Line scan sensors - has anybody tried this for macro?

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

nathanm
Posts: 222
Joined: Thu Jun 02, 2016 8:13 pm
Location: Bellevue, WA
Contact:

Line scan sensors - has anybody tried this for macro?

Post by nathanm »

In other threads I and others have been discussing the SK 85mm f/4.5 macro varon lens. It is an example of a modern macro lens designed to be used with high resolution line sensors.

The simplest line sensors are a single row of pixels, but there are fancier ones that have 3 or 4 rows so they can do color. The highest resolution line sensors these days have 16K pixels, which are typically 5 microns across, but they exist in all sizes.

Here are some examples

http://teledynedalsa.com/imaging/produc ... iranha-xl/

They are used for machine vision systems - an example is quality control inspection of manufactured circuit boards. The line sensors are cost effective because they are MUCH cheaper than a rectangular sensor of the same size. In these systems the subject to be photographed is mechanically scanned past the line sensor.

Well it finally dawned on me that instead of buying lenses made for line sensors to use with my camera sensor (for me these days PhaseOne 100 mpixel), I could also use one of these line sensors to make a scanning stacking camera.

This would mean moving the camera very accurately across the subject. But the requirements do not seem very difficult - on a single line camera, it would mean moving by one pixel width, which is 5 microns. Many of us have built stacking systems that have much more resolution (in depth).

With a multiline camera, it would be a multiple of that.

In the past there have been scanning backs for large format cameras like this. The Seitz 6x17 digital does something like this:

http://www.roundshot.com/xml_1/internet ... 2/f123.cfm

The difference is that this would be made for stacking, so it would be stepping in two axes - the depth axis as for normal stacking, and the side-to-side axis.

Obviously it would only work for static subjects, but stacking is already like that. One would need to figure out a way to focus and compose efficiently.

Has anybody on the forum tried something like that?

Comments? Suggestions?
nathanm

Sager
Posts: 52
Joined: Fri Jan 09, 2015 8:52 pm
Contact:

Post by Sager »

I have a few linescan camera's that I was thinking of experimenting with. Need to get the right capture cards. I think the ones I have are 12K monochromatic high-speed sensors. They came with Macro-Varon's I purchased a year or so ago.

Did you ever try anything out?

ray_parkhurst
Posts: 3416
Joined: Sat Nov 20, 2010 10:40 am
Location: Santa Clara, CA, USA
Contact:

Post by ray_parkhurst »

I suppose with smaller step size you could get a higher resolution in one axis. I'm not sure how far you could push it but you can certainly improve over a 5um resolution.

nathanm
Posts: 222
Joined: Thu Jun 02, 2016 8:13 pm
Location: Bellevue, WA
Contact:

Post by nathanm »

I looked into this in some detail.

Line scan cameras are made to sync with a moving conveyor belt. So one would either need to build a motorized stage for the subject, or a motorized stage for the camera. I was thinking of doing it with the camera.

Many line scan cameras don't work well at low speed, but some such as certain models of the Dalsa Piranha series http://www.teledynedalsa.com/imaging/pr ... =Line+Scan can ramp down to whatever speed you need all the way to dead stopped.. With one of those cameras you could get higher res in the scan direction.

Normally they do with fast (electronic) shutter speed, continuous light and continuous motion. But you could instead move the sensor with a stepper.

It is critical to have smooth and/or accurate motion in the scan direction. That is why I thought I would move the camera - it won't wiggle the way a macro subject might.

The next problem is that most color line scan cameras have 4 or more rows of sensors with something like a Bayer array of filters. Unfortunately this can cause parallax issues - these are hard to explain, but the key is that line scan cameras are generally set up to be fixed focus distance for a flat subject - such as a circuit board, or an LED display. A subject with varying height will get what looks like chromatic aberration if there is variation in height across the rows of pixels.

In principle one might be able to correct this in the stacking software, but it winds up mixing the demosaicing of the Bayer array with stacking in a complicated mix.

This is annoying in many line scan applications - so much so that some companies have color cameras by having 3 monochrome line scan sensors ganged together with prisms and dichroic filters. But none of those cameras have great resolution so what's the point.

I thought about building my own set of dichroic prisms, but that is starting to get rather exciting from a design/build standpoint.

You could also do 3 scans with a monochrome line scan and a filter wheel, but that gets more time consuming.

There is an odd digital panoramic camera called the Seitz 617
https://luminous-landscape.com/setiz-6x17-digital/ that essentially is a 6 x 17 format film camera converted to take a Dalsa line scan sensor. So it is certainly possible.

The bottom line is that if one wanted to make a super high resolution macro camera (16K pixels in the short direction, with scans as long as you like) then this is a possibility, but it is a lot of work.

The reason that I started this thread is that there are so many great line scan lenses.

Instead of this build, I would up buying one of those lenses - the super expensive Rayfact 2x-5x lens and I am making a camera for it which will use the PhaseOne 100 mpixel back. If I want a panorama I can scan with that. While that isn't as cheap as the quite reasonably priced line scan cameras, it is super high quality and I don't need to mess with all these other problems.

The lens is built to order (the US rep says I am the first customer in the US for this lens) - it should get here in another month.

Also, the camera that I am building for it is really complicated as it stands - just with all of the mechanical things I need to do. I will post on it when it is done. Using the line scan camera would be much the same, only way harder. So I will satisfy myself with that for a while.

Edit AdminCR: Link fixed
nathanm

kds315*
Posts: 224
Joined: Mon Feb 02, 2009 1:06 am
Contact:

Post by kds315* »

Looking forward to hearing about your camera Nathan!
Klaus

http://www.macrolenses.de for macro and special lens info
http://www.pbase.com/kds315/uv_photos for UV Images and lens/filter info
http://photographyoftheinvisibleworld.blogspot.com/ my UV diary

Post Reply Previous topicNext topic