Objective: to identify the image, and automatically on the moire fringe
As shown in figure, basically be to see the right, up and down on the right side of the stripe is not fully aligned, I need to do is to use the image recognition to calculate offset,

Question: look at the literature on a need [to extract phase information of the upper and lower], then the [] of phase difference of two images, the phase difference to calculate distance,
But do not know how to extract phase information of the image and,
Ask for advice, thank you!

CodePudding user response:

Pick up and down two parts RIO, projection on the x axis, to calculate again

CodePudding user response:

reference 1st floor sevancheng response:

select fluctuation two parts RIO, projection on the x axis, then calculate

What do you mean to do projection

CodePudding user response:

And I use matlab to do the Fourier transform to view their respective phase spectrum didn't completely understand,,, feel not the same meaning,,,

CodePudding user response:

With reference to this https://blog.csdn.net/mangobar/article/details/80286059

According to the x coordinate for peace

CodePudding user response:

reference 4 floor sevancheng response:

reference here https://blog.csdn.net/mangobar/article/details/80286059

According to the x coordinate sum

I think he comes out projection should be similar to the amplitude, but two value image of phase I still wouldn't _ (: з) _
<)Their trends can be fitting with a trigonometric function?
Phase difference distance is obtained only by two peak feeling?

CodePudding user response:

Choose n peaks, the ideal condition, ((P21 + + P23 P22) - (P11 + P12 + P13))/3, ask to the pixel

CodePudding user response:

refer to 6th floor sevancheng response:

choose n peaks, ideal condition, ((P21 + + P23 P22) - (P11 + P12 + P13))/3, ask to the pixel difference

Emmmm if or directly on the image measuring distance to calculate words I don't need to do the projection directly measured average fluctuation two groups of stripe center spacing is ok,
Because such precision is not enough that I want to by Fourier transform demodulation in frequency domain and then separate the phase component to calculate phase difference,