Method and apparatus with moving object detection

1. A processor-implemented method of detecting a moving object, the method comprising:

estimating a depth image of a current frame;

determining an occlusion image of the current frame by calculating a depth difference value between the estimated depth image of the current frame and an estimated depth image of a previous frame;

determining an occlusion accumulation image of the current frame by adding a depth difference value of the occlusion image of the current frame to a depth difference accumulation value of an occlusion accumulation image of the previous frame; and

outputting an area of a moving object based on the occlusion accumulation image,

wherein the determining of the occlusion accumulation image of the current frame comprises estimating a depth difference accumulation value of a newly discovered area in the occlusion accumulation image of the current frame based on a similarity between a depth value of the depth image of the current frame and the occlusion accumulation image of the previous frame.

2. The method of claim 1 , wherein the estimating comprises:

warping coordinates of the depth image of the previous frame into coordinates of the depth image of the current frame based on position information or pose information of a camera; and

estimating an area having an inestimable depth value in the depth image of the current frame using the warped depth image of the previous frame.

3. The method of claim 2 , wherein

the outputting of the area comprises outputting the area of the moving object in a color image, and

the area having the inestimable depth value is an area having a depth value that is not estimated by a depth sensor used to generate the depth images due to a position or angle difference between the depth sensor and another sensor used to estimate the color image.

4. The method of claim 1 , wherein the newly discovered area corresponds to an area outside of a boundary of the depth image of the current frame.

5. The method of claim 1 , wherein the determining of the occlusion accumulation image of the current frame comprises initializing a pixel with a depth difference accumulation value being less than or equal to a first threshold value in the occlusion accumulation image.

6. The method of claim 5 , wherein the initializing of the pixel comprises initializing the depth difference accumulation value of the pixel to a value of “0”.

7. The method of claim 1 , wherein the determining of the occlusion accumulation image of the current frame comprises initializing a pixel with a depth difference value being less than or equal to a second threshold value in the occlusion accumulation image.

8. The method of claim 1 , wherein the outputting comprises:

determining an area including one or more pixels having a depth difference accumulation value being greater than a first threshold value in the occlusion accumulation image; and

outputting a masking image representing an area of the moving object based on the determined area.

9. A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, configure the one or more processors to perform the method of claim 1 .

10. An apparatus for detecting a moving object, the apparatus comprising:

one or more processors configured to:

estimate a depth image of a current frame,

determine an occlusion image of the current frame by calculating a depth difference value between the estimated depth image of the current frame and an estimated depth image of a previous frame,

determine an occlusion accumulation image of the current frame by adding a depth difference value of the occlusion image of the current frame to a depth difference accumulation value of an occlusion accumulation image of the previous frame, and

output an area of a moving object based on the occlusion accumulation image,

wherein, for the determining of the occlusion accumulation image of the current frame, the one or more processors are configured to estimate a depth difference accumulation value of a newly discovered area in the occlusion accumulation image of the current frame based on a similarity between a depth value of the depth image of the current frame and the occlusion accumulation image of the previous frame.

11. The apparatus of claim 10 , wherein, for the estimating, the one or more processors are configured to:

warp coordinates of the depth image of the previous frame into coordinates of the depth image of the current frame based on position information or pose information of a camera, and

estimate an area having an inestimable depth value in the depth image of the current frame using the warped depth image of the previous frame.

12. The apparatus of claim 10 , wherein, for the determining of the occlusion accumulation image of the current frame, the one or more processors are configured to:

initialize a pixel with a depth difference accumulation value being less than or equal to a first threshold value in the occlusion accumulation image.

13. The apparatus of claim 10 , wherein, for the determining of the occlusion accumulation image of the current frame, the one or more processors are configured to:

initialize a pixel with a depth difference value being less than or equal to a second threshold value in the occlusion accumulation image.

14. The apparatus of claim 10 , wherein, for the outputting, the one or more processors are configured to:

obtain an area including one or more pixels having a depth difference accumulation value being greater than a first threshold value in the occlusion accumulation image, and

output a masking image representing an area of the moving object based on the obtained area.

15. The apparatus of claim 10 , further comprising a memory storing instructions that, when executed by the one or more processors, configure the one or more processors to perform the estimating of the depth image of the current frame, the determining of the occlusion image of the current frame, the determining of the occlusion accumulation image of the current frame, and the outputting of the area of the moving object.

16. A processor-implemented method of detecting a moving object, the method comprising:

acquiring a current depth image including an area in which a depth value is not obtained by a depth sensor;

estimating the current depth image, including the area not obtained by the depth sensor, by warping a previous depth image into coordinates of the current depth image; and

determining a current occlusion image by calculating a difference between the estimated depth image and the previous depth image;

determining a current occlusion accumulation image by adding the current occlusion image to a previous occlusion accumulation image; and

outputting an area of a moving object based on the current occlusion accumulation image,

wherein the determining of the occlusion accumulation image of the current frame comprises estimating a depth difference accumulation value of a newly discovered area in the occlusion accumulation image of the current frame based on a similarity between a depth value of the depth image of the current frame and the occlusion accumulation image of the previous frame.