How to partition the image into blocks of different sizes and In each block, take the angles of all black points with respect to the bottom left corner and calculate the summation and normalize it?

Please help me. I have pre-processed the image but i am not able to do this. If not possible then at least tell me how to partition the image and take the angles. Thanks in advance.

Answer 1

[Image Analyst]

For the partitioning part, see the FAQ: http://matlab.wikia.com/wiki/FAQ#How_do_I_split_an_image_into_non-overlapping_blocks.3F

For the angle, get the delta X and delta Y from the rows and columns of all the black pixels

[rows, columns] = find(~binaryImage);

Then get angles

angles = atand2d(deltaY/deltaX);

Of course you figure out what delta x and y are from the particular pixel to the lower right corner of the current blobk you're at. Not hard. Give it a try.

[Roger Stafford]

@IA: If you meant 'atan2d', it requires two arguments.

Answer 2

[VINSHI K K]

How can I divide an image of size [827 592] into overlaping blocks with block size=[252 252] and overlapping width=[64 64]

[Image Analyst]

See my demo of blockproc with overlapping block positions. Adapt as needed.