
Re: Generating Point Clouds
Posted:
Mar 14, 2011 1:03 PM


Image Analyst,
This is my prog on isolating laser stripe from the set of images. My result for this script gives me the isolated laser stripe as an image. But, I want point cloud in 3D space. Help me out.
source pics:
https://picasaweb.google.com/lh/sredir?uname=v.navaneethakannan&target=ALBUM&id=5579922441378566865&authkey=Gv1sRgCNCYsYik8c7obQ&feat=email
Script:
clear; clc; jpgFiles = dir('*.jpg'); for k = 1:length(jpgFiles) filename = jpgFiles(k).name; data1 = imread(filename); %imshow(data1); [p1,q1]=find(250<=data1(:,:,1)); [p2,q2]=find(250<=data1(:,:,2)); [p3,q3]=find(250<=data1(:,:,2)); r1=[p1 q1]; r2=[p2 q2]; r3=[p3 q3]; s1=intersect(r1,r2,'rows'); s2=intersect(s1,r3,'rows'); t1=setdiff(r1,s2,'rows');
b=zeros(size(data1)); u1=mean(t1,1); u2=std(t1,1)*0.6; ul=round(u1+u2); ll=round(u1u2); t1(t1(:,1)>ul(1),:)=[]; t1(t1(:,1)<ll(1),:)=[]; for i=1:length(t1) b(t1(i,1),t1(i,2),1)=data1(t1(i,1),t1(i,2),1); b(t1(i,1),t1(i,2),2)=data1(t1(i,1),t1(i,2),2); b(t1(i,1),t1(i,2),3)=data1(t1(i,1),t1(i,2),3); end %figure %imshow(b) c=im2bw(b); c=uint8(c); data1(:,:,1)=data1(:,:,1).*c; data1(:,:,2)=data1(:,:,2).*c; data1(:,:,3)=data1(:,:,3).*c; %figure(1),subplot(4,8,k),subimage(data1); imshow(data1) end;
"Navaneetha Kannan Viswanathan" <v.navaneethakannan@gmail.com> wrote in message <ilh0lo$90r$1@fred.mathworks.com>... > Hi ImageAnalyst, > > I am working as per your suggestions and now I am struck in generating a reliable method of turning the x,y location of laser in my image into distance measurement. pls help me out. > > Thanks in advance > > ImageAnalyst <imageanalyst@mailinator.com> wrote in message <270ad23756e34f2683f9af4e9e7f3167@22g2000prx.googlegroups.com>... > > I suppose but I don't think you'd do that. First you'll have to > > develop the algorithm where you turn the x,y location of the laser in > > the image into a distance measurement. After that, you probably won't > > do a point cloud but a full 3D model, at least if you want to do it > > like the pros do it: > > > > http://www.canfieldsci.com/imaging_systems/facial_systems/VECTRA_M3_Imaging_System.html > > > > http://www.3dmd.com/

