Definition:
A(i, j) = 1 is a midpoint of a cross if the elements
A(i-1, j) = 1
A(i+1, j) = 1
A(i, j+1) = 1
A(i, j-1) = 1.
Together the elements and the midpoint form a cross in a matrix A, where A is at least a 3-by-3 matrix and i, j ∈ ℕ\{0}.
Suppose the image above is the 8-by-8 matrix A with natural numbers 1, 2, 3 ... as elements. From this definition the matrix has a total of 3 crosses. The crosses have their midpoints on A(2,2), A(5, 4) and A(5, 5).
What I want to do is write a function that finds the number of crosses in the matrix A. I have an idea but I'm not sure it's the most optimal one. Here's the pseudocode for it:
ITERATE FROM row 2 TO row 7
ITERATE FROM column 1 TO column 8
IF current element contains 1
INCREMENT xcount by 1
IF xcount >= 3
CHECK IF counted 1:s is part of a cross
ELSE IF xcount IS NOT 0
SET xcount to 0
The idea is to iterate through every column from row 2 to row 7. If I find 3 consecutive 1:s on the same row I immediately check if the 1:s belongs to a cross. This should work, but imagine having a very large matrix A - how efficient would this code be in that situation? Couldn't this problem be solved using vector notation?
Any answer is very much appreciated. Thanks in advance!
Not near matlab at the moment, but this is what I'd do. Assuming A is binary (has only 0'a and 1's):
crs=[0 1 0 ; 1 1 1 ; 0 1 0]; % a minimal "cross" filter
C=conv2(A,crs./sum(crs(:)),'same'); % convolve A with it
[x y]=find(C>0.9); % find x,y positions of the crosses by looking
% for peak values of C
so you basically convolve with a "minimal" (normalized) cross (crs) and look for peaks using max. x and y are the coordinates of your cross positions. No need to use for loops, just the built in (and pretty fast) 2d convolution, and the max function.
The threshold condition C>0.9, is just to illustrate that there's need to be a threshold that is weighted by intensity of crs. In this case I have normalized crs in the colvolution line (crs/sum(crs(:))) so if A is a binary matrix as in the example, you find that the convolution of the minimal normalized cross will leave the value of the pixel where the cross is at 1, whereas other pixels will be less than 1 (that's why I arbitrarily chose 0.9) . So you can replace the threshold to C==1, if it's always a binary.
Another way to visulize the position of the cross is just to look at C.*(C==1). This will generate a matrix the size of A with 1s only where the crosses were...
EDIT:
For maximal speed, you may consider writing it as a one liner, for example:
[x y]=find(conv2(A,[0 1 0 ; 1 1 1 ; 0 1 0]./5,'same')==1);
Using bit masks:
ux = [false(size(A,1),1) (A(:,3:end) & A(:,2:end-1) & A(:,1:end-2)) false(size(A,1),1)]
uy = [false(1,size(A,2)); (A(3:end,:) & A(2:end-1,:) & A(1:end-2,:)); false(1, size(A,2))]
u = ux & uy
[x y] = find(u)
Related
I have two images which one of them is the Original image and the second one is Transformed image.
I have to find out how many degrees Transformed image was rotated using 3x3 transformation matrix. Plus, I need to find how far translated from origin.
Both images are grayscaled and held in matrix variables. Their sizes are same [350 500].
I have found a few lecture notes like this.
Lecture notes say that I should use the following matrix formula for rotation:
For translation matrix the formula is given:
Everything is good. But there are two problems:
I could not imagine how to implement the formulas using MATLAB.
The formulas are shaped to find x',y' values but I already have got x,x',y,y' values. I need to find rotation angle (theta) and tx and ty.
I want to know the equivailence of x, x', y, y' in the the matrix.
I have got the following code:
rotationMatrix = [ cos(theta) sin(theta) 0 ; ...
-sin(theta) cos(theta) 0 ; ...
0 0 1];
translationMatrix = [ 1 0 tx; ...
0 1 ty; ...
0 0 1];
But as you can see, tx, ty, theta variables are not defined before used. How can I calculate theta, tx and ty?
PS: It is forbidden to use Image Processing Toolbox functions.
This is essentially a homography recovery problem. What you are doing is given co-ordinates in one image and the corresponding co-ordinates in the other image, you are trying to recover the combined translation and rotation matrix that was used to warp the points from the one image to the other.
You can essentially combine the rotation and translation into a single matrix by multiplying the two matrices together. Multiplying is simply compositing the two operations together. You would this get:
H = [cos(theta) -sin(theta) tx]
[sin(theta) cos(theta) ty]
[ 0 0 1]
The idea behind this is to find the parameters by minimizing the error through least squares between each pair of points.
Basically, what you want to find is the following relationship:
xi_after = H*xi_before
H is the combined rotation and translation matrix required to map the co-ordinates from the one image to the other. H is also a 3 x 3 matrix, and knowing that the lower right entry (row 3, column 3) is 1, it makes things easier. Also, assuming that your points are in the augmented co-ordinate system, we essentially want to find this relationship for each pair of co-ordinates from the first image (x_i, y_i) to the other (x_i', y_i'):
[p_i*x_i'] [h11 h12 h13] [x_i]
[p_i*y_i'] = [h21 h22 h23] * [y_i]
[ p_i ] [h31 h32 1 ] [ 1 ]
The scale of p_i is to account for homography scaling and vanishing points. Let's perform a matrix-vector multiplication of this equation. We can ignore the 3rd element as it isn't useful to us (for now):
p_i*x_i' = h11*x_i + h12*y_i + h13
p_i*y_i' = h21*x_i + h22*y_i + h23
Now let's take a look at the 3rd element. We know that p_i = h31*x_i + h32*y_i + 1. As such, substituting p_i into each of the equations, and rearranging to solve for x_i' and y_i', we thus get:
x_i' = h11*x_i + h12*y_i + h13 - h31*x_i*x_i' - h32*y_i*x_i'
y_i' = h21*x_i + h22*y_i + h23 - h31*x_i*y_i' - h32*y_i*y_i'
What you have here now are two equations for each unique pair of points. What we can do now is build an over-determined system of equations. Take each pair and build two equations out of them. You will then put it into matrix form, i.e.:
Ah = b
A would be a matrix of coefficients that were built from each set of equations using the co-ordinates from the first image, b would be each pair of points for the second image and h would be the parameters you are solving for. Ultimately, you are finally solving this linear system of equations reformulated in matrix form:
You would solve for the vector h which can be performed through least squares. In MATLAB, you can do this via:
h = A \ b;
A sidenote for you: If the movement between images is truly just a rotation and translation, then h31 and h32 will both be zero after we solve for the parameters. However, I always like to be thorough and so I will solve for h31 and h32 anyway.
NB: This method will only work if you have at least 4 unique pairs of points. Because there are 8 parameters to solve for, and there are 2 equations per point, A must have at least a rank of 8 in order for the system to be consistent (if you want to throw in some linear algebra terminology in the loop). You will not be able to solve this problem if you have less than 4 points.
If you want some MATLAB code, let's assume that your points are stored in sourcePoints and targetPoints. sourcePoints are from the first image and targetPoints are for the second image. Obviously, there should be the same number of points between both images. It is assumed that both sourcePoints and targetPoints are stored as M x 2 matrices. The first columns contain your x co-ordinates while the second columns contain your y co-ordinates.
numPoints = size(sourcePoints, 1);
%// Cast data to double to be sure
sourcePoints = double(sourcePoints);
targetPoints = double(targetPoints);
%//Extract relevant data
xSource = sourcePoints(:,1);
ySource = sourcePoints(:,2);
xTarget = targetPoints(:,1);
yTarget = targetPoints(:,2);
%//Create helper vectors
vec0 = zeros(numPoints, 1);
vec1 = ones(numPoints, 1);
xSourcexTarget = -xSource.*xTarget;
ySourcexTarget = -ySource.*xTarget;
xSourceyTarget = -xSource.*yTarget;
ySourceyTarget = -ySource.*yTarget;
%//Build matrix
A = [xSource ySource vec1 vec0 vec0 vec0 xSourcexTarget ySourcexTarget; ...
vec0 vec0 vec0 xSource ySource vec1 xSourceyTarget ySourceyTarget];
%//Build RHS vector
b = [xTarget; yTarget];
%//Solve homography by least squares
h = A \ b;
%// Reshape to a 3 x 3 matrix (optional)
%// Must transpose as reshape is performed
%// in column major format
h(9) = 1; %// Add in that h33 is 1 before we reshape
hmatrix = reshape(h, 3, 3)';
Once you are finished, you have a combined rotation and translation matrix. If you want the x and y translations, simply pick off column 3, rows 1 and 2 in hmatrix. However, we can also work with the vector of h itself, and so h13 would be element 3, and h23 would be element number 6. If you want the angle of rotation, simply take the appropriate inverse trigonometric function to rows 1, 2 and columns 1, 2. For the h vector, this would be elements 1, 2, 4 and 5. There will be a bit of inconsistency depending on which elements you choose as this was solved by least squares. One way to get a good overall angle would perhaps be to find the angles of all 4 elements then do some sort of average. Either way, this is a good starting point.
References
I learned about homography a while ago through Leow Wee Kheng's Computer Vision course. What I have told you is based on his slides: http://www.comp.nus.edu.sg/~cs4243/lecture/camera.pdf. Take a look at slides 30-32 if you want to know where I pulled this material from. However, the MATLAB code I wrote myself :)
I have two images which one of them is the Original image and the second one is Transformed image.
I have to find out how many degrees Transformed image was rotated using 3x3 transformation matrix. Plus, I need to find how far translated from origin.
Both images are grayscaled and held in matrix variables. Their sizes are same [350 500].
I have found a few lecture notes like this.
Lecture notes say that I should use the following matrix formula for rotation:
For translation matrix the formula is given:
Everything is good. But there are two problems:
I could not imagine how to implement the formulas using MATLAB.
The formulas are shaped to find x',y' values but I already have got x,x',y,y' values. I need to find rotation angle (theta) and tx and ty.
I want to know the equivailence of x, x', y, y' in the the matrix.
I have got the following code:
rotationMatrix = [ cos(theta) sin(theta) 0 ; ...
-sin(theta) cos(theta) 0 ; ...
0 0 1];
translationMatrix = [ 1 0 tx; ...
0 1 ty; ...
0 0 1];
But as you can see, tx, ty, theta variables are not defined before used. How can I calculate theta, tx and ty?
PS: It is forbidden to use Image Processing Toolbox functions.
This is essentially a homography recovery problem. What you are doing is given co-ordinates in one image and the corresponding co-ordinates in the other image, you are trying to recover the combined translation and rotation matrix that was used to warp the points from the one image to the other.
You can essentially combine the rotation and translation into a single matrix by multiplying the two matrices together. Multiplying is simply compositing the two operations together. You would this get:
H = [cos(theta) -sin(theta) tx]
[sin(theta) cos(theta) ty]
[ 0 0 1]
The idea behind this is to find the parameters by minimizing the error through least squares between each pair of points.
Basically, what you want to find is the following relationship:
xi_after = H*xi_before
H is the combined rotation and translation matrix required to map the co-ordinates from the one image to the other. H is also a 3 x 3 matrix, and knowing that the lower right entry (row 3, column 3) is 1, it makes things easier. Also, assuming that your points are in the augmented co-ordinate system, we essentially want to find this relationship for each pair of co-ordinates from the first image (x_i, y_i) to the other (x_i', y_i'):
[p_i*x_i'] [h11 h12 h13] [x_i]
[p_i*y_i'] = [h21 h22 h23] * [y_i]
[ p_i ] [h31 h32 1 ] [ 1 ]
The scale of p_i is to account for homography scaling and vanishing points. Let's perform a matrix-vector multiplication of this equation. We can ignore the 3rd element as it isn't useful to us (for now):
p_i*x_i' = h11*x_i + h12*y_i + h13
p_i*y_i' = h21*x_i + h22*y_i + h23
Now let's take a look at the 3rd element. We know that p_i = h31*x_i + h32*y_i + 1. As such, substituting p_i into each of the equations, and rearranging to solve for x_i' and y_i', we thus get:
x_i' = h11*x_i + h12*y_i + h13 - h31*x_i*x_i' - h32*y_i*x_i'
y_i' = h21*x_i + h22*y_i + h23 - h31*x_i*y_i' - h32*y_i*y_i'
What you have here now are two equations for each unique pair of points. What we can do now is build an over-determined system of equations. Take each pair and build two equations out of them. You will then put it into matrix form, i.e.:
Ah = b
A would be a matrix of coefficients that were built from each set of equations using the co-ordinates from the first image, b would be each pair of points for the second image and h would be the parameters you are solving for. Ultimately, you are finally solving this linear system of equations reformulated in matrix form:
You would solve for the vector h which can be performed through least squares. In MATLAB, you can do this via:
h = A \ b;
A sidenote for you: If the movement between images is truly just a rotation and translation, then h31 and h32 will both be zero after we solve for the parameters. However, I always like to be thorough and so I will solve for h31 and h32 anyway.
NB: This method will only work if you have at least 4 unique pairs of points. Because there are 8 parameters to solve for, and there are 2 equations per point, A must have at least a rank of 8 in order for the system to be consistent (if you want to throw in some linear algebra terminology in the loop). You will not be able to solve this problem if you have less than 4 points.
If you want some MATLAB code, let's assume that your points are stored in sourcePoints and targetPoints. sourcePoints are from the first image and targetPoints are for the second image. Obviously, there should be the same number of points between both images. It is assumed that both sourcePoints and targetPoints are stored as M x 2 matrices. The first columns contain your x co-ordinates while the second columns contain your y co-ordinates.
numPoints = size(sourcePoints, 1);
%// Cast data to double to be sure
sourcePoints = double(sourcePoints);
targetPoints = double(targetPoints);
%//Extract relevant data
xSource = sourcePoints(:,1);
ySource = sourcePoints(:,2);
xTarget = targetPoints(:,1);
yTarget = targetPoints(:,2);
%//Create helper vectors
vec0 = zeros(numPoints, 1);
vec1 = ones(numPoints, 1);
xSourcexTarget = -xSource.*xTarget;
ySourcexTarget = -ySource.*xTarget;
xSourceyTarget = -xSource.*yTarget;
ySourceyTarget = -ySource.*yTarget;
%//Build matrix
A = [xSource ySource vec1 vec0 vec0 vec0 xSourcexTarget ySourcexTarget; ...
vec0 vec0 vec0 xSource ySource vec1 xSourceyTarget ySourceyTarget];
%//Build RHS vector
b = [xTarget; yTarget];
%//Solve homography by least squares
h = A \ b;
%// Reshape to a 3 x 3 matrix (optional)
%// Must transpose as reshape is performed
%// in column major format
h(9) = 1; %// Add in that h33 is 1 before we reshape
hmatrix = reshape(h, 3, 3)';
Once you are finished, you have a combined rotation and translation matrix. If you want the x and y translations, simply pick off column 3, rows 1 and 2 in hmatrix. However, we can also work with the vector of h itself, and so h13 would be element 3, and h23 would be element number 6. If you want the angle of rotation, simply take the appropriate inverse trigonometric function to rows 1, 2 and columns 1, 2. For the h vector, this would be elements 1, 2, 4 and 5. There will be a bit of inconsistency depending on which elements you choose as this was solved by least squares. One way to get a good overall angle would perhaps be to find the angles of all 4 elements then do some sort of average. Either way, this is a good starting point.
References
I learned about homography a while ago through Leow Wee Kheng's Computer Vision course. What I have told you is based on his slides: http://www.comp.nus.edu.sg/~cs4243/lecture/camera.pdf. Take a look at slides 30-32 if you want to know where I pulled this material from. However, the MATLAB code I wrote myself :)
My question is related to merging two matrices with different spacing on the field
Both matrices are in (frequency space domain (F-X)) and to illustrate lets consider :
Matrix One (A x B)
Matrix Two (A x C)
Where (A = 1500) is the frequency samples (they both share the same frequency samples number).
(B = 168 ) is the number of Receivers for the first data set (recording stations in the field) with spacing of 12 meters between B(1) and B(2), 12 meters between B(2) and B(3) and so on
(C = 196 ) is also the number of Receivers for the second data set (recording stations in the field) with different spacing than B. So C(1) to C(2) is 48 and C(2) to C(3) is 48 and so on.
The total length of receiver line in Matrix One is (168*12) = 2016 m.
The total length of receiver line in Matrix Two is (196*48) = 9408 m.
The line from matrix (one) sits on the middle part of line two.
Meaning they do over lap in many locations where both are present.(every forth receiver from the shorter line is overlapped by a receiver from the longer line)
I want to merge them (add them together) in a way where the final result represent the courser spacing (matrix two) added to it matrix A in the location where they overlap.
I hope I made this somehow clear
If A and B are your 2 matrices then merged would be the merged matrix. I used a size of 196x100 and 168x100 for the matrices just because I was to lazy to use 1500 ;).
C,D are integers ranging from 1 to the size of the first dimension.
A= randi(100, 168 ,100)+20;
B= randi(10, 196, 100);
C= 1:168; %ranging from 1 to number of receiver in Matrix 1(A)
D= 1:196; %ranging from 1 to number of receiver in Matrix 2(B)
C= C*12+4704; %Distance of each receiver to 0 and shifted to the middle of 2nd matrix
D= D*48; %Distance of each receiver to 0
A= [C', A]; %adding distance as first column to Matrix 1
B= [D', B]; %adding distance as first column to Matrix 2
C= [A;B]; %adding the 2 matrices
C=sortrows(C,1); %sorting the new matrix
merged = C(:,2:end); %deleting first column (distance)
What I am doing is generating the distance (C,D) appending them in column 1 appending the two matrices rowwise and sorting rowwise regarding the distance (new column 1). Then at the end I delete the distance because you didn't ask for it.
Just realized that your matrix is not 168x1500 but 1500x168, just convert it, using (A=A';)
If I'm understanding your question correctly, you want to add every 4th value of matrix 1 (I'll call this A) into matrix 2 (I'll call this B) at some defined position in the middle of 2. So we need to pick every 4th value of B and add it to A at some point of overlap p.
Here's how to select every 4th value of B along the 2nd dimension:
coarseB = B(:, 4:4:end);
Since matrices A and B are of different sizes, we have to expand the size of A with zeros so it can be added to B. It will help to know the size of coarseB along the 2nd dimension:
sizeB = size(B, 2);
coarseB = [zeros(1500, p - 1) coarseB zeros(1500, 168 - (p - 1) - sizeB)];
This will concatenate a zero array at the end and the beginning of coarseB to expand it to the size of A (1500 x 168).
Now coarseB can be added to A:
merged = coarseB + A;
Hope this helps!
I have a vector of 3D points lets say A as shown below,
A=[
-0.240265581092000 0.0500598627544876 1.20715641293013
-0.344503191645519 0.390376667574812 1.15887540716612
-0.0931248606994074 0.267137193112796 1.24244644549763
-0.183530493218807 0.384249186312578 1.14512014134276
-0.0201358671977785 0.404732019283683 1.21816745283019
-0.242108038906952 0.229873488902244 1.24229940627651
-0.391349107031230 0.262170158259873 1.23856838565023
]
what I want to do is to connect 3D points with lines which only have distance less than a specific threshold T. I want to get a list of pairs of points needed to be connected. Such as,
[
( -0.240265581092000 0.0500598627544876 1.20715641293013), (-0.344503191645519 0.390376667574812 1.15887540716612);
(-0.0931248606994074 0.267137193112796 1.24244644549763),(-0.183530493218807 0.384249186312578 1.14512014134276),.....
]
So as shown, I'll have a vector of pairs of points needed to be connected. So if anyone could please advise how this can be done in Matlab.
The following example demonstrates how to accomplish this.
%# Build an example matrix
A = [1 2 3; 0 0 0; 3 1 3; 2 0 2; 0 1 0];
Threshold = 3;
%# Calculate distance between all points
D = pdist2(A, A);
%# Discard any points with distance greater than threshold
D(D > Threshold) = nan;
If you wish to extract an index of all observation pairs that are linked by a distance less than (or equal to) Threshold, as well as the corresponding distance (your question didn't specify what form you wanted the output to take, so I am essentially guessing here), then instead use the following:
%# Obtain a list of linear indices of observations less than or equal to TH
I1 = find(D <= Threshold);
%#Extract the actual distances, as well as the corresponding observation indices from A
[Obs1Index, Obs2Index] = ind2sub(size(D), I1);
DList = [Obs1Index, Obs2Index, D(I1)];
Note, pdist2 uses Euclidean distance by default, but there are other options - see the documentation here.
UPDATE: Based on the OP's comments, the following code will express the output as a K*6 matrix, where K is the number of distance measures less than the threshold value, and the first three columns of each row is the first data point (3 dimensions) and the second three columns of each row is the connected data point.
DList2 = [A(Obs1Index, :), A(Obs2Index, :)];
SECOND UPDATE: I have not made any assumptions on the distance measure in this answer. That is, I'm deliberately using pdist2 in case your distance measure is not symmetric. However, if you are using a symmetric distance measure, then you could probably speed up the run-time by using pdist instead, although my indexing code would need to be adjusted accordingly.
Plot3 and pdist2 can be used to achieve what you want.
D=pdist2(A,A);
T=0.2;
for i=1:7
for j=i+1:7
if D(i,j)<T & D(i,j)~=0
i
j
plot3(A([i j],1),A([i j],2),A([i j],3));
hold on;
fprintf('line is plotted\n');
pause;
end
end
end
I am using Matlab to find the spectral radius of the Jacobi iteration matrix where A=[4 2 1;1 3 1;1 1 4].
I can't seem to input the correct commands to get the size of the error after 5 iterations. Can someone help me?
Here are a list of commands that I put into Matlab so far:
A=[4 2 1;1 3 1;1 1 4]
A =
4 2 1
1 3 1
1 1 4
D=diagonal(diagonal(A));L=(A,-1);U=(A,1);
b=([3 -1 4])
x0j=zeros([0 0 0]);
x=D\(-(U+L)*x0j+b);r=b-A*x %Jacobi iteration.
------------------------------------------------------------------------------
Error using *
Inputs must be 2-D, o enter code here r at least one input must be scalar.
To compute element wise TIMES, use TIMES (.*) instead.
The spectral radius of a matrix is the maximum of the modulus of its eigenvalues. It can be simply computed using max(abs(eig(·))).
However, as others have noticed, your whole code seems pretty mixed up and not actually valid Matlab code, so your problem is not really to compute the spectral radius, is it? The algorithm is very straightforward and easy to implement:
% diagonal part of A and rest
D = diag(diag(A));
R = A - D;
% iteration matrix and offset
T = - inv(D) * R;
C = inv(D) * b;
% spectral radius condition
rho = max(abs(eig(T)));
if rho >= 1
error('no convergence')
end
% initial guess
x = randn(size(b));
% iteration
while norm(A * x - b) > 1e-15
x = T * x + C;
end
Note that I used inv(D) to directly follow the description in Wikipedia, but the inverse of a diagonal matrix can be easily computed using diag(1 ./ diag(D)).
I don't really see why one would need to separate R into an upper and lower part. I suppose it has to do with numerical efficiency, but then, Matlab is a very efficient high-level language for matrix computations already. So actually there is no need to implement the Jacobi algorithm in it explicitly when you can simply write A \ b – except for educational purposes I guess.