How can I find the spatial calibration factor?

Asked by Tomas on 8 Dec 2012

Latest activity Commented on by Image Analyst on 8 Oct 2016

100 views (last 30 days)

Hi, According to this link http://www.azurs.net/photoblog/resolution-taille-pixels-image.html real heigth of the image (mm)= image heigth of the image (pixel)/image resolution (ppp)

So the spatial calibration factor can be find by this formula??

calibration factor=real heigth of the image (mm)/image heigth (pixel)

thaks in advance

0 Comments

3 Answers

Link

Answer by Image Analyst on 10 Dec 2012

Edited by Image Analyst on 10 Dec 2012

The calibration factor is (the actual "true" height of a known object) / (the height in pixels of that same object in your image).

Then when you measure any other distances, multiply by that factor. See demo code below:

% Code to spatially calibrate and image.
% Code asks user to draw a line and then to specify the length
% of the line in real world units.  It then calculates a spatial calibration factor.
% User can then draw lines and have them reported in real world units.
% Take out the next two lines if you're transferring this to your program.
close all;  % Close all figures (except those of imtool.)
clear;  % Erase all existing variables.
clc;    % Clear the command window.
workspace;  % Make sure the workspace panel is showing.
format longg;
format compact;
fontSize = 20;
% Check that user has the Image Processing Toolbox installed.
hasIPT = license('test', 'image_toolbox');
if ~hasIPT
  % User does not have the toolbox installed.
  message = sprintf('Sorry, but you do not seem to have the Image Processing Toolbox.\nDo you want to try to continue anyway?');
  reply = questdlg(message, 'Toolbox missing', 'Yes', 'No', 'Yes');
  if strcmpi(reply, 'No')
    % User said No, so exit.
    return;
  end
end
% Read in a standard MATLAB gray scale demo image.
folder = fullfile(matlabroot, '\toolbox\images\imdemos');
baseFileName = 'cameraman.tif';
% Get the full filename, with path prepended.
fullFileName = fullfile(folder, baseFileName);
% Check if file exists.
if ~exist(fullFileName, 'file')
  % File doesn't exist -- didn't find it there.  Check the search path for it.
  fullFileName = baseFileName; % No path this time.
  if ~exist(fullFileName, 'file')
    % Still didn't find it.  Alert user.
    errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
    uiwait(warndlg(errorMessage));
    return;
  end
end
grayImage = imread(fullFileName);
% Get the dimensions of the image.  
% numberOfColorBands should be = 1.
[rows columns numberOfColorBands] = size(grayImage);
% Display the original gray scale image.
subplot(2, 1, 1);
imshow(grayImage, []);
title('Original Grayscale Image', 'FontSize', fontSize);
% Enlarge figure to full screen.
set(gcf, 'units','normalized','outerposition',[0 0 1 1]);
% Give a name to the title bar.
set(gcf,'name','Demo by ImageAnalyst','numbertitle','off') 
% Initialize
units = 'pixels';
spatialCalibration = 1.0;
button = 1;
while button ~= 3
  % Get which action the user wants to do.
  button = menu('Choose an action', 'Calibrate', 'Measure', 'Exit');
  if button == 3
    % Bail out because they clicked Exit.
    break;
  end
  % Make caption the instructions.
  subplot(2, 1, 1);
  title('Left-click first point.  Right click last point.', 'FontSize', fontSize);
  % Ask user to plot a line.
  [x, y, profile] = improfile();
  % Restore caption.
  title('Original Grayscale Image', 'FontSize', fontSize);
  % Calculate distance
  distanceInPixels = sqrt((x(1)-x(end))^2 + (y(1)-y(end))^2);
  % Plot it.
  subplot(2,1,2);
  plot(profile);
  grid on;
  
  % Initialize
  realWorldNumericalValue = distanceInPixels;
  caption = sprintf('Intensity Profile Along Line\nThe distance = %f pixels', ...
    distanceInPixels);
  title(caption, 'FontSize', fontSize);
  ylabel('Gray Level', 'FontSize', fontSize);
  xlabel('Pixels Along Line', 'FontSize', fontSize);
  if button == 1
    % They want to calibrate.
    % Ask user for a number.
    userPrompts = {'Enter true size:','Enter units:'};
    defaultValues = {'180', 'cm'};
    titleBar = 'Enter known distance';
    caUserInput = inputdlg(userPrompts, titleBar, 2, defaultValues);
    if isempty(caUserInput),return,end; % Bail out if they clicked Cancel.
    
    % Initialize.
    realWorldNumericalValue = str2double(caUserInput{1});
    units = char(caUserInput{2});
    
    % Check for a valid integer.
    if isnan(realWorldNumericalValue)
      % They didn't enter a number.  
      % They clicked Cancel, or entered a character, symbols, or something else not allowed.
      message = sprintf('I said it had to be an number.\nI will use %d and continue.', distanceInPixels);
      uiwait(warndlg(message));
      realWorldNumericalValue = distanceInPixels;
      units = 'pixels';
      spatialCalibration = 1.0;
%       continue; % Skip to end of loop.
    end
    spatialCalibration = realWorldNumericalValue / distanceInPixels;
  end
  realWorldDistance = distanceInPixels * spatialCalibration;
  caption = sprintf('Intensity Profile Along Line\nThe distance = %f pixels = %f %s', ...
    distanceInPixels, realWorldDistance, units);
  title(caption, 'FontSize', fontSize);
  ylabel('Gray Level', 'FontSize', fontSize);
  xlabel('Pixels Along Line', 'FontSize', fontSize);
end

15 Comments

Show 12 older comments

Image Analyst on 16 Dec 2012

You can do that but I don't think it's more common. It's called binning, and it's where you combine adjacent pixels. It might depend on the camera sensor architecture. I just had a case where I wanted to reduce the resolution of a camera to get a higher frame rate so the users could position an object in front of the camera with no delay/lag. When set up with a lower resolution, it took the central portion of the sensor, so the field of view changed drastically. When I asked for the same field of view and use subsampling (say every 4th pixel) the engineer thought at first he could do it and then said the chip architecture wouldn't allow it but he could do binning. However with binning the image had to be monochrome because it combined adjacent pixels which are different colors (remember it couldn't do subsampling). And this was with a CMOS sensor which are supposed to be more addressable than CCD sensors. Anyway, neither of these was what I wanted so I just went with a lower resolution camera because the higher frame rate was really what was needed more than the resolution.

Catherine Perna on 22 May 2014

Hi there,

I am new to matlab code and I was wondering if it was possible to use this code (or something slightly different) to find the spatial calibration of a video? I have a moving object over a grid and I am trying to determine the speed of that object. I have most of the code worked out but struggling to determine the number of pixels/mm of my video, can anyone help?

Thanks

MOHIT on 19 Jul 2016

Suppose, I have this image and I want to measure the area of the table surrounded by a red line. This table is rectangular but not looking as a rectangle. Here we can't take pixel values uniform how can we measure the area of this table?

Link

Answer by Catherine Perna on 22 May 2014

Hi there,

Thanks

3 Comments

Image Analyst on 22 May 2014

Again, you usually have to have an object of known length in the scene that you can measure. You must know the grid spacing. And you can find the spacing in pixels, so that's your spatial calibration. If you watch a certain grid line and know how far it moved, then you can get the relative speed of the camera.

Hazim Nasir on 20 Aug 2016

Hi All I have this zigzag line and need to measure the total length of it,

 the length law L=sqrt((x(1)-x(2))^2+(y(1)-y(2))^2) can not used straight forward because the complicated shape of the line

Image Analyst on 20 Aug 2016

You should start your own question. But basically since that is a single wide binary image of a line, you can use sum to get the number of pixels in the line.

LineLengthInPixels = sum(binaryImage(:));

If you want Euclidean distance, so a two pixel long line at an angle of 45 degrees has a length of sqrt(2) instead of 2, you can do that with the sqrt() function and knowledge of where the vertices in your line are.

Link

Answer by Aiswarya M D on 7 Oct 2016

How to adapt different retinal image resolutions using spatial calibration with a field of view of 45°

1 Comment

Image Analyst on 8 Oct 2016

spatial_calibration_demo.m

I don't know what that means so all I can suggest is that you use my spatial calibration code to calibrate from known things. Perhaps your ophthalmoscope will tell you what the spatial resolution is.

You are now following this question