r0oland / fringeman Goto Github PK

I will take care of that, you can worry about how to get the correct phase map (instead of just taking the difference)

• load 2*4 images (dist and original phase shifted images)
• calc. phase map
• calc difference
  voila

camera images show artifacts, probably due to wrong configuration (fps or similar settings)

correct Gamma values of input images, otherwise output is very noisy and has steps

see:

• A fast and accurate gamma correction based on Fourier spectrum analysis for digital fringe projection profilometry:
  http://www.sciencedirect.com/science/article/pii/S0030401811012818
• Gamma Correction for Digital Fringe Projection Profilometry:
  http://www.opticsinfobase.org/ao/abstract.cfm?uri=ao-43-14-2906

take care of normalization of the unwrapped phase map, match it to U8 or U16 to have dynamic range completely covered

to ensure that recorded image matches created fringe completely in one direction (shorted side of the monitor)

Use of integers causes loss of precision, use float instead and normalize to a max. value of 255 to use in the images

• take images of mirror in focus, then screen in focus
• show tilt disalignment
• show fringes

match size of computer generated undistorted fringe images to that of the recorded distorded images

computer generated image size = screen resolution = screen_y_resolution * screen_x_resolution
recorded image size = camera resolution = camera_y_resolution * camera_x_resolution

size of actual useful part containing the fringes (due to round mirror):
y_resolution*y_resolution, where screen_y_resolution ≠ camera_resolution

-> only take useful (i.e. square part) and match the size, I guess scale the smaller one up...

• create correct phase map size from size of input image/array not from global variable screen size

• better detection of discontinuities
• highly cool but complicated method based on wavelet analysis:
  http://jnaiam.org/uploads/files/Paper5.pdf
• easy and almost ready to use (see vis in phase map folder)
  Edge Detection with 2D Convolution
  Edge Detection Example (based on IMAQ edge detection vi)