Optoelectronic Sensor and Method for Safe Detection of Objects of a Minimum Size

Patent metadata

Field	Value
Publication number	US-2017214902-A1
Application number	US-201615376873-A
Country	US
Kind code	A1
Filing date	Dec 13, 2016
Priority date	Jan 26, 2016
Publication date	Jul 27, 2017
Grant date	—

Abstract

An optoelectronic sensor ( 10 ) for safe detection of objects ( 32, 34 ) of a minimum size in a monitoring area ( 12 ), the sensor ( 10 ) having an image sensor ( 16 a - b ) for acquiring image data of the monitoring area ( 12 ) and an evaluation unit ( 24 ) configured to detect, in the image data, both finely detected objects ( 34, 36 ) with a fine detection capability and coarsely detected objects ( 32 ) with a coarse detection capability, the coarse detection capability being coarser than the fine detection capability, and to ignore finely detected objects ( 36 ) which are not in a vicinity ( 38 ) of a coarsely detected object ( 32 ).

First claim

1 . An optoelectronic sensor ( 10 ) for safe detection of objects ( 32 , 34 ) of a minimum size in a monitoring area ( 12 ), the sensor ( 10 ) having an image sensor ( 16 a - b ) for acquiring image data of the monitoring area ( 12 ) and an evaluation unit ( 24 ) configured to detect, in the image data, both finely detected objects ( 34 , 36 ) with a fine detection capability and coarsely detected objects ( 32 ) with a coarse detection capability, the coarse detection capability being coarser than the fine detection capability, and to ignore finely detected objects ( 36 ) which are not in a vicinity ( 38 ) of a coarsely detected object ( 32 ).

2 . The sensor ( 10 ) according to claim 1 ,
wherein the evaluation unit ( 24 ) is configured to form a common object ( 42 ) from a coarsely detected object ( 32 ) and finely detected objects ( 34 ) in its vicinity ( 38 ).

3 . The sensor ( 10 ) according to claim 1 ,
wherein the sensor ( 10 ) is a 3D camera which detects a depth map as image data by means of its image sensor ( 16 a - b ).

4 . The sensor ( 10 ) according to claim 1 ,
wherein the fine detection capability enables leg detection, arm detection, hand detection, or finger detection, and wherein the coarse detection capability enables body detection, leg detection, or arm detection.

5 . The sensor ( 10 ) according to claim 1 ,
wherein the fine detection capability is 14 mm, 20 mm, 40 mm, or 55 mm, and wherein the coarse detection capability is 70 mm, 120 mm, 150 mm, or 200 mm.

6 . The sensor ( 10 ) according to claim 1 ,
wherein the evaluation unit ( 24 ), when locating finely detected objects ( 34 ) in the vicinity ( 38 ) of a coarsely detected object ( 32 ), at first defines a circumscribing geometric shape ( 40 ) for the coarsely detected object ( 32 ) and then determines a search area ( 38 ) with a specified distance (r) around the geometric shape ( 40 ) as the vicinity ( 38 ).

7 . The sensor ( 10 ) according to claim 6 ,
wherein the geometric shape ( 40 ) is a circle, a rectangle, a hexagon, or another polygon.

8 . The sensor ( 10 ) according to claim 6 ,
wherein the specified distance (r) is derived from safety margins (C 1 , C 2 ) in dependence on the detection capability as defined in safety standards.

9 . The sensor ( 10 ) according to claim 8 ,
wherein the safety standard is ISO 13855, IEC 61496, or a relevant equivalent.

10 . The sensor ( 10 ) according to claim 8 ,
wherein the specified distance (r) is derived from a difference of the safety margin (C 1 ) in dependence on the coarse detection capability minus the safety margin (C 2 ) in dependence on the fine detection capability.

11 . The sensor according to claim 1 ,
wherein the evaluation unit ( 24 ), when locating finely detected objects ( 34 ) in the vicinity ( 38 ) of a coarsely detected object ( 32 ), places a mask ( 44 ) having a geometric shape with the size of a specified distance (r) at each image pixel of the coarsely detected object ( 32 ), and thus the sum of the masks ( 44 ) determines the vicinity ( 38 ).

12 . The sensor ( 10 ) according to claim 11 ,
wherein the geometric shape ( 40 ) is a circle, a rectangle, a hexagon, or another polygon.

13 . The sensor ( 10 ) according to claim 11 ,
wherein the specified distance (r) is derived from safety margins (C 1 , C 2 ) in dependence on the detection capability as defined in safety standards.

14 . The sensor ( 10 ) according to claim 13 ,
wherein the specified distance (r) is derived from a difference of the safety margin (C 1 ) in dependence on the coarse detection capability minus the safety margin (C 2 ) in dependence on the fine detection capability.

15 . The sensor ( 10 ) according to claim 1 ,
wherein the evaluation unit ( 24 ) is configured to evaluate whether a detected object ( 32 , 34 , 42 ) is located within a predefined protection field ( 28 ) or too close to a source of danger ( 26 ), and to output a safety-related shutdown signal via a safe output ( 30 ) in this case.

16 . The sensor ( 10 ) according to claim 15 ,
wherein the evaluation unit ( 24 ) is configured to evaluate different protection fields ( 28 ) for the coarse detection capability and the fine detection capability in that the protection field ( 28 ) is enlarged by a safety extension for the coarse detection capability.

17 . A method for safe detection of objects ( 32 , 34 ) of a minimum size in a monitoring area ( 12 ), wherein image data is acquired and evaluated to detect, in the image data, both finely detected objects ( 34 , 36 ) with a fine detection capability and coarsely detected objects ( 32 ) with a coarse detection capability, the coarse detection capability being coarser than the fine detection capability, and wherein finely detected objects ( 36 ) which are not in a vicinity ( 38 ) of a coarsely detected object ( 32 ) are ignored.

18 . The method according to claim 17 ,
wherein a common object ( 40 ) is formed from a coarsely detected object ( 32 ) and finely detected objects ( 34 ) in its vicinity ( 38 ), and wherein it is evaluated whether the common object ( 40 ) intrudes into a protection field ( 28 ) or is located too close to a source of danger ( 26 ).

19 . The method according to claim 17 ,
wherein the vicinity ( 38 ) is defined by a specified distance (r) with respect to each image pixel of a coarsely detected object ( 32 ) or a circumscribed geometric shape ( 40 ) of the coarsely detected object ( 32 ).

20 . The method according to claim 19 ,
wherein the specified distance (r) is derived from safety margins (C 1 , C 2 ) in dependence on the detection capability as defined in safety standards.

Assignees

• Sick Ag

Inventors

• Braune Ingolf

Classifications

• G06T7/50

  Depth or shape recovery · CPC title

• G08B13/19602

  Image analysis to detect motion of the intruder, e.g. by frame subtraction · CPC title

• H04N13/271 Primary

  wherein the generated image signals comprise depth maps or disparity maps · CPC title

• G06T2207/30196

  Human being; Person · CPC title

• G08B13/19626

  optical details, e.g. lenses, mirrors or multiple lenses · CPC title