Need Help from Microscopy Experts


Status
Not open for further replies.

richardg

New Member
Sep 3, 2006
330
0
0
Hi Experts,

Anyone ever took pics of 15 micron objects?


Need some help.
Thanks
Richard
 

Hi Experts,

Anyone ever took pics of 15 micron objects?


Need some help.
Thanks
Richard
Using microscope? I used to do that in my previous job too. We can discuss it here too. ;p
 

Need help/advise on technique or help and equipment to take the pics? :) Also, what kind of resolution you need? If just to see the object, then light microscopes would do, but if you need details, then electron microscopy might be needed.
 

I'm about to lay hands on a fullframe 4Megapixel BnW CCD with colourfilters, fluorescence imaging capabilities, camera mounted on a Zeiss Axiotron2 with Apo Chromat Objectives. Few more parts to be delivered to my lab and..viola! But duno what nice stuff to snap, as my lab doesn't do bio stuff..15microns can be imaged decently with light microscopes. If 3D, can use confocal microscopes.

SEM can define better,...but no colour!
 

I'm about to lay hands on a fullframe 4Megapixel BnW CCD with colourfilters, fluorescence imaging capabilities, camera mounted on a Zeiss Axiotron2 with Apo Chromat Objectives. Few more parts to be delivered to my lab and..viola! But duno what nice stuff to snap, as my lab doesn't do bio stuff..15microns can be imaged decently with light microscopes. If 3D, can use confocal microscopes.

SEM can define better,...but no colour!
What a waste of capabilities. ;p
 

wow! :)
looks like we have some experts with nice toys too :)

have been tasked :(
to do automated location of a 15 micron object.
no need details, just need to locate.

looking at FOV of like 0.2mm

The main design issue is that the whole contraption has to fit into 6 inch x 6 inch x 6 inch. :eek:

Looking like need a optical systems design engineer... :(

thanks in advance for any tips...

cheers
 

What's your timeline, and is this a one-off item, research project, or a product?

If it is one-off and for whatever industry your're in, why not contact one of the universities (esp. NTU) or polytechnics and discuss with them? Sounds like at least two final-year projects (one hardware, and one software for image processing/recognition). ;)
 


I would say it it a very tough job and require few engineers to do it. Nothing to do with photography. But any way, here is my advice

The cheapest way I could think is you buy a microscopy and then mould it with a camera. Hmm. I am sure the microscopy vendor would supply you such cheapo system.

To identify the location of a object the cheapest and easiest way is through Image processing, which based on the contrast of pixels to identify the object location. A simple programmer would be able to do it. So you only need to know a little about optical microscopy with visible light. They cost only about 5-10k ( if I am not wrong). FOV 2mx2mm is achieveable. But to locate the object of 15um it would be hard since the scanning area is too much larger compared to object. But it is still a doable. But it is the simplest and cheapest liao.

But after identify the location, what would you do since it is an automatic system, you might need do alignment ..Hmm..there it comes the expensive system.

Here is some system that do "searching" and "aligning". They work different mechanism because they scan lasers on the suface and detect the reflected beam to locate the object and measure its size.

1) To determine the location of any object in an area, you need a laser scanner system (microwave, acoustic, Xray are beyond of my knowledge). Frankly, I tell you..The system would be very expensive (my knowledge also is limited).

2) If you only need to identify how many "objects" within such area, then may be "cheap" system would do KLA-Tencor. It do scanner, identify the location,count number of object . The price also about S$900k. You could check with them for cheaper version. There are might other companies but I don't know. I am not working for Tencor so I wont benefit from this advice :D.

3)But if you need to identify accurately the location and do alignment. You need very very expensive scanner. In the microlectronic industry , lithography requires much more accurate dimension but there scan area is also limited (less than your FOV 2mmx2mm).There are only 2 players which could produce such system are Canon and Nikon, proudly. The cost is about S$11m per machine. So if you intend to buy this system, be prepare for the huge budget.


4)Important thing your budget lor. Time is an important factor you must consider too.
 

I would say it it a very tough job and require few engineers to do it. Nothing to do with photography. But any way, here is my advice

The cheapest way I could think is you buy a microscopy and then mould it with a camera. Hmm. I am sure the microscopy vendor would supply you such cheapo system.

To identify the location of a object the cheapest and easiest way is through Image processing, which based on the contrast of pixels to identify the object location. A simple programmer would be able to do it. So you only need to know a little about optical microscopy with visible light. They cost only about 5-10k ( if I am not wrong). FOV 2mx2mm is achieveable. But to locate the object of 15um it would be hard since the scanning area is too much larger compared to object. But it is still a doable. But it is the simplest and cheapest liao.

But after identify the location, what would you do since it is an automatic system, you might need do alignment ..Hmm..there it comes the expensive system.

Here is some system that do "searching" and "aligning". They work different mechanism because they scan lasers on the suface and detect the reflected beam to locate the object and measure its size.

1) To determine the location of any object in an area, you need a laser scanner system (microwave, acoustic, Xray are beyond of my knowledge). Frankly, I tell you..The system would be very expensive (my knowledge also is limited).

2) If you only need to identify how many "objects" within such area, then may be "cheap" system would do KLA-Tencor. It do scanner, identify the location,count number of object . The price also about S$900k. You could check with them for cheaper version. There are might other companies but I don't know. I am not working for Tencor so I wont benefit from this advice :D.

3)But if you need to identify accurately the location and do alignment. You need very very expensive scanner. In the microlectronic industry , lithography requires much more accurate dimension but there scan area is also limited (less than your FOV 2mmx2mm).There are only 2 players which could produce such system are Canon and Nikon, proudly. The cost is about S$11m per machine. So if you intend to buy this system, be prepare for the huge budget.


4)Important thing your budget lor. Time is an important factor you must consider too.

much useful info, thanks!
 

What's your timeline, and is this a one-off item, research project, or a product?

If it is one-off and for whatever industry your're in, why not contact one of the universities (esp. NTU) or polytechnics and discuss with them? Sounds like at least two final-year projects (one hardware, and one software for image processing/recognition). ;)

thanks for the suggestion..
unfortunately.. proposal including budget was due last year...
cant wait for de final years...
thanks!
 

wow! :)
looks like we have some experts with nice toys too :)

have been tasked :(
to do automated location of a 15 micron object.
no need details, just need to locate.

looking at FOV of like 0.2mm

The main design issue is that the whole contraption has to fit into 6 inch x 6 inch x 6 inch. :eek:

Looking like need a optical systems design engineer... :(

thanks in advance for any tips...

cheers
Can I rephrase your problem to be like this?
You need to locate a 15um object within a 0.2mmx0.2mm area? Or do you need to traverse to cover more area? If it's just simply a fixed FOV to cover 0.2mmx0.2mm, it will not be that difficult.

As for the 15um object, do you need some form of identification or that will be the only object in the FOV? If it's a certain feature you need to identify, then you will need an imaging system which can resolve enough details for you to perform the identification.

For 200um x 200um area, you can easily resolve to better than 1um with a 640x480 resolution camera. The rest is just digital image processing and feature extraction to give you the coordinates over your field of view.
 

Oh..my mistake. Sorry, I thought the area was 2mmx2mm. If the area is only 0.2mmx0.2mm which mean 200umx200um. Your object is about 15um. Then as others said, simple optical microscopy would do.You might need a ~50-100x microscopy. Hope my previous post can only be refered as a lousy referrence.
 

So now you need to:

1. Get a camera system to take a photo of a 0.2mm x 0.2mm area with enough resolution to capture a 15µm feature
2. Create an interface between this imaging system to whatever system you need to couple it to
3. Write a programe to control the camera: movement (if necessary), focusing (if necessary), and image recognition (your 15µm feature)

BC :think:
 

Can I rephrase your problem to be like this?
You need to locate a 15um object within a 0.2mmx0.2mm area? Or do you need to traverse to cover more area? If it's just simply a fixed FOV to cover 0.2mmx0.2mm, it will not be that difficult.

As for the 15um object, do you need some form of identification or that will be the only object in the FOV? If it's a certain feature you need to identify, then you will need an imaging system which can resolve enough details for you to perform the identification.

For 200um x 200um area, you can easily resolve to better than 1um with a 640x480 resolution camera. The rest is just digital image processing and feature extraction to give you the coordinates over your field of view.

yupz u got it.

software is our forte so ok there.
the main issue is the size :( , has to fit within a confined space like 6inch cube
most microscope objectives are too long...
also need d coax light :(
light path nids a few 90 degrees turns ...

thanks!
 

yupz u got it.

software is our forte so ok there.
the main issue is the size :( , has to fit within a confined space like 6inch cube
most microscope objectives are too long...
also need d coax light :(
light path nids a few 90 degrees turns ...

thanks!
I think the 6"x6"x6" should not be much of a problem. You should be able to get some shorter objectives and mount it straight to a CCD camera. You don't need a whole microscope. The coax light may be accomodated by one prism.

Request for a catalog from Edmund Optics and see if you can find the parts and try to fit into a 6"x6"x6" enclosure. You may also try the smaller CCD cameras which uses the webcam type of thread (I can't remember what it's called), those can save you quite a bit of space because you need less magnification and the optics are shorter also.

For the illumination, see if LEDs are sufficient, if they are then you don't need a halogen lamp and that will save you a lot of space. Also, you can see if using LEDs around the objective aligned to focus over your region of interest is able to meet your illumination demands. If so, then you save yourself the prism for coax lighting also.
 

Status
Not open for further replies.