Machine Vision

cancel
Showing results for 
Search instead for 
Did you mean: 

choice of camera ?

Hello ,
I read that somewhere: the defect To be detected, and therefore visible in the image, must at least fit into a pixel. But it is sometimes necessary to use two or even four, because of the background or the environment. It remains to calculate the minimum definition (D) of the sensor thanks to the formula:
D = (field of view x Number of pixels for a defect) / size of the smallest detectable defect
my question how am I going to define the numbers of pixels for a default
PS: I am on a project of acquisition and image processing, for the acquisition and image processing on small distances of (5cm), and in order to detect a  displacement (deformation) order of 1 micron on an image I would like to know what type of camera on resolution I have to choose sensor on a size marking (about 5X4 Cm), so I must calculet the resolution for the choice of my camera the best adapt thank you for your help ^^
0 Kudos
Message 1 of 2
(2,376 Views)

Just to get a resolution of 1 um, your field of view would be 50,000 x 40,000 pixels.  No camera is going to have that kind of resolution.  You need to rethink your application.  You would either need to look at a smaller area, or detect larger movement.

 

If you have an image with high contrast feature, you might be able to detect sub-pixel motion by detecting changes in grayscale at the edges of the features.  Perhaps you could use a 10,000 x 8,000 pixel camera or even a 5000 x 4000 pixel camera, but it would be a very challenging application.

 

Can you explain your project a little more so we might be able to make better suggestions?

 

Bruce

Bruce Ammons
Ammons Engineering
Message 2 of 2
(2,351 Views)