System and method for detecting object from depth image

Patent metadata

Field	Value
Publication number	US-2017193665-A1
Application number	US-201615007942-A
Country	US
Kind code	A1
Filing date	Jan 27, 2016
Priority date	Jan 4, 2016
Publication date	Jul 6, 2017
Grant date	—

Abstract

Provided is a system for detecting an object from a depth image. The system includes a communication module, a memory, and a processor. By executing the object detection program, the processor extracts a first object area and a second object area from the depth image, based on a predetermined floor plane and an outer plane which is set with respect to the predetermined floor plane. The processor extracts a target area including pixels of the second object area which are spaced apart from the first object area by a predetermined interval. The processor samples a pixel, which is not included in the target area, to extract a floor area from the second object area, calculates a boundary value of an object and a floor, based on the floor area and the target area, and extracts a foreground pixel from the target area, based on the calculated boundary value.

First claim

What is claimed is:

1 . A system for detecting an object from a depth image, the system comprising:
a communication module; a memory configured to store an object detection program; and a processor configured to execute the object detection program, wherein by executing the object detection program, the processor extracts a first object area and a second object area from the depth image, based on a predetermined floor plane and an outer plane which is set with respect to the predetermined floor plane, the processor extracts a target area including pixels of the second object area which are spaced apart from the first object area by a predetermined interval, the processor samples a pixel, which is not included in the target area, to extract a floor area from the second object area, calculates a boundary value of an object and a floor, based on the floor area and the target area, and extracts a foreground pixel from the target area, based on the calculated boundary value, and the second object area comprises the object and the floor and is disposed under the first object area.

2 . The system of claim 1 , wherein
the floor plane comprises an upper floor plane and a lower floor plane which are set above and under the floor plane in parallel to be spaced apart from each other by a predetermined interval, and the processor extracts the first object area, based on the upper floor plane and the outer plane.

3 . The system of claim 2 , wherein the processor extracts the second object area, based on the outer plane, the upper floor plane, and the lower floor plane.

4 . The system of claim 1 , wherein
the processor segments the target area into a plurality of block areas, and the processor samples a pixel, which is included in the second object area but is not included in the target area, with respect to one of the segmented block areas to extract a floor area from the second object area.

5 . The system of claim 4 , wherein the processor removes pixels, corresponding to a direction opposite to a direction of the first object area, from among pixels included in each of the block areas, calculates distances between a plane extracted from the floor area and pixels having the same direction as the direction of the first object area, and calculates a boundary value between the object and the floor, based on a distribution of the calculated distances.

6 . The system of claim 5 , wherein the foreground pixel is extracted from among pixels having the calculated boundary value or more in the plurality of block areas.

7 . The system of claim 1 , wherein the communication module receives the depth image in units of one frame from the depth camera.

8 . The system of claim 1 , wherein
the floor plane is set with respect to four pixels, and the outer plane is set as a plane contacting two pixels, which are successively disposed, among the four pixels in planes vertical to the floor plane.

9 . An object detection method performed by an object detection system, the object detection method comprising:
receiving a depth image from a depth camera; extracting a first object area from the depth image, based on a predetermined floor plane and an outer plane which is set with respect to the predetermined floor plane; extracting a second object area from the depth image, based on the predetermined floor plane and the outer plane; extracting a target area including pixels of the second object area which are spaced apart from the first object area by a predetermined interval; sampling a pixel, which is not included in the target area, to extract a floor area from the second object area; calculating a boundary value of an object and a floor, based on the floor area and the target area; and extracting a foreground pixel from the target area, based on the calculated boundary value, wherein the second object area comprises the object and the floor and is disposed under the first object area.

10 . The object detection method of claim 9 , wherein
the floor plane comprises an upper floor plane and a lower floor plane which are set above and under the floor plane in parallel to be spaced apart from each other by a predetermined interval, the extracting of the first object area comprises extracting the first object area, based on the upper floor plane and the outer plane, and the extracting of the second object area comprises extracting the second object area, based on the outer plane, the upper floor plane, and the lower floor plane.

11 . The object detection method of claim 9 , wherein the extracting of the floor area comprises:
segmenting the target area into a plurality of block areas; and sampling a pixel, which is included in the second object area but is not included in the target area, with respect to one of the segmented block areas to extract a floor area from the second object area.

12 . The object detection method of claim 11 , wherein
the calculating of the boundary value comprises: removing pixels, corresponding to a direction opposite to a direction of the first object area, from among pixels included in each of the block areas; extracting a plane corresponding to the floor area; calculating distances between the plane extracted from the floor area and pixels having the same direction as the direction of the first object area; and calculating the boundary value between the object and the floor, based on a distribution of the calculated distances, and the extracting of the foreground comprises extracting the foreground pixel from among pixels having the calculated boundary value or more in the plurality of block areas.

13 . The object detection method of claim 9 , wherein the extracting of the floor area, the calculating of the boundary value, and the extracting of the foreground pixel are performed for each of the block areas.

Assignees

• Electronics & Telecommunications Res Inst

Inventors

• Lee Su Woong
• Lee Jun Suk
• Jee Hyung Keun

Classifications

• G06T7/0079 Primary

  Physics · mapped topic

• G06K9/00335

  Physics · mapped topic

• G06T2207/20144

  Physics · mapped topic

• G06T2207/10028

  Range image; Depth image; 3D point clouds · CPC title

• G06T7/194

  involving foreground-background segmentation · CPC title