Journal de bord
Team leader: Yanis Lesné
Two sub-teams -> Matlab and Zemax
Day 1 : Monday
Bibliography: study of different phase masks
-> Yanis, Yilin, Raul, Shutian, Xiaolin
Search for references for the infrared sensor
-> Romain, Gabriel
Brainstorming: overall vision of project progress
-> Romain, Gabriel, Yanis, Yilin, Raul, Shutian, Xiaolin
Day 2 : Tuesday
Optical design on Zemax :
-> Romain, Yanis, Gabriel
We have started to design the optical system, starting with 1 lens in Germanium working with one wavelength. Correction of the spherical aberrations of the 3rd order on the axis. For an object field at 2°, the system is limited by diffraction for N = 1.5.
Then to correct the off-axis aberrations, we add another lens and we have started to increase the field in the image plane and the aperture.
Filter on Matlab :
-> Raul, Yilin, Gabriel
We have simulated the imaging of an ideal object on the sensor plane, affected by light diffraction and noise during propagation, and used Wiener deconvolution to eliminate these imaging defects. Furthermore, we simulated the optimal form of the Wiener filter under defocus conditions to further optimize the phase mask.
Phase mask on Matlab :
-> Shutian, Xiaolin
We have simulated four different phase masks and integrated the deconvolution process for parameter optimization, aiming to determine the optimal phase conditions.
We choose to work with an exponential phase mask.
Day 3: Wednesday
Everybody : We choose the scene and the noise we will use. It will be a photon noise modelised by a Gaussian law and for the scene, we will use 2 scenes : a black and white street image and an image with high and low frequencies.
Filter and mask matlab:
-> Shutian, Xiaolin, Raul, Yilin
We added the standard of optimization MSE and optimized the parameters of the phase mask. We successfully deconvolute the noisy image to a clear one.
Optical design on Zemax :
-> Romain, Yanis, Gabriel
We started again from an optical system with only 1 lens to increase step by step the field and the aperture by adding more lenses. We achieved the optimization of the optical system to the maximum FOV = 6 mm. We added the protection of the sensor.
Day 4: Thursday
Everybody : We work on the presentation and the Powerpoint.
Optical design on Zemax :
-> Romain, Yanis, Gabriel
We increased the aperture to N = 1.37 (aperture value = 11) thanks to the aspherisation of the 2 surfaces of the first lens and with conditions on the distance between the 2 lenses and between the second lens and the protection of the sensor.For the second week : look at materials in order to correct chromatic aberrations (Zinc sulfide or Zinc selenide ?).
Phase mask matlab:
-> Shutian, Xiaolin, Raul, Yilin
We corrected some errors and worked on the exponential phase mask optimization (There are still some issues in the optimization). We created a graph with the influence of different parameters of the phase mask to demonstrate that our results are valid.
Day 5 : Friday
Presentation training + feedbacks
Week 2 (16-20 December)
Day 1 : Monday
Brainstorming: overall vision of project progress
=> We continue what we were doing the first week.
Optical design on Zemax :
-> Romain, Yanis, Gabriel
We continued to work on our previous design, trying to maximize the field and the aperture. We separated the work in two groups depending on where we put the phase mask in the design (in the front or inside).
We tried to do a bit of tolerance and discovered that our system was very sensitive (too much).
So we went to the library and found a new interesting optical design to test in Handbooks of optic by Gross, that should be more resilient. Then, we separated again in two groups, one working on it and the other on a doublet.
Exponential phase mask encoding:
-> Raul
Encoding the exponential phase mask in C++ and trying to compile it into .dll to make it usable in ZEMAX. Exploration of numerous way to do it (different application, compilator, etc) but it doesn’t work yet.
Matlab:
-> Ylin, Shutian, Xiaolin
We modified the calculation method for the Enhanced Phase Mask (EPM) and recalculated the values of α and β. We successfully generated the EPM mask image, but we found that the coefficients of the cubic phase mask were incorrect. Tomorrow, we plan to explore another optimization method.
Zemax : We finished to work on a three lenses optical system and we tolerance it. The optical system is not robust for irregularities on surfaces.
We started to design a two lenses optical system.
We made detailed specifications.
Exponential phase mask encoding:
-> Raul
Encoding the exponential phase mask in C++ to make it usable in Zemax. The code is now compiling properly and seems to be working. The shape of the phase mask surface changes according to the mask parameters, and it is possible to observe alterations in the optical system's behavior (wavefront map, spot diagram, etc.). However, further tests are still needed to verify if it behaves as expected.
-> Shutian Yilin Xiaolin
I am studying how to use Matlab to obtain the optimal coefficients of the cubic phase mask, and how to output the MTF image in Matlab to compare the effects of the phase mask.
Day 3 : Wednesday
Optical design on Zemax :
-> Romain, Yanis, Gabriel
We put a phase mask on the 3 lenses optical system on Zemax. We compare the cubic and exponential phase mask.We try to do a 2 lenses system but it's quite hard to increase the field with only 2 lenses.
-> Raul
-> Raul
I corrected some small bugs in the exponential phase mask code, and now it's working perfectly. I (with Romain, Yanis, and Gabriel) analyzed how both phase masks (cubic and exponential) perform in our optical system using some optimal parameters provided by MATLAB simulations (made by Ylin, Shutian and Xiaolin).
-> Shutian Yilin Xiaolin
Matlab simulation to generate correct expenantial Phase Mask, and results are great
Day 4 : Thursday.
-> Yanis, Gabriel
Optical design : We have taken some data for the tolerancing of the optical system with the cubic phase mask and we started to take data for the optical system with the exponential phase mask
Report : we started to build the report
Image processing:
-> Raul, Romain
We deconvolute the image of the street using the PSF given by Zemax and an average Wiener filter. We need now to use more different PSF for the average filter and to center the PSF.
->Shutian Yilin Xiaolin
We take zernike parameters into account, to have slight better results of MTF in zemax, which means the MTF of different object distances are combined together when spatial frequences increases.
Day 5 :Friday.
->Shutian Yilin Xiaolin
We finally generate correct "centered" OTFs and then get correct average winer filter.
-> Raul, Romain
We center the PSF and compute an average filter with 9 different PSF for the cubic phase mask. We start to compare the MSE of the blurred image and the recovered image and also depending of the defocus.