Method and apparatus for disparity computation in stereo images

1. A method for generating a disparity map, comprising the steps of:

segmenting a first reference image acquired at a first time into one or more segments, each segment comprising a plurality of pixels,

segmenting a corresponding second reference image acquired at the first time into one or more segments, each segment comprising a plurality of pixels;

determining a disparity for one or more segments between the first reference image and the corresponding second reference image by overlaying the first reference image on the second reference image and performing subtraction between at least one of the one or more segments in the first reference image and corresponding segments in the second reference image;

generating a first disparity map based upon the subtraction between the first image and the corresponding second image;

segmenting a third reference image acquired at a second time image into one or more segments, each segment comprising a plurality of pixels,

segmenting a corresponding fourth reference image acquired at the second time into one or more segments, each segment comprising a plurality of pixels;

determining one or more altered segments comprising a difference between one of the first and second images acquired at the first time and one of the corresponding third and fourth images acquired at the second time;

determining a disparity only for the determined one or more altered segments between the third reference image and the corresponding fourth reference image;

generating a disparity map update including the one or more determined disparities for the one or more altered segments between the third reference image and the corresponding fourth reference image;

generating a composite disparity map by combining the disparity map update and the first disparity map; and

generating a depth map from the composite disparity map.

2. The method of claim 1 , wherein a determined disparity for a segment is assigned to all pixels in the segment.

3. The method of claim 1 , wherein combining the disparity map update and the first disparity map further comprises the steps of:

identifying one or more locations in the first disparity map corresponding to the determined one or more altered segments; and

substituting for each identified location in the first disparity map a corresponding portion of the disparity map update.

4. A system for generating a disparity map, comprising:

a camera for acquiring a first plurality of images at a first time and for acquiring a second plurality of corresponding images at a second time; and

a processor for generating a first disparity map for at least two of the first plurality of images, determining one or more objects that change position between at least one of the first plurality of images and corresponding at least one of the second plurality of images, generating disparity map update segments corresponding to the determined objects, and combining the first disparity map and the disparity map update segments to generate a composite disparity map;

wherein a determination of disparity between the at least two images acquired at the second time is performed only on portions thereof that have changed from one or more of the at least one images acquired at the first time.

5. The system of claim 4 , wherein the camera is a stereo camera, and wherein the first and second plurality of images each comprise a left and a right image acquired by the stereo camera.

6. The system of claim 4 , wherein the processor, when combining the first disparity map and the disparity map update segments, is further adapted to:

determine the position of the objects in the second plurality of images; and

replace these positions in the first disparity map with the corresponding disparity map update segments.

7. The system of claim 4 , wherein the objects comprise portions of an image that have changed.