Three dimensional (3D) object detection

1. A method comprising:

receiving a first image that includes a set of pixels;

determining a set of optical flow vectors for the first image with respect to a second image that precedes the first image in time, wherein the set of optical flow vectors includes a respective vector for each pixel of the set of pixels;

determining a set of magnitudes for the set of optical flow vectors;

dividing the set of pixels into a first set of blocks having a first size;

determining a first set of histograms of magnitudes for the first set of blocks, wherein the first set of histograms of magnitudes includes a respective histogram for each block of the first set of blocks;

dividing the set of pixels into a second set of blocks having a second size that is different from the first size;

determining a second set of histograms of magnitudes for the second set of blocks, wherein the second set of histograms of magnitudes includes a respective histogram for each block of the second set of blocks; and

determining whether an object is present in the first image based on the first set of histograms of magnitudes and the second set of histograms of magnitudes.

2. The method of claim 1 , wherein the determining of whether an object is present further includes:

determining a set of gradients of the set of optical flow vectors;

determining a histogram of gradients of a subset of the set of gradients; and

determining whether the object is present further based on the histogram of gradients.

3. The method of claim 2 , wherein:

the determining of whether an object is present further includes determining a region of interest of the first image based on the first set of histograms of magnitudes and the second set of histograms of magnitudes; and

the subset of the set of gradients is associated with the region of interest.

4. The method of claim 3 , wherein:

the determining of whether an object is present further includes determining a mask based on the region of interest; and

the determining of the histogram of gradients includes excluding a portion of the first image outside the region of interest based on the mask.

5. The method of claim 1 , wherein the determining of whether an object is present further includes:

determining a set of gradients of the set of optical flow vectors;

dividing the set of pixels into a third set of blocks having a third size;

determining a first set of histograms of gradients for the third set of blocks;

dividing the set of pixels into a fourth set of blocks having a fourth size; and

determining a second set of histograms of gradients for the fourth set of blocks.

6. The method of claim 5 , wherein:

the first size and the third size are the same; and

the second size and the fourth size are the same.

7. The method of claim 1 , wherein:

the first set of blocks do not overlap; and

the second set of blocks do not overlap.

8. The method of claim 1 , wherein the determining of whether an object is present further includes determining whether a region of the first image includes: a three-dimensional object, a flat feature, or no feature based on the first set of histograms and the second set of histograms.

9. The method of claim 1 , wherein the determining of the set of optical flow vectors is performed based on at least one of: a phase correlation algorithm, a sum of squared differences algorithm, a sum of absolute difference algorithm, normalized cross-correlation algorithm, a differential optical flow algorithm, or a discrete optimization optical flow algorithm.

10. A system comprising:

an image capture device configured to provide an image that includes a set of pixels;

a processor coupled to the image capture device and configured to:

determine a set of optical flow vectors for the image;

determine a set of magnitudes for the set of optical flow vectors; and

a histogram circuit coupled to the processor and configured to:

determine a first block of pixels of the set of pixels, wherein the first block has a first size;

determine a first histogram of magnitudes for the first block of pixels;

determine a second block of pixels of the set of pixels, wherein the second block has a second size that is different from the first size; and

determine a second histogram of magnitudes for the second block of pixels, wherein the processor is further configured to determine whether an object is present in the image based on the first histogram of magnitudes and the second histogram of magnitudes.

11. The system of claim 10 , wherein:

the processor is further configured to determine a set of gradients for the set of optical flow vectors;

the histogram circuit is further configured to:

determine a third block of pixels of the set of pixels, wherein the third block has a third size;

determine a first histogram of gradients for the third block of pixels;

determine a fourth block of pixels of the set of pixels, wherein the fourth block has a fourth size that is different from the third size; and

determine a second histogram of gradients for the fourth block of pixels; and

the processor is further configured to determine whether the object is present in the image based on the first histogram of gradients and the second histogram of gradients.

12. The system of claim 11 , wherein the processor is configured to determine a region of interest of the image based on the first histogram of magnitudes and the second histogram of magnitudes.

13. The system of claim 12 , histogram circuit is configured to determine the first histogram of gradients and the second histogram of gradients based on the third block of pixels and the fourth block of pixels being with the region of interest.

14. The system of claim 11 , wherein:

the first size and the third size are the same; and

the second size and the fourth size are the same.

15. The system of claim 10 , wherein the histogram circuit is configured to:

determine a first set of blocks of pixels of the set of pixels that includes the first block such that blocks of the first set of blocks do not overlap; and

determine a second set of blocks of pixels of the set of pixels that includes the second block such that blocks of the second set of blocks do not overlap.

16. The system of claim 10 , wherein the processor is configured to determine the set of optical flow vectors based on at least one of: a phase correlation algorithm, a sum of squared differences algorithm, a sum of absolute difference algorithm, normalized cross-correlation algorithm, a differential optical flow algorithm, or a discrete optimization optical flow algorithm.

17. A method comprising:

receiving an image that includes a set of pixels;

determining a set of optical flow vectors for the image;

determining a set of magnitudes for the set of optical flow vectors;

determining a first block of pixels of the set of pixels, wherein the first block has a first size;

determining a first histogram of magnitudes for the first block of pixels;

determining a second block of pixels of the set of pixels, wherein the second block has a second size that is different from the first size;

determining a second histogram of magnitudes for the second block of pixels; and

performing object detection on the image based on the first histogram of magnitudes and the second histogram of magnitudes.

18. The method of claim 17 further comprising:

determining a set of gradients for the set of optical flow vectors;

determining a third block of pixels of the set of pixels, wherein the third block has a third size;

determining a first histogram of gradients for the third block of pixels;

determining a fourth block of pixels of the set of pixels, wherein the fourth block has a fourth size that is different from the third size;

determining a second histogram of gradients for the fourth block of pixels; and

performing object detection on the image further based on the first histogram of gradients and the second histogram of gradients.