Removing image's initial linear background with linearfit
1. Open the image that you want to remove the background. Usually, is the resulting image you got after preprocess / integration.
2. Execute the script SCRIPT -> Image Analysis -> Noise Evaluation (CFA Bayer)
3. We are going to select the lower noise channel as a base image for our linear fit, based in the bayer pattern of our color camera. In this example, the camera is an ASI 2600 MC Pro, and its bayer pattern is RGGB.
As per the few little documentation I found of the script, this would be the relation of the results with our bayer pattern.
Note: RGGB is a common bayer pattern usually used for color cameras, like Canon DSLR as well. If you don't know which is the bayer pattern corresponding to your camera, you just have to do a google search with the model and the keywords "bayer pattern".
The lower noise channel is the channel 0 (R) = 1.458e-01.
Note: In the past, I have had problems executing the Noise Evaluation Script for CFA images, often thrown an error. As a workaround, I did split first the channels (see step 4) and evaluated each channel individually with the script
SCRIPT -> Image Analysis -> Noise Evaluation.
4. Extract the image RGB channels clicking on the toolbar button Split RGB Channels. Three new images, each corresponding to a RGB channel will appear.
5. Select the menu PROCESS -> Color Calibration -> Linear Fit.
6. Click on the right icon on the field Reference Image and select the view that we identified in the step 2 as the lower noise channel image and then click the OK button. In our case, we will select the R image channel.
Fitting the other channels
7. Next, we are going to fit the remaining 2 channels (B,G) with the (R) channel. Simply drag the lower left triangle over the desired channel or click the target image to select it and click on the square channel to apply.
8. Once you finish fitting the remaining 2 channels (B,G), select PROCESS -> ChannelManagement -> ChannelCombination. Make sure RGB option is selected in the Color Space options group.
9. Click on the image selector icon on the right of each one of the three Channels / Source image fields and select the corresponding image channel. Make sure you match it with the last letter of the name name of the image.
10. Click on the Apply Global button (F6 or the blue circle icon in the lower left corner of the window). The new corrected image will be created.
11. Click on the STF Autostretch button on the toolbar to see the results.
Some remarks / information
- I made LinearFit my personal preferred choice to execute after the image integration was done. Some people goes directly with the DBE/ABE tools, but for me this has worked better because if I see this remove well the background, I can skip using DBE / ABE background removal tools.
- Some times the image will still have gradients even after executing LinearFit, at this moment you can continue applying your preferred background removal process/tool.