Machine Vision

cancel
Showing results for 
Search instead for 
Did you mean: 

find the value of grayscale

Hello Guys.

I'm a newbie in here and also a newbie in labview stuff. I'm currently working on my final project and face a handicap in image processing. I need to get characteristic extraction as a variable in artificial neural networks. Does anyone knows how to get characteristic extraction from image grayscale image? To find its characteristic of energy, contrast, homogenity, entropy, level of grayscale, and standard deviation.

Thanks in advance

0 Kudos
Message 1 of 7
(5,094 Views)

@kak_iguuun wrote:

Hello Guys.

I'm a newbie in here and also a newbie in labview stuff. I'm currently working on my final project and face a handicap in image processing. I need to get characteristic extraction as a variable in artificial neural networks. Does anyone knows how to get characteristic extraction from image grayscale image? To find its characteristic of energy, contrast, homogenity, entropy, level of grayscale, and standard deviation.

Thanks in advance



I am not sure about what you mean by energy. You can try imaq histogram function. that will give you gray scale value and standard deviation.

0 Kudos
Message 2 of 7
(5,070 Views)

Hello,

 

to my knowledge there is no Labview function that can perform what you want directly. You would have to code it yourself.

 

Whenever I am in doubt and need some extra info about algorithms regarding the image processing I tend to check Matlab documentation first (the also cite different papers). In my opinion no other SW environment comes close to the number of implemented algorithms in image processing (there are also a lot of users/programers that implement all kinds of algorithms). So:

 

http://www.mathworks.com/help/images/ref/graycoprops.html

 

http://www.mathworks.com/help/images/ref/entropy.html

 

You would need to implement the algorithms yourself in Labview or create a .dll in Matlab, that can be called in Labview (have never done this). Assuming you have Matlab. If not, your only option is the former.

 

Best regards,

K


https://decibel.ni.com/content/blogs/kl3m3n



"Kudos: Users may give one another Kudos on the forums for posts that they found particularly helpful or insightful."
0 Kudos
Message 3 of 7
(5,065 Views)

Hi, as muks advice you, using imaq histogram you can find some data . If you send some details about the desired algorithm may be i can hshare you some idea.

0 Kudos
Message 4 of 7
(4,934 Views)

Sorry, I missed this when you posted it a couple of weeks ago.  Have you looked at IMAQ Cooccurrence Matrix?  There are a number of Haralick features you can evaluate, including entropy, contrast, homogeneity, energy and so on.  As well as giving you the GLCM.  There might be other useful routines under Image Processing/Texture.

 

The VI help is really poor (not even listing the features available) but in the Vision Concepts manual, there is a good explanation of the features (though the equations all state "Entropy" -- if someone official could flag this as a CAR that would be good!!).

0 Kudos
Message 5 of 7
(4,919 Views)

While I'm at it with the errors in the Concepts Manual, the previous page on What To Expect also has errors, particularly in the section on Scale Variation which talks a couple of times about rotations of 10 degrees - presumably should be scale changes of 10%.  Add that to the CAR please!

0 Kudos
Message 6 of 7
(4,918 Views)

@GregSands wrote:

Sorry, I missed this when you posted it a couple of weeks ago.  Have you looked at IMAQ Cooccurrence Matrix?  There are a number of Haralick features you can evaluate, including entropy, contrast, homogeneity, energy and so on.  As well as giving you the GLCM.  There might be other useful routines under Image Processing/Texture.

 

The VI help is really poor (not even listing the features available) but in the Vision Concepts manual, there is a good explanation of the features (though the equations all state "Entropy" -- if someone official could flag this as a CAR that would be good!!).


Nice one! Was not even aware of this function!

 

Thanks for pointing this out!

 

Best regards,

K


https://decibel.ni.com/content/blogs/kl3m3n



"Kudos: Users may give one another Kudos on the forums for posts that they found particularly helpful or insightful."
0 Kudos
Message 7 of 7
(4,906 Views)