how to measure distance and centroid of moving object with Matlab stereo computer vision? -


which matlab functions or examples should used (1) track distance moving object stereo (binocular) cameras, , (2) track centroid (x,y,z) of moving objects, ideally in range of 0.6m 6m. cameras?

i've used matlab example uses peopledetector function, becomes inaccurate when person within 2m. because begins clipping heads , legs.

the first thing need deal with, in how detect object of interest (i suppose have resolved issue). there lot of approaches of how detect moving objects. if cameras stand in fix position can work 1 camera , use background subtraction objects appear in scene (some info here). if cameras are moving, think best approach work optical flow of 2 cameras (instead use previous frame flow map, stereo pair images used optical flow map in each fame).

in matlab, there option called disparity computation, try detect objects in scene, after need add stage extract objects of interest, can use thresholds. once have desired objects, need put them in binary mask. in mask can use image momentum (check this , this) extractor calculate centroids. if images in binary mask noissy can use morphological operations improve reults (watch this).


Comments

Popular posts from this blog

javascript - Using jquery append to add option values into a select element not working -

Android soft keyboard reverts to default keyboard on orientation change -

Rendering JButton to get the JCheckBox behavior in a JTable by using images does not update my table -