Machine Vision

cancel
Showing results for 
Search instead for 
Did you mean: 

Detect pixel saturation

I am trying to take an image and easily detect the pixels that are saturated within the image.  So far I have extracted a single plane or band from the image and then converted it into an array.  One question I have is why does the display show only the middle half of the image alongthe x axis? Secondly, I can't figure out how to get it to display just the image based on its dimensions.  The x and y values that it automatically detects are not the image dimensions (1912 x 1076)and I can't figure out what they are representing.  Third, is there a way to change the displays color ramp from something other than shades of blue?

 

Thanks in advance!

0 Kudos
Message 1 of 4
(4,073 Views)

Hi,

 

Is there a reason why you are not using an image display (from the Vision palette)?  With that, the image will display as a grayscale.  

 

Take a look at the attachment.  Is this more what you are looking for?

Message 2 of 4
(4,053 Views)

I wasn't sure what display to use for what I was trying to do.  I want to be able to specify the ranges of values to be displayed; say 0-100 as black, 101-200 as grey, and 201-255 as white for example so that the saturated or near saturated pixels will be easily detected telling me I need to adjust the integration times on the camera.  I'll play around with what you did and see if it does what I'm looking for. 

 

Thanks for the response!

0 Kudos
Message 3 of 4
(4,049 Views)

After using the vi I originally posted, it seems that the "Saturated" display is mixing up the x and y values, so it displays the width as being 1076 pixels wide instead of 1912.  If I can figure out how to swap the x and y input values, then this way of displaying the images will do.  Using a rotate function rotated the image but it didn't change the display function properties.  It would be nice also if I had a little more control on the colors used in the color ramp ie. red for the red band, green for the green band, and blue for the blue band.

 

Any thoughts?

0 Kudos
Message 4 of 4
(4,037 Views)