Method and device for detecting an object in a vehicle in particular for occupant protection systems

• US 20040040772A1
• Filed: 06/19/2003
• Published: 03/04/2004
• Est. Priority Date: 12/20/2000
• Status: Active Grant

• Current Assignee: Continental Automotive GmbH (Continental AG)
• Original Assignee: Siemens AG
• Inventors: Ertl, Ludwig, Kohler, Throsten
• Granted Patent: US 7,376,248 B2
• Time in Patent Office: Days
• US Class Current: 180/271
• CPC Class Codes: B60R 21/01538 for image processing, e.g. cameras or sensor arrays

Abstract

In optical image processing methods for determining the occupation of a vehicle seat, it is advantageous to know the position of the vehicle seat (14), even in the case when said seat is occupied. The above is achieved, whereby the side surface of the backrest (11), facing the camera (30) is part of a boundary plane. A section (T, B), between a separating plane, cutting vertically through the vehicle seat and the boundary plane (B) is at least partly not covered by a seat occupant in the direction of the camera. By means of an image processing method the position of the seat may be determined.

Claims

1. Method for detecting an object in a vehicle, in particular for occupant protection systems, wherein

an image area in the interior of the vehicle (1) is registered by a camera (30),

the image information (13) registered by the camera (30) is transmitted to an evaluation unit (31),

a specified section (T, B) which contains a characteristic feature of an element (10, 12, 13, 40) of the vehicle is selected from the image area, and

the position of the characteristic feature of the element (1, 40, 41) of the vehicle (1) is determined from the image information (13) registered in the specified section (T, B).

2. Method according to claim 1, characterized in that

the specified section (T, B) encompasses an area which is enclosed by two planes (T, B) which contain and/or touch a side surface of the element (vehicle seat (1, 10, 11)) facing the camera (30).

3. Method according to one of the claims 1 or 2, characterized in that

the specified section (T, B) encompasses an area which is enclosed by a separating plane (T) and a boundary plane (B), wherein

the boundary plane (B) contains and/or touches the edge of the backrest (1, 10, 11) facing the camera (30) and runs substantially parallel to the side surface (112) of the backrest (11), and

the separating plane (T) vertically transects the vehicle seat (14).

4. Method according to one of the preceding claims, characterized in that

the boundary plane (B) is substantially supported by the z-axis and the x-axis of the vehicle coordinate system,

wherein the position of the boundary plane (B) in the y direction is dependent on the mounting position of the vehicle seat (14) in the Y direction.

5. Method according to one of the preceding claims, characterized in that

the separating plane (T) is substantially supported by the z-axis and the x-axis of the vehicle coordinate system,

wherein the separating plane (T) vertically transects the vehicle seat and is located in the y direction between the headrest (12) and the boundary plane (B).

6. Method according to one of the preceding claims, characterized in that

the specified section (T, B) is selected such that the specified section (T, B) registered by the camera (30) is at least not totally obscured by an object (6).

7. Method according to one of the preceding claims, characterized in that

a specified section (T1, B1) encompasses an area which is enclosed by two planes (T1, B1) which contain the side surface (121) of the headrest that faces the camera (1).

8. Method according to one of the preceding claims, characterized in that

a distance image with image points is produced from the image information, said image being transformed into a preferably Cartesian system of coordinates (xyz) having a fixed vehicle reference whose coordinate axes point into the longitudinal axis of the vehicle (x-axis), at right angles to the longitudinal axis of the vehicle (y-axis), and vertically (z-axis), wherein

the specified section (T, B) is subdivided into volume pixel elements (13), an initial specified value being assigned (deactivated) to the volume pixel elements (13) in which no image points of the element (10, 11, 12, 14) and/or of the object (6) were registered and at least a further specified value being assigned (activated) to the volume pixel elements (13) in which at least one image point of an element (10, 11, 12, 14) and/or object (6) was registered.

9. Method according to one of the preceding claims, characterized in that

the values assigned to the volume pixel elements (13) are compared with at least one specified volume pixel element pattern which for example has an element and/or the combination of an object with an element.

10. Method according to one of the preceding claims, characterized in that

the volume pixel elements (13) are formed as a rectangular solid or cube with a specified edge length, preferably between around 2 (two) and 4 (four) cm, which border each other directly without overlapping, wherein

the surfaces of the volume pixel elements (13) are in each case formed parallel to the coordinate axes of the coordinate system (xyz).

11. Method according to one of the preceding claims, characterized in that

the specified section (T, B, T1, B1, T2, B2) is subdivided in the y direction into volume pixel element planes, preferably into 3 and 5 volume pixel element planes, wherein

the volume pixel elements (13) which in these volume pixel element planes contain another specified value (are activated) are projected into an evaluation plane, wherein

the evaluation plane only registers the volume pixel elements (13) for further evaluation that lie within the scope of the adjustment range (S1, S2, S3) of the element (10, 11, 12, 14, 40, 41) of the vehicle (1).

12. Method according to one of the preceding claims, characterized in that

the evaluation plane only registers those volume pixel elements (13) whose other values have remained unchanged at least a specified number of times, preferably about 3 (three) times, during a specified period of lapsed time, preferably a period covered by about the last 20 (twenty) measurements.

13. Method according to one of the preceding claims, characterized in that

an element (10, 11, 12, 14, 40, 41) is adjustable with reference to the vehicle and an object (6) is movable/removable and/or mobile with reference to the vehicle.

14. Method according to one of the preceding claims, characterized in that

the surface shape and/or surface position of the element is determined with the aid of the specified geometric dimensions of the element (10, 11, 12, 14, 40, 41) and of the position of the characteristic feature of the element (10, 11, 12, 14, 40, 41), and that the volume distribution of the object (6) located in front of the element (10, 11, 12, 14, 40, 41) is determined with the aid of optical detection methods.

15. Method according to one of the preceding claims, characterized in that

from the characteristic feature the position of the start point (111, 122) of an element (11, 12), preferably the backrest (11) and/or headrest (12), is determined which corresponds to the position of the volume pixel element (13) in the evaluation plane that has been provided with an additional value (has been activated) and is preferably located closest to the camera (30) and the vehicle roof (2).

16. Method according to one of the preceding claims, characterized in that

the mean is taken across several measurements of the position of the start point (111) and/or of the characteristic feature of the element (11, 12).

17. Method according to one of the preceding claims, characterized in that

from the position of the start point (111) at least an approximate mean is taken

of the inclination of the backrest (11), and/or

of the height position of the vehicle seat (14), and/or

of the adjustment position of the vehicle seat (1, 10) along the seat positioning track (20), after which

the surface shape and/or the surface position of the vehicle seat (1) and/or of the backrest and/or of the seat base (10) and/or of the headrest (12) is calculated.

18. Method according to one of the preceding claims, characterized in that

an element (10, 11, 12, 14, 40, 41) is implemented as a vehicle seat (14) and/or seat base (10) and/or backrest (11) and/or headrest (12) and/or window frame (40) and/or adjustable window pane (41) and/or vehicle mirror.

19. Method according to one of the preceding claims, characterized in that

an object (6) takes the form of an occupant (6) and/or a part of an occupant'"'"'s body, such as an upper body area, an arm, a leg and/or a head, and/or of a child car seat and/or of an object on the vehicle seat (14).

20. Method according to one of the preceding claims, characterized in that

a further boundary plane (B1) is specified by the Y position of the side surface (121) of the headrest (12) facing the camera (30) and another separating plane (T1) vertically transects the headrest, and that

the position of the headrest (12) is determined from the volume pixel elements (13) that have been provided with an additional value that are located between the boundary plane (B1) and the separating plane (T1).

21. Method according to one of the preceding claims, characterized in that

by means of multiple volume pixel elements (13) of the evaluation plane that have been provided with an additional value, which register the backrest (11) and/or the headrest (12), a virtual straight line is determined which crosses the plane of the seat positioning track (20) in a position point from which the overall position of the vehicle seat (14) along the seat positioning track (20) is determined.

22. Method according to one of the preceding claims, characterized in that

the position of the seatrest is determined from a preceding measuring cycle or a standard inclination with an angle of inclination of 105°

to 110°

is applied if the position point cannot be determined during a measuring cycle.

23. Method according to one of the preceding claims, characterized in that

the position of the vehicle seat is measured with the aid of a seat position sensor (50), preferably a Hall sensor (20), wherein the measurement resolution is at least limited to at least two, preferably three sections (A, B, C) along the seat positioning track (A, B, C).

24. Method according to one of the preceding claims, characterized in that

the weight and/or the weight distribution of an object (6) on the element is determined by a weight sensor as additional information.

25. Method according to one of the preceding claims, characterized in that

at least one switch is attached to a mechanical and/or electrical seat adjustment device which in each case supplies information about any change to the seat height, to the angle of inclination of the backrest and/or to the seat position along the seat positioning track (20) (S1, S2, S3).

26. Method according to one of the preceding claims, characterized in that

the information from the evaluation plane is compared with the grayscale image registered by the camera.

27. Method according to one of the preceding claims, characterized in that

the surface shape (contour) of the vehicle seat is determined when the vehicle seat is occupied.

28. Method according to one of the preceding claims, characterized in that

the specified section (T, B) encompasses an area which is enclosed by two planes (T, B) which contain the top edge, which faces the camera (30), of the window frame (40) which supports the window pane (41),

the top edge of the window pane (41) is optically registered by the camera,

the closing motion of the window pane (41) is dependent on the position of the vehicle occupant and on the position of the top edge of the window pane (41).

29. Method according to one of the preceding claims, characterized in that

the position of the plane of the side surface of the vehicle seat (14) which faces the camera (30) is specified by the specified position of the vehicle seat in the Y direction and remains substantially constant when the vehicle seat is moved in the y direction.

30. Method according to one of the preceding claims, characterized in that

the registered image information takes the form of image data which is transmitted by the camera to the evaluation unit and processed there.

31. Method according to one of the preceding claims, characterized in that

the camera (30) is implemented as a 3D camera for registering stereo images and preferably equipped with two image sensors.

32. Method according to one of the preceding claims, characterized in that

the aperture angle of the camera (30) is approximately 100×

90 degrees.

33. Method for detecting an object in a vehicle, in particular for occupant protection systems, characterized in that a camera (30) registers an image area in the interior of the vehicle (1) and the image information registered by the camera (1) is processed by an evaluation unit (31) according to one of the method steps specified in the preceding method claims.