpose of knowledge sharing and mutual learning only, and may not be reproduced without permission.