EDGE function ERROR

I'm trying to perform the following operation

%Load the image into matlab

    A = imread('before3.jpg');

%Show the image
  figure(1)
  imshow(A)

    I = edge(A, 'sobel');

Using this high pass filter I expect to find the edges of my image but I get the following error while using the edge function.

??? Function EDGE expected its first input, I, to be two-dimensional.

Error in ==> edge>parse_inputs at 564
iptcheckinput(I,{'numeric'},{'nonsparse','2d'},mfilename,'I',1);

Error in ==> edge at 197
[a,method,thresh,sigma,thinning,H,kx,ky] = parse_inputs(varargin{:});

Error in ==> image1 at 21
I = edge(A,'sobel');

The image 'before3.jpg' has value

   <1728x2304x3 uint8>

The error says to use a 2 dimensional input but I have not been able to find a way to change my image to a 2 dimensional matrix and keep at the same time the original edges.

I tried

[X, map] = rgb2ind(A,256);

I used as the map a 'gray' scale but it did not work.

Please if you have a different approach that you can recommend me I will be very grateful.

Answer 1

[JOEL]

EFECTIVAMENTE AMIGO, LA FUNCION edge SOLO ADMITE IMAGEN ES ESCALAS GRISES, REALIZA LO SIGUIENTE:

%para leer una imagen
I = imread('casa.jpg');
%% convertir a escalas grises
I=rgb2gray(I);
%hallando contorno
BW = edge(I);
%mostrando contorno
imshow(BW)

slds,

Joel Villavicencio

[bharat arora]

thanks

[Muddasar Khan]

Yes, You are right. We should convert colored image to gray image then edges should be detected

[princess sofiya]

thanks too.

Answer 2

[Walter Roberson]

rgb2gray() rather than rgb2ind()

Answer 3

[Jan]

In the help text of edge I find:

 W = edge(I) takes a grayscale or a binary image I as its input...

I'd expect rgb2gray to perform a proper conversion.

[SUBBARAMIREDDY YERUVA]

Thanks alot

Answer 4

[Alex]

according to the documentation, using imread with a jpeg object results in RGB scale. As in, each pixel is represented by a 3 element vector ['red value' 'green value' 'blue value'];

[Walter Roberson]

Not universally true. There are true grayscale jpeg images. But I have only ever encountered one, apart from test images to prove that they exist. Jpeg is nearly always rgb. Never pseudocolor though.

Answer 5

[Eulises Ulloa]

Thank you guys, that really helped.

However, using the edge function after converting my image to gray scale does not give good result in my case. I have seem some examples on the web and even on the documentation of the function in where they use the edge function right after "imread". I always get the needs 2 dimensional input error. Can you tell me if you know other approach for finding the edges of a image.

e.g I would like to perform the edge detection operation on the original image.

[Walter Roberson]

The places that used edge() right after imread() were places where the image being read was grayscale.

[Jan]

If you want to use the edge-detection on an RGB image, you cannot use the EDGE function. There are a lot of edge-detection methods in the FEX: http://www.mathworks.com/matlabcentral/fileexchange/?term=edge

There are other methods to convert the image to gray-scale also. Consider to sharpen the edges in the HSV color-space at first.

[Eulises Ulloa]

Thanks

Answer 6

[Priscil23]

Just as everyone said, first convert your rgb2gray then to view the output after your desired edge functions use imshow to view the image. I faced the same problem.

Answer 7

[priyanka nale]

close all;
clear ;
I=imread('4.jpg');
figure,imshow(I)
I=rgb2gray(I);
c= edge(I,'sobel');
figure,imshow(c);