290 likes | 462 Views
Measurements in Fluid Mechanics 058:180:001 (ME:5180:0001) Time & Location: 2:30P - 3:20P MWF 218 MLH Office Hours: 4:00P – 5:00P MWF 223B-5 HL. Instructor: Lichuan Gui lichuan-gui@uiowa.edu http://lcgui.net. Lecture 38. Stereo high-speed imaging. Stereo High-speed Imaging.
E N D
Measurements in Fluid Mechanics058:180:001 (ME:5180:0001)Time & Location: 2:30P - 3:20P MWF 218 MLHOffice Hours: 4:00P – 5:00P MWF 223B-5 HL Instructor: Lichuan Gui lichuan-gui@uiowa.edu http://lcgui.net
Stereo High-speed Imaging - High-speed camera Gray value resolution 10-bit CMOS sensor 17.5µm pixels Frame rate & digital resolution 10241024 @ 2000 fps 1024256 @ 8000 fps
Stereo High-speed Imaging - Optical configuration outside the camera Mirror image Mirror Lens Front view Alate Side view Mirror Mirror image
Lens adaptor Block Image sensor (1024256) Slit apertures Mirrors Stereo High-speed Imaging - Optical configuration inside the camera Lens Mirror images
High-speed camera Mirrors Back lighting Tethered fire ant alate Microphones Stereo High-speed Imaging - Experimental setup with sound recording
Stereo High-speed Imaging - Experimental setup with sound recording Sound recorded from bottom & rear with MCDL
Raw image Background Processed Stereo High-speed Imaging - Sample image & postprocessing
Stereo High-speed Imaging - Synchronize motion w. sound in one period 70 frames in the period of 8.8 ms (114Hz), bar chart represents sound pressure
Stereo High-speed Imaging - Improved system for wing motion reconstruction • New optic table and magnetic holders for higher precision • New template for adjusting mirror angles to ensure 90o • difference between front and side- views - Tested ant body carefully oriented
O: root of the wing T: tip of the wing 3: the 3rd point at the wing surface Stereo High-speed Imaging - Position & orientation of the wing • the wing assumed to be a planar surface without thickness
Stereo High-speed Imaging - Wing root position More frames may be necessary for higher accuracy & reliability
Axis's Axis of the wing: OT Axis of ant body: OZ Planes Pink: wing surface Angles Yellow: OT & OZ xy: wing angle in xy-plane zy: wing angle in yz-plane : angle between wing surface & plane through OT & OZ Stereo High-speed Imaging - 3D view of the wing surface
Wing surface function: OT&OZ surface function: Wing rotation angle: Stereo High-speed Imaging - Data reduction equations
Stereo High-speed Imaging - Example I: BIFA male test on July 8, 2006 Body weight: 5.5 mg Body length: 6 mm Image size: 590190 pixels, digital resolution: 22.35 pixel/mm
Stereo High-speed Imaging - Wing tip position of the BIFA male
17 Stereo High-speed Imaging - Wing angles of the BIFA male
Stereo High-speed Imaging - High resolution picture of the forewing Image size: 850×300 pixels Digital resolution: 128 pixel/mm O T
H L Stereo High-speed Imaging - High resolution picture of the hindwing Image size: 850×300 pixels Digital resolution: 128 pixel/mm O T Gray value distribution: G(L,H)
Front view image: gf(x,y)=G(L,H) H Side view image: (x,y,z) gs(z,y)=G(L,H) L Top view image: gt(x,z)=G(L,H) Stereo High-speed Imaging - Simulating images form three view angles
Stereo High-speed Imaging - Original (top) and simulated (bottom) images
Stereo High-speed Imaging - Reconstructed 3D view of the male
23 Stereo High-speed Imaging - Overlapped image of 74 frames
Stereo High-speed Imaging • Reference • Gui L, Fink T, Cao Z, Sun D, Seiner JM and Streett DA (2010) Fire ant alate wingmotion data and numerical reconstruction. Journal of Insect Science 10:9,available online: http://insectscience.org/10.19
Correlation Interrogation & FFT Acceleration % Computer correlation function with FFT gfft1=fft2(g1); % FFT of g1 gfft2=conj(fft2(g2)); % complex conjugate of FFT for g2 C=real(ifft2(gfft1.*gfft2)); % correlation coefficient for i=1:M for j=1:N m(i,j)=i-M/2; % horizontal coordinate in correlation plane n(i,j)=j-N/2; % vertical coordinate in correlation plane end end for i=1:M/2 % periodical reconstruction of correlation function C1(i,1:N)=C(i+M/2,1:N); end for i=M/2+1:M C1(i,1:N)=C(i-M/2,1:N); end for j=1:N/2 C2(1:M,j)=C1(1:M,j+N/2); end for j=N/2+1:N C2(1:M,j)=C1(1:M,j-N/2); end C=C2; % normalize and display correlation function Cmax=max(max(C)); Cmin=min(min(C)); D=(C-Cmin)./(Cmax-Cmin)*245+5; imshow(int8(D)) clear; % input image file A=imread('image01.bmp'); % gray value distribution of the image % nx=124; ny=124; for x=1:nx for y=1:ny G(x,y)=A(ny-y+1,x); end end % select evaluation point x1=nx/2; y1=ny/2; % determine interrogation window size M=64; N=64; % select evaluation sample for i=1:M for j=1:N g(i,j)=G(i+x1-M/2,j+y1-N/2); end end g1=g; g2=g;
Correlation Interrogation & FFT Acceleration C Fast computation of evaluation function C D Acceleration with radix-2 based FFT algorithm - Periodical reconstruction of the correlation function A B g1(i,j) g2(i,j) % Computer correlation function with FFT gfft1=fft2(g1); % FFT of g1 gfft2=conj(fft2(g2)); % complex conjugate of FFT for g2 C=real(ifft2(gfft1.*gfft2)); % correlation coefficient
Correlation Interrogation & FFT Acceleration C Fast computation of evaluation function C D Acceleration with radix-2 based FFT algorithm - Periodical reconstruction of the correlation function A B g1(i,j) g2(i,j) C D C1 A B % periodical reconstruction of correlation function for i=1:M/2 C1(i,1:N)=C(i+M/2,1:N); End for i=M/2+1:M C1(i,1:N)=C(i-M/2,1:N); end
Correlation Interrogation & FFT Acceleration C1 Fast computation of evaluation function Acceleration with radix-2 based FFT algorithm - Periodical reconstruction of the correlation function g1(i,j) g2(i,j) C C D D C2 A A B B % periodical reconstruction of correlation function for j=1:N/2 C2(1:M,j)=C1(1:M,j+N/2); end for j=N/2+1:N C2(1:M,j)=C1(1:M,j-N/2); end
Correlation Interrogation & FFT Acceleration C2 Fast computation of evaluation function Acceleration with radix-2 based FFT algorithm - Periodical reconstruction of the correlation function g1(i,j) g2(i,j) C C D D C A A B B % periodical reconstruction of correlation function C=C2; for i=1:M for j=1:N m(i,j)=i-M/2; % horizontal coordinate in correlation plane n(i,j)=j-N/2; % vertical coordinate in correlation plane end end