JP2010121986A - Calibration target and calibration system for on-board camera - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a calibration target allowing travel of a vehicle between lines in adjacent manufacturing lines. <P>SOLUTION: The calibration target 8 includes a first slope 801B that is used on a predetermined installation surface for calibrating an on-board camera 2A2 and inclines at a predetermined angle to the installation surface, and first marks 81d-81f that are drawn on the first slope 801B and are used for calibrating the on-board camera 2A2. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a calibration target and a calibration system used for calibration when a camera installed in a vehicle deviates from a predetermined mounting position and mounting direction.

A conventional in-vehicle camera calibration system calibrates an in-vehicle camera attached to a vehicle by determining a rotation matrix of the camera and determines success or failure. At this time, the calculation is performed based on the coordinates in the reference coordinate system of the mark portions of the calibration targets arranged at least at two different locations. (For example, refer to Patent Document 1).
Japanese Patent Laid-Open No. 2008-131250 (page 18, FIG. 1)

  For an in-vehicle camera system that is mounted on a vehicle and displays an image around the vehicle, such as an image from an upper viewpoint, calibration may be performed in the vehicle manufacturing process. Further, in order to increase productivity, there is a manufacturing process in which manufacturing lines A and B are arranged in parallel as shown in FIG.

  Further, in the manufacturing process, since the operation of the vehicles α1 and α2 is restricted according to the line, the usable space is limited, and the calibration targets 8a to 8d used for calibration also have height and size corresponding to the space restriction. It is required to adjust to the size.

  In such a manufacturing process, when the calibration system for the camera 2A1 of the vehicle α1 is operated by using the mark portions of the calibration targets arranged at two different places 8a and 8c, as shown in FIG. The calibration target 8b to be present is within the angle of view of the in-vehicle camera 2A1.

  Therefore, the calibration system of the vehicle α1 has a problem of erroneously detecting the calibration target.

  As a proposal for solving this, there is a method of installing a screen-like shield γ between adjacent calibration targets 8a and 8b as shown in FIG. 12, but in the manufacturing process as shown in FIG. Will be hindered.

  An object of the present invention is to provide a calibration target and a calibration system that do not hinder the operation of a vehicle in the manufacturing process of adjacent vehicles.

  In order to achieve the above object, the present invention provides a calibration target used on a predetermined installation surface for calibration of an in-vehicle camera, wherein the calibration target is at a predetermined angle with respect to the installation surface. And a first slope portion drawn on the first slope portion and used for calibration of the in-vehicle camera.

  By providing the calibration target with the first slope portion that is inclined at a predetermined angle as in the above configuration, the vehicle can get over the calibration target, thereby preventing the operation of the vehicle from being hindered. It becomes possible.

(First embodiment)
Hereinafter, a calibration system for an in-vehicle camera according to an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing an overall configuration of a calibration system 1 for an in-vehicle camera according to the first embodiment.

  As shown in FIG. 1, a calibration system 1 inputs a plurality of cameras 2 (illustrated four units 2A to 2D) attached to a vehicle, a display device 6, and image information captured by each camera 2. And a camera image control unit 7 that performs a predetermined process on the image information, generates a video signal indicating an overhead image overlooking the vehicle periphery, and outputs the video signal to the display device 6.

  As shown in FIG. 2, the camera 2 is attached so as to photograph the periphery of the vehicle α (illustrated by a dotted line). In this description, the camera 2A is attached to the rear end of the vehicle α, the camera 2B is attached to the left side of the vehicle α, the camera 2C is attached to the front end of the vehicle α, and the camera 2D is attached to the right side of the vehicle α.

  FIG. 2 further shows a range captured by the cameras 2A to 2D in the peripheral region of the vehicle.

  The camera 2A images the situation of the imaging range CAM0 behind the vehicle α. Similarly, the camera 2B images the imaging range CAM1 on the left side of the vehicle α, the camera 2C images the imaging range CAM2 in front of the vehicle α, and the camera 2D images the imaging range CAM3 on the right side of the vehicle α. Here, in order to cover the entire periphery of the vehicle α with the cameras 2A to 2D, the imaging range CAM0 overlaps with the imaging ranges CAM1 and CAM3, respectively, and the imaging range CAM2 also overlaps with the imaging ranges CAM1 and CAM3, respectively. Thereafter, the overlapping portion of the imaging ranges CAM0 and CAM1 is the overlap region F01, the overlapping portion of the imaging ranges CAM1 and CAM2 is the overlap region F12, the overlapping portion of the imaging ranges CAM2 and CAM3 is the overlap region F23, and the imaging range CAM3 The overlapping portion between CAM0 and CAM0 is called an overlap region F30.

  In FIG. 2, the display device 6 displays the video indicated by the input video signal on a screen such as a cathode ray tube or a liquid crystal.

  The camera image control unit 7 includes a memory 3 (four in the figure, 3A to 3D), an image processing unit 4, and a nonvolatile memory 5. Memory 3A-3D consists of RAM etc., and records the picked-up image data of cameras 2A-2D. The image processing unit 4 includes a DSP or FPGA, and performs image processing by taking in pixel data recorded in the memories 3A to 3D. The nonvolatile memory 5 is referred to by the image processing unit 4.

  The memory 3A is connected to the camera 2A and temporarily holds video information shot by the camera 2A. Similarly, the memories 3B to 3D are connected to the cameras 2B to 2D, and temporarily hold video information from the cameras 2B to 2D. These pieces of video information are referred to by the image processing unit 4.

As shown in FIG. 3, the image processing unit 4 includes an image extraction unit 41, a correction value calculation unit 42, a correction image generation unit 43, and an image video superimposition unit 44.

  The image extraction unit 41 extracts a captured image held therein from any of the memories 3A to 3D. The extracted captured image is referred to by the correction value calculation unit 42 and the correction image generation unit 43.

  The correction value calculation unit 42 calculates a correction amount using the captured image extracted by the image extraction unit 41. Specifically, the correction value calculation unit 42 extracts a plurality of calibration point designs (see FIG. 9) from the currently referenced captured image, and coordinates information on the image of each figure and the three-dimensional space. The parameter correction amount for the camera 2 is calculated so that the coordinate information is correctly associated.

  The corrected image generation unit 43 corrects the captured image extracted by the image extraction unit 41 with the correction amount calculated by the correction value calculation unit 42. Specifically, the corrected image generation unit 43 refers to the parameter correction amount calculated by the correction value calculation unit 42 and corrects the captured image received from the image extraction unit 41. The corrected image generation unit 43 refers to the nonvolatile memory 5 and corrects the distortion of the captured image after correction. Such a corrected photographed image is output to the image video superimposing unit 44.

  The image video superimposing unit 44 superimposes a predetermined image on the corrected captured image generated by the corrected image generating unit 43. More specifically, the image video superimposing unit 44 superimposes, for example, a guideline indicating the distance from the vehicle or a portion without video (for example, an image showing the host vehicle) on the output captured video of the corrected image generating unit 43. To do.

  FIG. 4 is a diagram illustrating the positional relationship between the camera 2A1 attached to the vehicle α1 and the calibration target 8. As shown in FIG. FIG. 4 particularly shows the positional relationship between the camera 2A1 for imaging the rear of the vehicle α1 and the inclined surfaces 801A and 801B of the calibration target 8 captured by the camera 2A2 for imaging the rear of the vehicle α2. Since the lines A and B are adjacent to each other, the slope portion 801B of the calibration target 8 of the manufacturing line B adjacent to the manufacturing line A is within the imaging range CAM0a of the camera 2A1.

  FIG. 5 is a diagram showing a detailed configuration of the slope portions 801A and 801B of the calibration target 8. As shown in FIG. In FIG. 5, the calibration target 8 includes slope portions 801A and 801B.

  Further, mark portions 81a to 81f used for calibration are respectively drawn on the surfaces of the slope portions 801A and 801B of the calibration target 8. The mark parts 81a to 81f are figures to be extracted by the image extracting part 41, and may be composed of a plurality of figures, for example, a figure such as a circle.

  Here, the area in which the mark portions 81a to 81c can be drawn on the slope portion 801A of the calibration target 8 in the present embodiment will be described.

The area S1 in which the mark portion in FIG. 5 can be drawn is represented by the depth D1, the height H1, and the width W of the calibration target 8,
S1 = D1 × (H1 / sin (Atan (H1 / (W / 2))) (Formula 1)
Thus, as shown in FIG. 6, the mark portion in the camera video can be shown large.

  Here, H1 needs to be less than 140 mm so that the vehicle does not hinder the operation.

FIG. 7 shows the installation position point P of the calibration target 8. An angle formed by the optical axis of the camera 2A1 with respect to the xy plane on the xz plane is defined as θ1. The angle formed by the inclined surface 801B of the calibration target not used for the camera 2A1 with respect to the xy plane is also θ1. At this time, the position of the point P is composed of the installation height H3 of the camera 2A1 and the right triangle formed by the distance L from the coordinate origin to the point P in FIG. 12, the height H2 of the calibration target 8, and the calibration target 801B. Due to the similarity of right triangles,
L = (H3 / H2) * (W2 / 2) (Equation 2) is determined.

  Thereby, it becomes possible to prevent the slope portion 801B of the calibration target from being erroneously detected as an invisible portion from the camera 2A1.

  The operation of the calibration using the calibration target 8 configured as described above will be described with reference to the flowchart of FIG.

  Any of the cameras 2A to 2D captures an image so that the calibration target 8 enters the image at a predetermined position (step 1001). The photographed image is stored in the memory 3 subsequent to the camera 2 that has photographed this time (step 1002). In the image processing unit 4, the image extraction unit 41 determines a search range on a predetermined two-dimensional image from the acquired captured image (step S 1003), and a mark portion drawn on a predetermined slope portion of the calibration target 8. Is extracted (step S1004).

Now, the number of extracted mark parts is n, and i is a natural number from 1 to n. The correction value calculator 42 calculates the center-of-gravity coordinates MA _i (X _i , Y _i ) of the design of each extracted calibration point (step S1005), and calculates the three-dimensional coordinate system from the MA _i (X _i , Y _i ). The point CA _i (Xw _i , Yw _i , Zw _i ) is converted (step S1006). After calculating the points CA _i in the design of all the calibration points, the correction value calculation unit 42 for each i has a three-dimensional coordinate value TA _i (Xt _i , Yt _i , Zt) where the center of gravity coordinates CA _i are originally located. Camera parameters for correcting the point CA _i are calculated by optimization processing so as to be closest to _i ) (step S1007). The above processing is performed for n mark portions.

In the calibration system using the calibration target 8 configured as described above, the size of the mark portion of the calibration target is secured without hindering the operation of the vehicle by suppressing the height, and the image of the mark portion. While facilitating extraction, it is possible to prevent erroneous detection of calibration targets of adjacent production lines.
(Second Embodiment)
Next, a second embodiment of the present invention will be described.

  FIG. 9 is a diagram showing a calibration target 8a according to the second embodiment. In FIG. 9, two on-vehicle cameras 2A1 and 2A2 on two different vehicle production lines are shown, and a calibration target that forms one slope for one line is shown. Specifically, the in-vehicle cameras 2A1 and 2A2 use slope portions 801A and 801B that form slopes with respect to the xy plane, respectively.

More specifically, the slope 801A of the calibration target that draws the mark portions 81a to 81c of the calibration target 8a is installed so as to fall within the imaging range CAM0a of the in-vehicle camera 2A1, and the calibration that draws the mark portion of the calibration target. The target 801B is installed so as to fall within the shooting range CAM0b of the in-vehicle camera 2A2. However, unlike the first embodiment, the incident line of the light beam for the in-vehicle camera 2A1 is the camera 2A1 out of the group of lines projected by dropping the incident line of the light beam for the in-vehicle camera 2A1 on the xy plane of FIG. The line segments formed by the point Q that intersects with the center of the two mark portions 81a and 81c that take the maximum value and the minimum value on the y-axis coordinate on the calibration target 801A are denoted by β1 and β2, respectively. Installed at a position where β1 = β2.

  As a result, as shown in FIG. 10, compared to the first embodiment, the mark portion 81a that is farthest from the camera 2A1 on the y-axis in FIG. 9 can be shown larger.

  In the second embodiment, the height H3 of the in-vehicle camera 2A1 in FIG. 7 and the angle formed by the optical axis of the camera 2A1 with respect to the xy plane from the value of θ1, H2 is not necessarily the maximum height. When there is no need to do so, in addition to the above installation, the angle of the slope portion of the calibration target can be adjusted to make the mark portion appear larger.

  The calibration target according to the present invention enables the operation of the vehicle between processes by suppressing the height without erroneously detecting the design of the calibration point in the adjacent manufacturing process, and the design of the calibration point design. This is useful for calibration of an in-vehicle camera that is required to ensure the size.

1 is a block diagram showing an overall configuration of a calibration system 1 according to a first embodiment The figure which illustrates the calibration system 1 in a vehicle The block diagram which shows the detailed structure of the image process part 4 The figure which shows the example of arrangement | positioning of the calibration target 8 in 1st Embodiment. The figure which shows the detailed structure of the calibration target 8 The figure which illustrates the picked-up image of the calibration target 8 in 1st Embodiment The figure which shows the installation position of the calibration target 8 in 1st Embodiment. The flowchart which shows the procedure of the calibration which concerns on 1st Embodiment The figure which shows the example of arrangement | positioning of the calibration target 8A in 2nd Embodiment. The figure which illustrates the picked-up image of the calibration target 8 in 2nd Embodiment The figure which shows the example of arrangement | positioning of calibration target 8a, 8b in an adjacent manufacturing line The figure which shows the example of prevention of an erroneous detection in the calibration targets 8a and 8b in an adjacent manufacturing line. A diagram illustrating the operation of a vehicle between adjacent production lines

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Calibration system 2, 2A, 2B, 2C, 2D, 2A1, 2A2 Car-mounted camera 3 Memory 4 Image processing part 41 Image extraction part 42 Correction value calculation part 43 Correction image generation part 44 Image video superimposition part 5 Non-volatile memory 6 Display device 7 Camera image control units 8, 8a Calibration targets 801A, 801B Slope portions 81a, 81b, 81c, 81d, 81e, 81f Calibration target mark portions

Claims (7)

A calibration target used on a predetermined installation surface for calibration of an in-vehicle camera,
The calibration target is
A first slope portion inclined at a predetermined angle with respect to the installation surface;
A first mark portion drawn on the first slope portion and used for calibration of the in-vehicle camera;
Calibration target. A second inclined surface portion inclined at the predetermined angle with respect to the installation surface and connected to the first inclined surface portion;
A second mark portion drawn on the second slope portion and used for calibration of another in-vehicle camera;
The calibration target according to claim 1. A second mark portion used for calibration of another in-vehicle camera is further provided on a surface different from the first slope portion,
The second mark part is invisible from an in-vehicle camera using the first mark part.
The calibration target according to claim 1. The angles formed by the optical axis of the in-vehicle camera and the second inclined surface portion with the installation surface are set to be the same.
The calibration target according to claim 3. The calibration target is installed at a position satisfying (H3 / H2) × (W2 / 2) with reference to a point where the installation position of the in-vehicle camera is projected on the installation surface.
The H2 is the height of the calibration target, the H3 is the height at which the in-vehicle camera is installed, and the W2 is the width of the calibration target.
The calibration target according to claim 4. The height of the first slope portion is less than 140 mm on the basis of the installation surface.
The calibration target according to claim 1. An in-vehicle camera that images the calibration target according to claim 1;
An extraction unit that extracts the first mark portion in the calibration target from a captured image acquired by the in-vehicle camera;
A coordinate calculation unit for calculating coordinates of a predetermined position of the first mark portion extracted by the extraction unit;
A coordinate converter that converts the coordinates calculated by the coordinate calculator into predetermined three-dimensional coordinates;
A calibration system comprising: an optimization unit that optimizes camera parameters of the in-vehicle camera based on the three-dimensional coordinates obtained by the coordinate conversion unit.