Camera calibration method and calibration system

Abstract

Disclosed is a camera calibration method that eliminates the need to acquire pattern images from multiple viewpoints in order to acquire optical parameters of a camera, and that improves robustness with respect to imaging position deviation. The disclosed calibration method includes: acquiring pattern images obtained by using cameras 10, 12, 14 to image calibration boards 30, 32 comprising two pattern boards disposed spaced apart from each other; acquiring principal point coordinates indicating the position of a lens optical axis on an imaging element on the basis of patterns displayed on the two pattern boards captured in the pattern images; and using the principal point coordinates and relative positional relationships between the cameras 10, 12, 14 and the calibration boards 30, 32 to acquire optical parameters including the focal length and not including the principal point coordinates, on the basis of the pattern images.

Description

Camera Calibration Method and Calibration System

[0001] The present invention relates to a camera calibration method and a calibration system, and particularly to a calibration method and a calibration system suitable for calibration of a camera used in a sensing system mounted on a moving object such as an automobile.

[0002] In recent years, driving support systems and safety support systems that are mounted on vehicles such as automobiles and assist drivers in driving operations have become widespread. In these driving support systems and safety support systems, for example, processing such as obstacle recognition is performed using an image captured by a camera mounted on a vehicle. In such a system, it is necessary to obtain the distance to an obstacle with high accuracy, and it is desirable that the optical parameters of the camera used for the recognition processing have correct values and are calibrated.

[0003] For example, Non-Patent Document 1 discloses that a pattern drawn on a plane is photographed with a camera to be calibrated, and calibration of internal parameters and external parameters of the camera is performed based on the photographed image. In this method, pattern images obtained by photographing a pattern of a known size from a plurality of different viewpoints with a camera to be calibrated are used, and the optical parameters of the camera are obtained from the correspondence between two-dimensional feature points on the pattern image and three-dimensional points in the world coordinate system. The calibration method disclosed in Non-Patent Document 1 is generally widely used, for example, in OpenCV, which is an open-source software library developed for computer vision.

[0004] 「A Flexible New Technique for Camera calibration」(IEEE Transactions on Pattern Analysis and Machine Intelligence, 22(11):1330-1334, 2000.)

[0005] Calibration methods such as those described in Non-Patent Document 1 allow for accurate calibration of the camera's optical parameters. However, calibration requires capturing known patterns from multiple different viewpoints, which may be inconvenient depending on the environment in which the optical parameters are acquired.

[0006] For example, when acquiring optical parameters from a camera mounted on a vehicle, a board with a known pattern drawn on it is used, but it becomes necessary to photograph the board while moving either the board or the vehicle. Moving a vehicle is usually not easy. On the other hand, in order to efficiently acquire optical parameters, a sufficiently large pattern is required, resulting in a large board. Furthermore, if a robust material that does not deform due to temperature and humidity differences is used for the board to prevent shrinkage or distortion of the drawn pattern due to temperature and humidity changes, the board will be heavy and difficult to handle easily. In addition, with a wide-angle camera that can capture a wide area, the range of board handling also increases.

[0007] In the method described in Non-Patent Document 1, the viewpoint from which the pattern is photographed is not specified, so the movement path of the board does not need to be determined, but there is a possibility that the reproducibility of the acquired parameters will be lost. On the other hand, if the movement path is determined, the equipment for calibration will become large and complex. The acquired optical parameters include external parameters related to the position and orientation of the camera, and internal parameters that indicate the camera's inherent characteristics. For example, the former includes a total of six parameters: three-dimensional coordinates and rotations with each of the three-dimensional axes as the central axis, while the latter includes four parameters: focal length in the x and y directions and principal point coordinates, as well as five or more parameters that express distortion. In the method described in Non-Patent Document 1, it is possible to acquire all of these optical parameters, but the more optical parameters to acquire, the lower the accuracy becomes, and there are drawbacks such as obtaining local optima.

[0008] This invention has been made in view of the above-mentioned problems, and its purpose is to eliminate the need to acquire pattern images from multiple viewpoints when acquiring optical parameters in camera calibration, thereby improving robustness against shifts in shooting position.

[0009] A camera calibration method according to the present invention, in one preferred embodiment, includes the process of capturing a pattern image by photographing a first pattern board on which a first pattern is displayed and a second pattern board on which a second pattern is displayed, which is positioned behind the first pattern board in the optical axis direction of the camera as viewed from the camera to be calibrated and spaced apart from the first pattern board; the process of obtaining the principal point coordinates, which are the position of the optical axis of the lens on the imaging surface of the image sensor provided by the camera, based on the pattern image; and the process of obtaining the focal length of the camera based on the pattern image using the principal point coordinates, the distance between the first pattern board and the second pattern board, and the distance between the camera and the first pattern board.

[0010] Furthermore, in one preferred embodiment, the calibration system of the present invention includes a calibration board comprising a first pattern board on which a first pattern is displayed, and a second pattern board spaced apart from the first pattern board in the direction in which the normal to the display surface of the first pattern extends, on which a second pattern is displayed; a camera interface that is communicatively connected to a camera to be calibrated and acquires a pattern image obtained by photographing the calibration board with the camera; and a parameter acquisition unit configured to acquire the principal point coordinates of the camera based on the first and second patterns depicted in the pattern image acquired via the camera interface, and to acquire optical parameters other than the principal point coordinates, including the focal length of the camera, based on the pattern image using the acquired principal point coordinates and the relative positional relationship between the camera and the calibration board.

[0011] According to the present invention, the acquisition of pattern images from multiple viewpoints during camera calibration becomes unnecessary, improving robustness against shifts in shooting position. Other novel features of the present invention and the technical problems solved thereby will become apparent from the description and drawings herein.

[0012] This is a schematic diagram showing an example of the configuration of the calibration equipment. This is a schematic diagram showing the configuration of the calibration board. This is a schematic diagram showing the configuration of the calibration board. This is a flowchart showing the processing flow during calibration. This is a schematic diagram showing the positional relationship between the vehicle and the calibration board when the calibration board is photographed. This is a schematic diagram showing the positional relationship between the vehicle and the calibration board when the calibration board is photographed. This is a flowchart showing the processing flow executed by the parameter acquisition unit. This is a schematic diagram of the pattern image taken when the rear pattern board is hidden. This is a schematic diagram of the pattern image taken when a pattern is displayed on the rear pattern board. This is a schematic diagram showing the pattern image of the rear pattern board generated from the pattern images shown in Figures 8 and 9. This is a graph showing the relationship between the parameter calculation error and the positional relationship between the pattern board and the vehicle. This is a schematic diagram showing a modified arrangement of the calibration board.

[0013] Hereinafter, representative embodiments of the present invention will be described with reference to the drawings. Note that the embodiments and drawings described below are illustrative examples for illustrating the present invention, and have been omitted or simplified as appropriate for clarity of explanation. Furthermore, please note that the positions, sizes, shapes, and extents of the components shown in the drawings may not necessarily accurately represent them, in order to facilitate understanding of the invention.

[0014] Figure 1 is a schematic diagram showing an example of the configuration of a calibration facility to which the present invention is applied.

[0015] Cameras 10, 12, and 14 are cameras mounted on the vehicle 100 that are subject to calibration. Each camera 10, 12, and 14 is configured with a lens and an image sensor (not shown). The lenses and image sensors of each camera 10, 12, and 14 may be different in type, characteristics, and performance, or they may be the same.

[0016] Cameras 10, 12, and 14 are installed on the right side of vehicle 100 to photograph the area to the right of vehicle 100. Cameras 10 and 14 are installed so that their optical axes are approximately parallel to each other and pointing to the right of vehicle 100. Camera 12 is installed with its optical axis offset from that of cameras 10 and 14. The camera installation positions and shooting directions shown here are merely examples; cameras can, of course, be installed at any position on vehicle 100, such as the left side or front / rear, and the shooting direction can also be any direction.

[0017] The calibration device 20 is installed, for example, on a car manufacturer's production line or repair shop together with calibration boards 30 and 32, and is capable of calibrating cameras 10, 12 and 14 while they are mounted on a vehicle 100. The calibration device 20 includes a parameter acquisition unit 200 that calculates optical parameters based on images captured by cameras 10, 12 and 14, a camera interface (camera IF) 210 to which cameras 10, 12 and 14 are connected, and a board interface (board IF) 220 to which calibration boards 30 and 32 are connected.

[0018] The calibration device 20 can be implemented using a general-purpose computing device, such as a personal computer, server computer, or tablet device, which has an interface that can be used as a camera IF 210 or board IF 220. In this case, the parameter acquisition unit 200 can be implemented in software by executing a program stored in memory using a processor such as a CPU (Central Processing Unit) or GPU (Graphics Processing Unit) provided by the computing device. The calibration device 20 may be configured as a dedicated device, or it may be implemented in hardware using an FPGA (Field-Programmable Gate Array) or ASIC (Application-Specific Integrated Circuit).

[0019] Calibration boards 30 and 32 display patterns captured by cameras 10, 12, and 14 during calibration. As shown in Figures 2 and 3, calibration boards 30 and 32 include pattern boards 300 and 310, which are spaced apart from each other in the direction in which the normals of the pattern display surfaces extend. Pattern board 310 is positioned behind pattern board 300 (towards the rear from the perspective of cameras 10, 12, and 14) during calibration. Furthermore, pattern boards 300 and 310 are superimposed on each other when viewed from the optical axis direction of cameras 10, 12, and 14, which are used for capturing images, and are positioned so that their pattern display surfaces are approximately parallel to each other.

[0020] The pattern board 300 is made of a transparent, flat board such as an acrylic sheet, and is positioned on the vehicle 100 side of the pattern board 310 during calibration, so that it appears to overlap with the pattern board 310 when viewed from the vehicle 100. The pattern board 300 has parallel vertical lines drawn at equal intervals, for example, as shown in Figure 2.

[0021] The pattern board 310 uses a display device such as a liquid crystal display or an organic EL display and is connected to the calibration device 20, displaying a pattern under control from the calibration device 20. On the pattern board 300, parallel horizontal lines are displayed at equal intervals, as shown in Figure 2, under control from the calibration device. Since the pattern board 300 is made of a transparent material, the pattern displayed on the pattern board 310 can be viewed through the pattern board 300 during calibration. In addition, the background of the pattern displayed on the pattern board 310 is, for example, plain white, so as not to mix with the pattern on the pattern board 300.

[0022] The distance Z between pattern board 300 and pattern board 310 should preferably be 30% or more of the distance between the front pattern board 300 and the camera position during calibration, in order to prevent large parameter acquisition errors caused by measurement errors, as will be described later.

[0023] In this embodiment, parallel lines in the vertical direction are used as the pattern displayed on pattern board 300, and parallel lines in the horizontal direction are used as the pattern displayed on pattern board 310, but these may be swapped. Also, the lines displayed on pattern boards 300 and 310 do not have to be vertical and horizontal in this way; for example, they may be rotated by 45 degrees, or they may be rotated by any angle as long as their extension directions are different.

[0024] During calibration, it is desirable that the calibration boards 30 and 32 be positioned so as to face the camera being calibrated. In other words, it is desirable that the optical axis of the camera be approximately parallel to the normal of the display surface of the board. As shown in Figure 1, in this embodiment, during calibration, the calibration board 30 is positioned so as to face the cameras 10 and 14, and the calibration board 32 is positioned so as to face the camera 12. Then, the calibration board 30 is used to calibrate the cameras 10 and 14, and the calibration board 32 is used to calibrate the camera 12.

[0025] Figure 4 is a flowchart showing the processing flow during the calibration of cameras 10, 12, and 14. During calibration, for example, the car is moved forward (from bottom to top in the diagram) while each camera captures images of the calibration board 30 or 32 and acquires optical parameters.

[0026] At the start of the calibration process, the vehicle 100 is moved to a position where one of the multiple cameras faces one of the multiple calibration boards. In this embodiment, when the vehicle 100 is moved in the direction of the calibration boards 30 and 32, the camera 10 first faces the calibration board 30. Therefore, the vehicle 100 is moved to a predetermined position where the camera 10 and the calibration board 30 can be faced directly (step S100).

[0027] After the vehicle is positioned in the designated location, the calibration device 20 instructs the camera 10 via the camera IF 210 to photograph the calibration board 30, and acquires the image captured by the camera 10. The parameter acquisition unit 200 switches the display state of the pattern board 310 via the board IF 220, as shown in Figure 2, between a state where the pattern is displayed and a state where only a white background image is displayed without the pattern, as shown in Figure 3, and controls the camera 10 to take pictures in each state. The parameter acquisition unit 200 acquires the optical parameters of the camera 10 using the image taken with the pattern on the pattern board 310 displayed and the image taken with the pattern on the pattern board 310 erased. The acquired optical parameters are sent to the camera 10 via the camera IF 210 and set as parameters of the camera 10 (step S110).

[0028] After the calibration of camera 10 is completed, the vehicle 100 is moved to a position where the next camera is directly facing one of the calibration boards. Here, as shown in Figure 5, the vehicle 100 is moved to a position where camera 12 is directly facing calibration board 32 (step S120). Then, in the same manner as in step S110, the calibration of camera 12 is performed and the acquired optical parameters are set in camera 12 (step S130).

[0029] Subsequently, the vehicle 100 is moved in the same manner as shown in Figure 6 until the camera 14 is directly facing the calibration board 30 (step S140), and the optical parameters of the camera 14 are acquired and set to the camera 14 in the same manner as in step S110 (step S150).

[0030] If other cameras are installed on the vehicle 100, the same process as in steps S100 and S110 is carried out, and the process ends when the calibration of all cameras is completed.

[0031] Figure 7 is a flowchart showing the flow of processing performed by the parameter acquisition unit 200 in the process of acquiring the optical parameters of camera 10 in step S110. Here, we will explain the acquisition of optical parameters of camera 10, but the optical parameters of camera 14 can be acquired in the same manner. Furthermore, optical parameters of camera 12 can also be acquired in the same manner as camera 10, except that the calibration board used is calibration board 32.

[0032] When the vehicle 100 is moved to a predetermined position where the camera 10 is directly facing the calibration board 30, processing by the parameter acquisition unit 200 is started. Whether the vehicle 100 has moved to the predetermined position can be detected, for example, by a sensor installed on the production line if it is a production line, and the processing of the parameter acquisition unit 200 can be started using this detection signal as a trigger. Alternatively, the system may be configured so that processing is started manually when an operator confirms that the vehicle 100 has moved to the fixed position.

[0033] When the parameter acquisition unit 200 starts processing, it uses the board IF 220 to display only a white background on the pattern board 310 behind the calibration board 30, hiding the pattern and bringing it to the state shown in Figure 3 (step S200).

[0034] The parameter acquisition unit 200 instructs the camera 10 to take a picture via the camera IF 210 and acquires the captured image. If cameras 10, 12, and 14 are wide-angle cameras capable of shooting at a wide angle of view, the lens has distortion aberration, so the acquired image will be like image 400 shown in Figure 8, for example (step S210).

[0035] Next, the parameter acquisition unit 200 displays a pattern on the pattern board 310 (step S320), instructs the camera 10 to take a picture again, and acquires the captured image. The image acquired at this time will look like image 410 shown in Figure 9 due to lens distortion (step S330). Subsequently, the parameter acquisition unit 200 acquires an image of horizontal parallel lines displayed on the pattern board 310, such as image 420 shown in Figure 10, based on images 400 and 410. Image 420 can be acquired, for example, by subtracting the pixel values ​​of image 400 from the pixel values ​​of image 410 (step S340).

[0036] Next, the parameter acquisition unit 200 acquires principal point coordinates (u1, v1) indicating the position of the lens optical axis on the image sensor (the position where the imaging plane of the image sensor and the optical axis of the lens intersect) based on images 400 and 420. Here, the horizontal coordinate axis of the image is referred to as the u-axis, and the vertical coordinate axis as the v-axis. Note that the u-axis and v-axis do not necessarily have to be axes aligned with the Bayer angle of the image sensor. The principal point coordinates can be obtained by estimating the position of a virtual straight line that would have been captured as a straight line without the influence of distortion if there was a straight line between two curves in images 400 and 420 where the direction of curvature changes, and obtaining the coordinates of the point where the virtual straight lines in image 400 and image 420 intersect.

[0037] The parameter acquisition unit 200 first estimates the position of a virtual line estimated on the image 400 and obtains the principal point coordinates u1 from the position of the estimated virtual line in the u-axis direction on the image. Specifically, for example, the error between each curve and its first-order approximation is used as an index of curvature, and the position where the index is minimized is acquired as a line, and its position in the u-axis direction is acquired as u1 (step S250). Subsequently, the parameter acquisition unit 200 estimates the position of a virtual line estimated on the image 420 based on the image 420, similar to step S250, and obtains the principal point coordinates v1 from the position of the estimated virtual line in the v-axis direction on the image (step S260).

[0038] Next, the parameter acquisition unit 200 acquires the distance wn on the image between two adjacent lines sandwiching the acquired principal point coordinates based on the image 400. This distance wn corresponds to the distance between the patterns (parallel lines) displayed on the pattern board 300. Since the influence of distortion increases as the distance from the optical axis increases, it is desirable that the distance wn between the two lines be acquired around the coordinates (u1, v1) obtained in steps S250 and S260 (step S270). Subsequently, in the same manner as step S270, based on the image 420, the distance wf on the image between two adjacent lines sandwiching the principal point coordinates is acquired. The distance wf corresponds to the distance between the patterns (parallel lines) displayed on the pattern board 310. Similar to the distance wn, it is desirable that the distance wf be acquired around the principal point coordinates (u1, v1) (step S280).

[0039] The parameter acquisition unit 200 acquires the focal length of the camera 10 using the distances wf and wn obtained in the above-described process. The focal length fc can be obtained using the distances wf and wn as

[0040]

[0041] Here, as described above, Z is the distance between the pattern board 300 and the pattern board 310 of the calibration board 30. Also, W1 is the distance between the lines of the pattern displayed on the rear pattern board 310, and W2 is the distance between the lines of the pattern displayed on the front pattern board 300 of the calibration board 30.

[0042] Also, the distance L from the camera 10 to the front pattern board 300 of the calibration board 30 can be obtained in the same manner as

[0043]

[0044] as

[0045] Here, since the distance from the camera 10 to the calibration board 30 (pattern board 300 or pattern board 310) is not included in Equation (1), it is not necessary to accurately maintain the distance between the vehicle 100 and the calibration board 30 during calibration. Further, since wn and wf are measured around the optical axis (principal point coordinates) obtained as described above, the focal length can be calculated while suppressing the influence of distortion.

[0046] Note that, as shown in FIG. 11, the calculation error of the parameter increases as the ratio of Z to L decreases. This is because the calculation error of the parameter due to the measurement error of wn and wf increases as the ratio of Z to L decreases. From FIG. 11, it is understood that the increase in the calculation error of the parameter becomes larger particularly in the region where the ratio of Z to L is less than 30%. Therefore, as described above, it is desirable that the distance Z between the pattern board 300 and the pattern board 310 on the calibration boards 30 and 32 is 30% or more of the distance L between the front pattern board 300 and the camera (step S290).

[0047] Thus, by first obtaining the principal point coordinates, even if there is a relative displacement between the lens and the imaging element and a deviation in the stop position of the vehicle during calibration, these can be tolerated and the optical parameters can be obtained.

[0048] In the process described above, the acquisition of principal point coordinates and focal length was explained as optical parameters. However, if the camera has a sufficient field of view, the image captured in steps S210 and S230 will include not only the pattern of the calibration board facing the camera, but also the pattern of the calibration board that is not facing the camera. Since the calibration board that is not facing the camera is captured in the image at a position away from the principal point coordinates, it is possible to use this to acquire some distortion parameters. For example, in Figure 1, the distortion parameters of camera 10 can be acquired using calibration board 32. Similarly, in the state shown in Figure 5, the distortion parameters of camera 12 can be acquired using calibration board 30, and in the state shown in Figure 6, the distortion parameters of camera 14 can be acquired using calibration board 32. In particular, it is easy to know the relative positional relationship between each calibration board and each camera, and using this information, along with the principal point coordinates and focal length acquired using the calibration board facing the camera, it is possible to acquire the axially symmetric component of the distortion parameters.

[0049] In this embodiment, calibration board 30 and calibration board 32 are positioned such that the normals of the board surfaces of their respective pattern boards intersect with the calibration boards 30 and 32 on the vehicle 100 side. Calibration boards 30 and 32 can be swapped from their arrangement in this embodiment, for example, as shown in Figure 12. In this case as well, the optical parameters of each camera can be obtained using the method described above. However, when obtaining distortion parameters using calibration boards that are not directly facing each other, arranging the calibration boards so that the normals of the board surfaces of the multiple calibration boards intersect on the vehicle 100 side, as in this embodiment, allows for more effective use of the patterns of the calibration boards that are not directly facing each other.

[0050] Furthermore, it is possible to obtain distortion parameters by photographing the calibration board from a position other than the one where the vehicle 100 is stopped in order to obtain the principal point coordinates and focal length. For example, at the stationary position shown in Figure 1, the calibration boards 30 and 32 can be photographed by cameras 12 and 14, and the distortion parameters of cameras 12 and 14 can be obtained using the images taken at this time. In this case, since images of known patterns within a wider field of view can be obtained, it is possible to improve the accuracy of calculating distortion parameters.

[0051] If the vehicle's stopping position is significantly off, causing the relative positional relationship between the camera and the calibration board to fall outside a predetermined range, the optical parameters may not be acquired as intended. In such cases, the calibration device 20 may detect that it could not calculate the optical parameters as intended, or it may detect the amount of positional deviation of the vehicle 100's stopping position and output information indicating that the relative positional relationship is outside a predetermined range, prompting the user to correct the vehicle 100's stopping position. One example of a case where optical parameters cannot be acquired as intended is when a virtual straight line cannot be obtained from the image captured by the camera. Furthermore, the amount of positional deviation of the vehicle's stopping position can be estimated by, for example, displaying unique marks on the calibration boards 30 and 32 and obtaining an approximate amount of positional deviation from the coordinate positions of the marks on the captured image.

[0052] As described above, according to this embodiment, the principal point coordinates of the camera can be obtained first using images of patterns drawn on two pattern boards that have normals substantially parallel to the optical axis direction of the camera and are spaced apart in the optical axis direction, and other optical parameters such as the focal length can be obtained using the obtained principal point coordinates. This makes it possible to reduce the number of pattern captures required during camera calibration and improve the robustness of the capture position.

[0053] Although the present invention has been described above using representative embodiments as examples, the present invention is not limited thereto and can be implemented in various ways without departing from the spirit of the invention as described in the claims.

[0054] For example, in the embodiment described above, a display device is used on the rear pattern board, but the rear pattern board may be, for example, an acrylic board with a pattern drawn on it, which cannot be switched on or off. In this case, for example, lines in the vertical and horizontal directions can be extracted from an image like the one shown in Figure 9, in which the patterns of two pattern boards are superimposed, and images like those shown in Figures 8 and 10 can be obtained to acquire parameters.

[0055] Furthermore, a transparent display device that allows the rear pattern board to be seen may be used as the front pattern board. By using display devices for both pattern boards and alternately displaying and hiding the patterns on each display device, it becomes possible to directly acquire images of both patterns shown in Figures 8 and 10.

[0056] Furthermore, in the embodiments described above, the background of the rear pattern board was described as white, but it is not limited to white; any color that provides a clear contrast with the patterns drawn on the front and rear pattern boards is acceptable.

[0057] The embodiments described above are explained in detail for the purpose of clearly illustrating the present invention, and are not necessarily limited to those comprising all the configurations described.

[0058] 10, 12, 14... Camera, 20... Calibration device, 30, 32... Calibration board, 200... Parameter acquisition unit, 210... Camera IF, 220... Board IF, 100... Vehicle

Claims

1. A camera calibration method comprising: a process of acquiring a pattern image by photographing a first pattern board on which a first pattern is displayed and a second pattern board on which a second pattern is displayed, which is positioned behind the first pattern board in the optical axis direction of the camera as viewed from the camera to be calibrated, and spaced apart from the first pattern board; a process of acquiring the principal point coordinates, which are the position of the optical axis of the lens on the imaging surface of the image sensor provided by the camera, based on the pattern image; and a process of acquiring the focal length of the camera based on the pattern image using the principal point coordinates, the distance between the first pattern board and the second pattern board, and the distance between the camera and the first pattern board.

2. The camera calibration method according to claim 1, wherein the camera has distortion aberration in the lens.

3. The camera calibration method according to claim 1, wherein the process of capturing the pattern image is performed while the camera is mounted on a vehicle.

4. The camera calibration method according to claim 1, wherein the first pattern board and the second pattern board are arranged such that the display surface of the first pattern and the display surface of the second pattern are parallel to each other.

5. The camera calibration method according to claim 1, wherein the first pattern board is a transparent plate-shaped member on which the first pattern is displayed.

6. The camera calibration method according to claim 1, wherein at least one of the first pattern board and the second pattern board is a display device capable of switching between displaying and not displaying a pattern.

7. The camera calibration method according to claim 1, wherein the first pattern includes a plurality of parallel lines, and the second pattern includes a plurality of parallel lines that extend in different directions from the lines of the first pattern.

8. The camera calibration method according to claim 3, wherein the vehicle is equipped with a plurality of cameras that are subject to calibration, each having different optical axis directions, and the first pattern board and the second pattern board are arranged in a plurality corresponding to the optical axis directions of each of the plurality of cameras, such that the optical axis direction and the normals of the display surfaces of the first pattern and the second pattern are parallel.

9. The camera calibration method according to claim 1, wherein the distance between the first pattern board and the second pattern board is 30% or more of the distance between the camera and the first pattern board.

10. The camera calibration method according to claim 1, wherein the process of acquiring the pattern image includes a process of detecting that the relative positional relationship between the camera and the first pattern board and the second pattern board is outside a predetermined range.

11. A calibration system comprising: a calibration board comprising: a first pattern board on which a first pattern is displayed; a second pattern board positioned spaced apart from the first pattern board in the direction in which the normal of the display surface of the first pattern extends, on which a second pattern is displayed; a camera interface that is communicatively connected to a camera to be calibrated and acquires a pattern image obtained by photographing the calibration board with the camera; and a parameter acquisition unit configured to acquire the principal point coordinates of the camera based on the first pattern and the second pattern depicted in the pattern image acquired via the camera interface, and to acquire optical parameters other than the principal point coordinates, including the focal length of the camera, based on the pattern image using the acquired principal point coordinates and the relative positional relationship between the camera and the calibration board; and a calibration device.

12. The calibration device further comprises a board interface to which the calibration board is connected, at least one of the first pattern board and the second pattern board is a display device capable of switching the display of patterns on and off, and the parameter acquisition unit is configured to control the display of patterns on at least one of the pattern boards via the board interface, the calibration system according to claim 11.

13. The calibration system according to claim 11, further comprising another calibration board having a third pattern board on which a third pattern is displayed, a fourth pattern board that is positioned on which a fourth pattern is displayed, and another calibration board that is positioned in a direction different from that of the calibration device during calibration.

14. The calibration system according to claim 13, wherein the parameter acquisition unit is configured to acquire distortion parameters of the camera based on the third pattern or the fourth pattern captured in the pattern image, using the acquired principal point coordinates and focal length.

15. The calibration system according to claim 11, which is configured to detect, based on the pattern image, that the relative positional relationship between the camera and the calibration board is outside a predetermined range, and to output information indicating that the relative positional relationship is outside the predetermined range.

Images