Imaging simulations

[meller92]

Hi, I don't know Python a lot but i think that adding something like this to the zernike.py module could be useful:

import opticspy
import numpy as np
from PIL import Image
from scipy import signal
import matplotlib.pyplot as plt

Z = opticspy.zernike.Coefficient(Z1=0.2)
Z.zernikesurface()
psfmatrix = Z.psf(matrix=True)
Z.mtf()
abspsf=abs(psfmatrix/psfmatrix.max())

im = Image.open("img.bmp")
pic = np.array(im)
imgplot=plt.imshow(pic)
imgplot.set_cmap("gray")
plt.colorbar()
plt.show()

abspic=(pic/pic.max())
outpic=signal.fftconvolve(abspsf,abspic,mode="same")
outpic=np.trunc(255*abs(outpic/outpic.max()))
imgplot=plt.imshow(outpic)
imgplot.set_cmap("gray")
plt.colorbar()
plt.show()

[Sterncat]

Hi meller,

Thank you!

I will test it and add it in the module! Cool!

Marvin

[dguerra]

Hi Marvin, any improvement on this? I think it could be a nice feature. Thanks

[brandondube]

Hi @dguerra, you may be interested in prysm, an alternative python optics module. Among its features are image simulation, with respect for units and a lack of necessity for the PSF and image to have equal spacing. I do need to polish the code for that type of thing more, but you can find an example of using image simulation here.

[goldengrape]

failed in
psfmatrix = Z.psf(matrix=True)
return
TypeError: psf() got an unexpected keyword argument 'matrix'

I dive into the code, for .psf

def psf(self,r=1,lambda_1=632*10**(-9),z=0.1):
		"""
		------------------------------------------------
		psf()
		Return the point spread function of a wavefront described by
		Zernike Polynomials
		------------------------------------------------
		Input:
		r: exit pupil radius(mm)
		lambda_1: wavelength(m)
		z: exit pupil to image plane distance(m)
		"""
		print(r,lambda_1,z)
		PSF = self.__psfcaculator__(r=r,lambda_1=lambda_1,z=z)
		fig = __plt__.figure(figsize=(9, 6), dpi=80)
		__plt__.imshow(abs(PSF),cmap=__cm__.RdYlGn)
		__plt__.colorbar()
		__plt__.show()
		return 0

Z.psf() always return 0, not PSF from psfcaculator

[Sterncat]

Maybe one int 0 get into the calculation.