I came across this method of calculating the DOF that involves pre-calculating the camera CoC.
The same method can be used for lineScan cameras?
What is the Coc of 16K linescan camera (Cmos Sensor 4x 16384 Pixels, 5 x 5μm) ?
I came across this method of calculating the DOF that involves pre-calculating the camera CoC.
The same method can be used for lineScan cameras?
What is the Coc of 16K linescan camera (Cmos Sensor 4x 16384 Pixels, 5 x 5μm) ?
The circle of confusion is what you accept as tolerable blur.
In the olden days of 35mm film photography, a common convention was to accept a 30µm diameter as "sharp enough". Translating this into the digital world, this roughly corresponds to a 1 megapixel resolution.
The DoF calculations have nothing to do with the medium that captures the image, only with the optical system that produces the image, and your acceptable circle of confusion. So yes, generic DoF formulas can be used for linescan sensors as well.
So, if you need the full resolution of your linescan camera (and your optics are good enough!), then use the single-pixel size of 5µm for your DoF calculations. If you can e.g. accept a degradation to effectively 4000 pixels (blur size = 4 pixels), then use 20µm.
The depth of field is dependent on the size of the output image, and the distance from which it is viewed.
The typical/accepted standard is a 0.2mm blur point in an 8x10 image and viewed so that it occupies the human's ~60˚ primary FoV... which is from a distance equivalent to the diagonal of the output image (12").
If you know the dimensions of the scan, the modern accepted/standard CoC is d/1500 where d is the diagonal of the original image (sensor) size.
E.g. for a 35mm FF sensor the diagonal is 43.2mm. 43.2/1500 = .0288 rounded to .03 (30um). Enlarging the 36mm sensor width to 10 inch requires ~ 7x. 7 x .03 = .21mm blur point in the final image. And when an 8x10 is viewed from the diagonal distance of ~12" it will occupy ~ 60˚ FOV.
You know the width 16364 x .005 = 81.92mm and the height is 4 x .005 = .02mm... all you need to determine is how many scans will be combined.
But you also need to decide if the standard CoC is suitable; it's actually a very low standard for resolution. d/3000 would be more suited for 20/20 corrected vision and high contrast. And if you want the maximum possible, then 2x the pixel pitch is a good choice to keep it below the diffraction limit of the sensor (some use 2.5x).