%------------------------------------------------------------------------ %'phys_XYZ':transforms image (px) to real world (phys) coordinates using geometric calibration parameters % function [Xphys,Yphys,Zphys]=phys_XYZ(Calib,X,Y,Zindex) % %OUTPUT: % Xphys,Yphys,Zphys: vector of phys coordinates corresponding to the input vector of image coordinates %INPUT: % Calib: Matlab structure containing the calibration parameters (pinhole camera model, see % http://servforge.legi.grenoble-inp.fr/projects/soft-uvmat/wiki/UvmatHelp#GeometryCalib) and the % parameters describing the illumination plane(s) % .Tx_Ty_Tz: translation (3 phys coordinates) defining the origine of the camera frame % .R : rotation matrix from phys to camera frame % .fx_fy: focal length along each direction of the image % X, Y: vectors of X and Y image coordinates % ZIndex: index defining the current illumination plane in a volume scan %======================================================================= % Copyright 2008-2024, LEGI UMR 5519 / CNRS UGA G-INP, Grenoble, France % http://www.legi.grenoble-inp.fr % Joel.Sommeria - Joel.Sommeria (A) univ-grenoble-alpes.fr % % This file is part of the toolbox UVMAT. % % UVMAT is free software; you can redistribute it and/or modify % it under the terms of the GNU General Public License as published % by the Free Software Foundation; either version 2 of the license, % or (at your option) any later version. % % UVMAT is distributed in the hope that it will be useful, % but WITHOUT ANY WARRANTY; without even the implied warranty of % MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the % GNU General Public License (see LICENSE.txt) for more details. %======================================================================= function [Xphys,Yphys,Zphys]=phys_XYZ(Calib,Slice,X,Y,Zindex) %------------------------------------------------------------------------ testangle=0;% =1 if the illumination plane is tilted with respect to the horizontal plane Xphys Yphys test_refraction=0;% =1 if the considered points are viewed through an horizontal interface (located at z=Calib.InterfaceCoord(3)') Zphys=0; %default output if isempty(Slice) Slice=Calib;%old convention < 2022 end if exist('Zindex','var')&& isequal(Zindex,round(Zindex))&& Zindex>0 && isfield(Slice,'SliceCoord')&&size(Slice.SliceCoord,1)>=Zindex if isfield(Slice, 'SliceAngle') && size(Slice.SliceAngle,1)>=Zindex && ~isequal(Slice.SliceAngle(Zindex,:),[0 0 0]) testangle=1; norm_plane=angle2normal(Slice.SliceAngle(Zindex,:));% coordinates UVMAT-httpsof the unit vector normal to the current illumination plane end Z0=Slice.SliceCoord(Zindex,3);%horizontal plane z=cte Z0virt=Z0; if isfield(Slice,'InterfaceCoord') && isfield(Slice,'RefractionIndex') H=Slice.InterfaceCoord(3);% z position of the water surface if H>Z0 Z0virt=H-(H-Z0)/Slice.RefractionIndex; %corrected z (virtual object) test_refraction=1; end end else Z0=0; Z0virt=0; end if ~exist('X','var')||~exist('Y','var') Xphys=[]; Yphys=[];%default return end %coordinate transform if ~isfield(Calib,'fx_fy') Calib.fx_fy=[1 1]; end if ~isfield(Calib,'Tx_Ty_Tz') Calib.Tx_Ty_Tz=[0 0 1]; end if ~isfield(Calib,'Cx_Cy') Calib.Cx_Cy=[0 0]; end kc=[0 0]; if isfield(Calib,'kc') kc(1:numel(Calib.kc))=Calib.kc; end % if ~isfield(Calib,'kc2') % Calib.kc2=0; % end if isfield(Calib,'R') R=(Calib.R)'; c=Z0virt; cvirt=Z0virt; if testangle % equation of the illumination plane: z=ax+by+c a=-norm_plane(1)/norm_plane(3); b=-norm_plane(2)/norm_plane(3); if test_refraction avirt=a/Slice.RefractionIndex; bvirt=b/Slice.RefractionIndex; else avirt=a; bvirt=b; end cvirt=Z0virt-avirt*Slice.SliceCoord(Zindex,1)-bvirt*Slice.SliceCoord(Zindex,2);% Z0 = (virtual) z coordinate on the rotation axis (assumed horizontal) % c=z coordinate at (x,y)=(0,0) c=Z0-a*Slice.SliceCoord(Zindex,1)-b*Slice.SliceCoord(Zindex,2); R(1)=R(1)+avirt*R(3); R(2)=R(2)+bvirt*R(3); R(4)=R(4)+avirt*R(6); R(5)=R(5)+bvirt*R(6); R(7)=R(7)+avirt*R(9); R(8)=R(8)+bvirt*R(9); end Tx=Calib.Tx_Ty_Tz(1); Ty=Calib.Tx_Ty_Tz(2); Tz=Calib.Tx_Ty_Tz(3); Dx=R(5)*R(7)-R(4)*R(8); Dy=R(1)*R(8)-R(2)*R(7); D0=(R(2)*R(4)-R(1)*R(5)); Z11=R(6)*R(8)-R(5)*R(9); Z12=R(2)*R(9)-R(3)*R(8); Z21=R(4)*R(9)-R(6)*R(7); Z22=R(3)*R(7)-R(1)*R(9); Zx0=R(3)*R(5)-R(2)*R(6); Zy0=R(1)*R(6)-R(3)*R(4); B11=R(8)*Ty-R(5)*Tz+Z11*cvirt; B12=R(2)*Tz-R(8)*Tx+Z12*cvirt; B21=-R(7)*Ty+R(4)*Tz+Z21*cvirt; B22=-R(1)*Tz+R(7)*Tx+Z22*cvirt; X0=R(5)*Tx-R(2)*Ty+Zx0*cvirt; Y0=-R(4)*Tx+R(1)*Ty+Zy0*cvirt; %px to camera: Xd=(X-Calib.Cx_Cy(1))/Calib.fx_fy(1); % sensor coordinates Yd=(Y-Calib.Cx_Cy(2))/Calib.fx_fy(2); dist_fact=1+kc(1)*(Xd.*Xd+Yd.*Yd);% distortion factor, first approximation Xu,Yu=Xd,Yd test=0; niter=0; while test==0 && niter<10 dist_fact_old=dist_fact; Xu=Xd./dist_fact;%undistorted sensor coordinates, second iteration Yu=Yd./dist_fact; R2=Xu.*Xu+Yu.*Yu; dist_fact=1+kc(1)*R2+kc(2)*R2.*R2;% distortion factor,next approximation test=max(max(abs(dist_fact-dist_fact_old)))<0.00001; % reducing the relative error to 10^-5 forthe inversion of the quadraticcorrection niter=niter+1; end denom=Dx*Xu+Dy*Yu+D0; Xphys=(B11.*Xu+B12.*Yu+X0)./denom;%world coordinates Yphys=(B21.*Xu+B22.*Yu+Y0)./denom; if testangle Zphys=c+a*Xphys+b*Yphys; else Zphys=Z0; end else Xphys=-Calib.Tx_Ty_Tz(1)+X/Calib.fx_fy(1); Yphys=-Calib.Tx_Ty_Tz(2)+Y/Calib.fx_fy(2); end