Locating objects within depth images
In accordance with some embodiments, connected-component labeling is performed in both the screen dimensions (which may be referred to as the x and y dimensions) and a depth dimension to label objects in a depth image. Then the contour of labeled blobs may be used to identify an object in the depth image. Using contours may be advantageous in some embodiments because it reduces the amount of data that must be handled and the extent of computations, compared to conventional techniques which use bit map based operations.
1. A method comprising:
performing on a computer connected components labelling in both screen dimensions and a depth dimension to label objects in a depth image;
using a contour of said labels to identify an object in the depth image using said computer;
identifying an object within a labeled blob by finding a convex hull that surrounds a blob labelled by said connected components labelling; and
finding a defect in said blob and a closest opposite point on the convex hull and using a line from said defect to said closest opposite point to segment the object to be identified from a remainder of the blob.
2. The method of claim 1 including performing connected components labelling by allowing two labels to be neighbors that touch one another.
3. The method of claim 1 including identifying an occlusion.
4. The method of claim 3 including determining which of at least two objects, a group of pixels is connected to, using shape erosion followed by propagation.
5. The method of claim 3 including correcting for occlusion by relabeling only moving pixels.
6. The method of claim 1 including developing an object movement threshold to determine whether a depicted object is an object to be identified.
7. The method of claim 6 including determining the threshold based on an expected maximum speed of an object to be identified.
8. The method of claim 7 including basing the threshold on the maximum number of desired labels.
9. One or more non-transitory computer readable media storing instructions to perform a sequence comprising:
performing using a computer connected components labelling in both screen dimensions and a depth dimension to label objects in a depth image using said computer;
using a contour of said labels to identify an object in the depth image;
identifying an object within a labeled blob by finding a convex hull that surrounds a blob labelled by said connected components labelling; and
finding a defect in said blob and a closest opposite point on the convex hull and using a line from said defect to said closest opposite point to segment the object to be identified from a remainder of the blob.
10. The media of claim 9 , further storing instructions to perform a sequence including performing connected components labelling by allowing two labels to be neighbors that touch one another.
11. The media of claim 9 , further storing instructions to perform a sequence including identifying an occlusion.
12. The media of claim 11 , further storing instructions to perform a sequence including determining which of at least two objects, a group of pixels is connected to, using shape erosion followed by propagation.
13. The media of claim 11 , further storing instructions to perform a sequence including correcting for occlusion by relabeling only moving pixels.
14. The media of claim 9 , further storing instructions to perform a sequence including developing an object movement threshold to determine whether a depicted object is an object to be identified.
15. The media of claim 14 , further storing instructions to perform a sequence including determining the threshold based on an expected maximum speed of an object to be identified.
16. The media of claim 15 , further storing instructions to perform a sequence including basing the threshold on the maximum number of desired labels.
17. An apparatus comprising:
a processor to perform connected components labelling in both screen dimensions and a depth dimension to label objects in a depth image and use a contour of said labels to identify an object in the depth image, identify an object within a labeled blob by finding a convex hull that surrounds a blob labelled by said connected components labelling, find a defect in said blob and a closest opposite point on the convex hull and using a line from said defect to said closest opposite point to segment the object to be identified from a remainder of the blob; and
a memory coupled to said processor.
18. The apparatus of claim 17 , said processor to perform connected components labelling by allowing two labels to be neighbors that touch one another.
19. The apparatus of claim 17 , said processor to identify an occlusion.
20. The apparatus of claim 19 , said processor to determine which of at least two objects, a group of pixels is connected to, by using shape erosion followed by propagation.
21. The apparatus of claim 19 , said processor to correct for occlusion by relabeling only moving pixels.
22. The apparatus of claim 17 , said processor to develop an object movement threshold to determine whether a depicted object is an object to be identified.
23. The apparatus of claim 22 , said processor to determine the threshold based on an expected maximum speed of an object to be identified.
24. The apparatus of claim 23 , said processor to base the threshold on the maximum number of desired labels.
25. The apparatus of claim 17 including a battery.