============================================================================== Computer Vision and Image Processing functions and scripts

Automatically crop valid ROI

Assuming the pixel (1,1) has the color of the border to remove from one image, crop the image returning only the rectangle containing valid pixels.

File: autocropimage.m

Collect row, column and value of the selected points

Collect row, column and value of the selected points in an image given the handle to the figure.

File: collectcursorinfo.m

Create a video from a list of images

Create a video from a list of black and white or colored images, with the option to specify the frame rate and whether to compress the video or not.

File: createvideofromimages.m

Sample: test_createvideofromimages.m

Save data in a OpenCV XML file

Save data according to OpenCV XML format. Th input is a struct with multiple fields that can be strings, numbers or matrices.

File: data2opencvxml.m

Convert disparity map to depth map and vice versa

Compute a depth map given a disparity map and the product baseline * focal length. It can be used also to compute the opposite, given a depth map, compute the disparity map.

File: disparity2depth.m

Fit a plane

Fit a plane on a 3D point cloud (x,y,z).

File: fitplate.m

Read 4 channels PNG image as float matrix

Interpret RGB and alpha channels as 32 bits elements of a matrix.

File: imread32f.m

Save a float matrix in a 4 channels PNG image

Save a 32 bit image (float) into a PNG image using RGB and alpha channel.

File: imwrite32f.m

Load data in the xml file stored with OpenCV

Create a structure with all the entries in the xml file. Deal with Mat, string and single numbers. Each entry in the structure has the name of the variable in the XML file.

File: loadopencvxmlfile.m

Load matrix stored in a binary file

Load a matrix stored in a binary file, given rows columns and type of tha data (e.g. 'float').

File: loadrawmat.m

Create ply point cloud

Converts 3D point cloud (x,y,z) to '.ply' format. The point cloud can have additional columns for the color information.

File: mat2ply.m

Sample: test_stereobm.m

Otsu's segmentation method with 3 gray levels

Example that shows the Otsu's algorithm for segmenting an image whose color distribution presents 3 peaks. The algorithm find the thresholds such that the within-class variance is minimized. This is an extension of the original algorithm for two peaks distribution.

Directory: otsu3

Load ply point cloud

Load a '.ply' file and store 3D point cloud in a N x 3 matrix (x,y,z). If the original point cloud has also color information, this will be saved as well.

File: ply2mat.m

Compute Block Matching

Class to produce a disparity map by computing stereo block matching given a couple of rectified images.

File: stereobm.m

Sample: test_stereobm.m

Back project 2D points to 3D

Project a 2D depth map in the 3D space.

File: to3d.m