JP2010107348A - Calibration target and in-vehicle calibration system using it - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a calibration target for preventing an adjacent vehicle calibration target from being erroneously detected. <P>SOLUTION: The calibration target includes: a marker board frame 101; two markers 102 disposed within the frame 101, separated from the frame 101, and having the same color as the frame 101 and a predetermined area; a marker board 104 having a color different from the frame 101 and the markers 102, and including an extra-marker region 103 between the frame 101 and the markers 102; and an inverted marker board 105 disposed symmetric to the marker board with reference to an optical axis of an in-vehicle imaging apparatus. The inverted marker board 105 includes: a marker board 108 and markers 109 having the same color as the extra-marker region 103; and an extra-marker region 110 having the same color as the markers 102 or the marker board 101. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a calibration target for calibrating a mounting state of a camera for imaging a surrounding situation of a running vehicle with respect to the vehicle and a calibration for an in-vehicle camera.

    In recent years, in the field of in-vehicle cameras, there is an increasing need for in-vehicle camera calibration technology that determines the installation position and installation angle of an in-vehicle camera with respect to a road surface. For example, in order to superimpose a line indicating the distance from the bumper on the camera image behind the vehicle at the time of parking, it is necessary to perform camera calibration in advance.

  Many proposals have been made for a method of using the camera calibration result in realizing the white line detection function and the obstacle detection function. Here, a conventional on-vehicle camera calibration system will be described with reference to the drawings.

  FIG. 5 is a plan view showing a conventional on-vehicle camera calibration system and its use state.

  In FIG. 5, a vehicle-mounted camera 502 is provided at the rearmost part of a vehicle 501. On the road surface behind the vehicle 501, a calibration target 503 having a plurality of reference position patterns indicating positions serving as calibration references is arranged. The reference position pattern of the calibration target 503 indicates a reference position for camera calibration, and a pattern in which feature points are easily extracted visually is usually used. In the figure, the checkerboard pattern is different in the color of adjacent squares.

  As a calibration method for the in-vehicle camera 502, first, the positions of the vehicle 501 and the calibration target 503 are accurately aligned. For example, the centers of the vehicle 501 and the calibration target 503 are aligned. The coordinates are three-dimensional coordinates of the actual road surface (hereinafter referred to as “world coordinates”). For example, in FIG. 5, the intersection of the rear end of the vehicle 501 and the central axis is the origin, the road surface is the X and Y axes, and the height direction is the Z axis. After positioning the vehicle 501 and the calibration target 503, the calibration target 503 is imaged by the in-vehicle camera 502. The captured image is displayed on a monitor (not shown), and a square vertex on the image is input as a reference position by an input means such as a mouse. The coordinates of the vertex are expressed by two-dimensional coordinates on the xy plane (hereinafter referred to as “image coordinates”). The image coordinates of the vertices may be automatically obtained by pattern recognition.

  Thereafter, the correspondence between the image coordinates of the square vertices and the actual world coordinates is performed. Such association between image coordinates and world coordinates is performed for a plurality of vertices, and parameters for camera calibration (hereinafter simply referred to as “camera parameters”) are calculated using these coordinates as input.

Details of the above processing are described in Non-Patent Document 1, for example.
“A Versatile Camera Calibration Technology for High-Accuracy 3D Machine Vision Metrology Using Off-the-Shelf TV Cameras and Lenses,” Roger. Tsai, IEEE Journal of Robotics and automation Vol. RA-3. No. 4 August 1987, pp 323-344

  Such in-vehicle camera calibration is usually performed on a production line at the time of vehicle production. However, in consideration of vehicle productivity, a parallel line is installed in a limited space to perform calibration. . Therefore, it is inevitable that when the calibration target is imaged, the adjacent target is simultaneously imaged in addition to the target for the own vehicle. Then, if the target for the adjacent vehicle is erroneously detected and the camera calibration of the own vehicle is performed using the position information, the camera application includes a large error in the calculated installation position and installation angle of the camera. End up.

  The present invention provides a calibration target that can prevent erroneous detection of an adjacent calibration target when performing calibration of an in-vehicle camera mounted on each vehicle in a state where a plurality of vehicles are arranged in parallel. It aims at providing the calibration system for vehicle-mounted cameras.

  In order to achieve the above object, the present invention is directed to a calibration target used for calibration of an in-vehicle imaging device capable of imaging the periphery of a vehicle, and the calibration target includes a first marker board frame, A first marker board which is inside the first marker board frame, has the same color as the first marker board frame and does not contact the first marker board frame, and has a predetermined area; A marker board having a frame and a color different from each of the first markers and including a first marker board frame and a first outside marker area between the first markers; An inversion marker board is provided on the opposite side of the marker boat with the optical axis of the imaging device as a boundary, and the inversion marker board has the same color as the first marker outside region. Of including a marker board frame and a second marker, and a first marker or the first marker board frame and the second marker extracellular region of the same color.

  With the above-described configuration, it is possible to provide an in-vehicle camera calibration system that can prevent erroneous detection of a calibration target for an adjacent vehicle in the calibration of an in-vehicle camera that is performed on vehicles arranged in parallel in a factory.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
(First embodiment)
FIG. 1 shows a calibration target according to the first embodiment of the present invention. As shown in the dotted line frame in FIG. 1, the calibration target first includes a marker board frame 101, a marker 102, an outside marker area 103, and a marker board 104.

  The marker board frame 101 forms a closed section with a line segment having a specific color and thickness.

  Each marker 102 is inside the marker board frame 101 and does not contact the marker board frame 101 and has the same color as the marker board frame 101 and a predetermined area. Here, the number of markers is two in the illustrated example, but is determined by the number of camera parameters of the in-vehicle camera to be obtained by calibration.

  The marker outside region 103 has a color different from that of the marker board frame 101 and both markers 102, and exists between the marker board frame 101 and both markers 102. Further, the area of the marker outside region 103 is selected in advance so as to be different from the total area of the markers 102. In the present embodiment, the area of the marker outside region 103 is larger.

  The marker board 104 includes a marker board frame 101, both markers 102, and an outside marker area 103.

  The calibration target further includes an inversion marker board 105 disposed on the opposite side of each configuration of the marker board 104 with the optical axis 107 of the imaging device 106 attached to the vehicle and capable of imaging the surroundings of the vehicle as a boundary. Preferably, the inverted marker board 105 is arranged in line symmetry with the marker board 104 with respect to the optical axis 107. However, in the inverted marker board 105, the marker board frame 108 and the marker 109 are the same color as the marker outside area 103, and the marker outside area 110 is the same color as the marker 102 or the marker board frame 101.

  According to the calibration target described above, by providing the marker outside areas 103 and 110 having different color schemes from the markers 102 and 109 outside the markers 102 and 109 necessary for calibration, the marker boards 104 and 105 are in any environment. Even if installed, the markers 102 and 109 are not buried in the surrounding background color.

  Furthermore, by providing the marker board frames 101 and 108 having different colors outside the marker outer regions 103 and 110, the marker outer regions 103 and 110 are not buried in the surrounding background color. As a result, the markers 102 and 109 in the areas can be reliably detected by extracting the areas outside the markers 103 and 110 regardless of the environment.

Further, by arranging the marker boards 104 and 105 with the reversed color arrangement on both the left and right sides of the optical axis 107 of the imaging device 106, adjacent markers even in an environment where calibration is performed on a plurality of parallel vehicles. The boards 104 and 105 can be easily identified because the color schemes are always different.
(Second Embodiment)
Next, the configuration of the in-vehicle camera calibration system according to the second embodiment of the present invention is shown in FIG.

  In the block diagram of FIG. 2, the calibration target 201 uses the calibration target shown in the first embodiment. Therefore, description of the detailed configuration of the calibration target 201 is omitted. The in-vehicle camera calibration system is installed in a vehicle and includes an imaging device 202, a marker detection unit 203, a marker centroid calculation unit 204, a camera parameter calculation unit 205, and a marker coordinate storage unit 206.

  This target 201 is imaged by the imaging device 202 to be calibrated. The marker detection unit 203 processes the captured image and extracts a marker region in the image.

  A specific processing flow of the marker detection unit 203 will be described with reference to the flowchart of FIG.

  In FIG. 3, first, binarization processing is performed on an input image captured by the imaging device 202 (step S301). Since the input image is a color image, the image is binarized by comparing its luminance value and saturation value with a reference threshold value, for example. At this time, as a method for determining the binarization threshold, for example, a method of setting a fixed threshold or a method of discriminant analysis can be used. However, the threshold value must be determined so that the marker board frame and the marker in the calibration target have the same pixel value and a pixel value different from the area outside the marker by this binarization processing.

  Next, a board labeling process is executed (step S302). This is a process of grouping regions by adding the same label to the pixels having the same pixel value to be connected to the image in which the pixel values are classified into binary values by the binarization process. As a specific method, there are a 4-neighbor labeling process in which four neighborhoods to be processed are labeled as connected components, and an 8-neighbor labeling process in which labeling is performed with 8 neighborhoods to be processed as connected components. There are other methods using a line block method and a histogram, which are general methods in image processing, and are all effective for the board labeling processing of the present invention. As a result of the board labeling process, the marker board frame, the area outside the marker, and the individual markers are grouped as different areas in the image.

  Here, as described in the first embodiment, the calibration target of the present invention is arranged so that the color arrangement is reversed on the left and right of the optical axis of the imaging device 202. In addition, since the area outside the marker in the marker board and the marker are color-coded so as to have different pixel values in the previous binarization process, using this information, the area that exists on the left or right of the center of the image, For example, for the marker board on the right side of the target arranged as shown in FIG. 1, when the area outside the marker is classified as black, for example, a group of black areas having a large number of pixels is included. What is necessary is just to select as an area | region outside a marker (step S303).

  If the area outside the marker including the marker necessary for calibration is selected in this way, as shown in FIG. 4, the colored area of the marker board in which the adjacent color is inverted becomes a marker, and the number of internal pixels is small. It will not be selected by mistake. Even if the size of the marker is large and it is selected by mistake, an erroneous selection can be found by the processing following the selection of the area outside the marker. Specifically, the same labeling process as that of the marker board is performed for the markers in the area outside the marker selected by the marker labeling process (step S304). By this process, the number of markers included in each marker board is individually recognized.

  If this number matches the predetermined number of markers (YES in step S305), the marker board including the area outside the marker is detected as a marker to be processed, and if it does not match (NO in step S305). ), Re-select the area outside the marker as the marker board for the adjacent vehicle.

  As described above, by processing by the marker detection unit 203, only the marker necessary for calibration can be reliably detected from the captured image.

  Next, the marker centroid calculation unit 204 calculates the marker centroid coordinates on the image detected by the marker detection unit 203. Specific methods include using a coordinate average value of all the pixels included in the marker, obtaining the center position from the end points (upper end, lower end, left end, right end) on the marker image, and further, if the marker is a polygon. In the case of, there is a method of obtaining the barycentric coordinates using the coordinates of the vertices and diagonal lines.

  Based on the marker center-of-gravity coordinates calculated in this way and the three-dimensional coordinates of the marker center in the actual environment where the marker board is actually stored, stored in the marker coordinate storage unit 206, the external parameters of the imaging device 202 are stored. Is calculated by the camera parameter calculation unit 205. A specific method of parameter calculation is detailed in Non-Patent Document 1. Note that the number of markers in the marker board may be appropriately determined according to the number of camera parameters calculated at this time.

  The in-vehicle camera calibration system according to the present invention has an effect that it is possible to prevent erroneous detection of a calibration target for an adjacent vehicle in the in-vehicle camera calibration performed for vehicles arranged in parallel in a factory. This is useful as a vehicle-mounted camera calibration method for calibrating the installation state of a camera installed in a vehicle to image the vehicle surroundings during traveling.

The schematic diagram which shows the calibration target in the 1st Embodiment of this invention The block diagram of the calibration system for vehicle-mounted cameras in the 2nd Embodiment of this invention The flowchart which shows the process of the marker detection part 203 in the 2nd Embodiment of this invention. The figure for demonstrating operation | movement of the marker detection part 203 in the 2nd Embodiment of this invention. The figure explaining the conventional calibration system for in-vehicle cameras

Explanation of symbols

101, 108 Marker board frame 102, 109 Marker 103, 110 Outside marker area 104 Marker board 105 Reverse marker board 106 Imaging device 107 Optical axis

Claims (4)

A calibration target used for calibration of an in-vehicle imaging device capable of imaging the surroundings of a vehicle,
The calibration target is
A first marker board frame;
A first marker that is inside the first marker board frame, has the same color as the first marker board frame, and has a predetermined area without being in contact with the first marker board frame;
A marker board having a color different from that of the first marker board frame and the first marker, and including a first outside marker area located between the first marker board frame and the first marker; Prepared,
The calibration target further includes an inversion marker board disposed on the opposite side of the marker boat with the optical axis of the in-vehicle imaging device as a boundary,
The reverse marker board is
A second marker board frame and a second marker of the same color as the area outside the first marker;
A second marker outside region of the same color as the first marker or the first marker board frame,
Calibration target. The area of the first marker outer region is different from the total area of the second markers,
The calibration target according to claim 1. The area of the first marker outer region is larger than the sum of the areas of the second markers,
The calibration target according to claim 2. A calibration target according to claim 1;
An imaging device attached to the vehicle for imaging the surroundings of the vehicle;
A marker detection unit for detecting a marker necessary for calibration from an image captured by the imaging device;
A marker centroid calculating unit that calculates the centroid coordinates on the image with respect to the marker detected by the marker detecting unit;
A marker coordinate storage unit for preliminarily storing the three-dimensional coordinates of the calibration target placed in a real environment;
A camera parameter calculation unit that calculates an external parameter including an attachment angle and a position of the imaging device to the vehicle based on the marker gravity center coordinates calculated by the marker gravity center calculation unit and the coordinates stored in the marker coordinate storage unit; Prepare
Car camera calibration system.