A camera intrinsic parameter evaluation method, device, equipment and medium

Abstract

The embodiment of the specification discloses a camera internal parameter evaluation method, device, equipment and medium. Including: obtaining the first distance between the first camera optical center position when the to-be-evaluated camera collects the first image for the calibration pattern and the second camera optical center position when the second image is collected; obtaining the world coordinate set of the target vertex in the calibration image, the first pixel coordinate in the first image and the second pixel coordinate in the second image; using the world coordinate set of the target vertex, the first pixel coordinate and the second pixel coordinate, generating the first estimated world coordinate of the first camera optical center position and the second estimated world coordinate of the second camera optical center position, calculating the second distance between the first estimated world coordinate and the second estimated world coordinate, and generating the camera internal parameter evaluation result for the to-be-evaluated camera according to the first distance and the second distance. The scheme can improve the accuracy of the camera internal parameter evaluation result.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer vision, and in particular to a camera intrinsic parameter evaluation method and device, equipment and a medium. BACKGROUND

[0002] The existing camera intrinsic parameter evaluation method mainly uses the re-projection error to evaluate the camera intrinsic parameter, and when the re-projection error is small, it means that the camera intrinsic parameter is good.

[0003] However, the re-projection error result is mainly affected by the camera intrinsic parameter, the camera extrinsic parameter, and the distance between the camera and the calibration object, and other factors. Therefore, when the re-projection error is small, it does not necessarily mean that the camera intrinsic parameter is good, and it may be that the re-projection error is small due to the camera extrinsic parameter or the distance between the camera and the calibration object. Therefore, the accuracy of the camera intrinsic parameter evaluation result obtained by using the re-projection error is not high.

[0004] Therefore, how to evaluate the camera intrinsic parameter without the influence of other factors to improve the accuracy of the camera intrinsic parameter evaluation result has become a technical problem to be solved. SUMMARY

[0005] The camera intrinsic parameter evaluation method, device, equipment and medium provided by the embodiments of the present application solve the technical problem of low accuracy of the camera intrinsic parameter evaluation result in the prior art when the re-projection error is used to evaluate the camera intrinsic parameter.

[0006] To solve the above technical problems, the embodiments of the present application are implemented as follows:

[0007] The camera intrinsic parameter evaluation method provided by the embodiments of the present application comprises,

[0008] Obtaining a first distance between a first optical center position of a camera to be evaluated when collecting a first image of a calibration pattern and a second optical center position when collecting a second image.

[0009] According to the world coordinate set of the target vertex in the preset world coordinate system, the first pixel coordinate of the target vertex in the first image and the second pixel coordinate of the target vertex in the second image in the preset pixel coordinate system, and the camera intrinsic parameter of the camera to be evaluated, the conversion relationship between the world coordinate system and the pixel coordinate system is used to generate a camera extrinsic parameter set.

[0010] According to the world coordinate set and the camera extrinsic parameter set, the conversion relationship between the world coordinate system and the camera coordinate system is used to generate a first estimated world coordinate of the first camera optical center position and a second estimated world coordinate of the second camera optical center position.

[0011] generate a second distance between the first camera optical center position and the second camera optical center position according to the first estimated world coordinate and the second estimated world coordinate.

[0012] generate the evaluation result of the camera intrinsic parameter of the camera to be evaluated by using the first distance and the second distance.

[0013] Preferably, the camera extrinsic parameter set is generated by using the conversion relationship between the world coordinate system and the pixel coordinate system according to the world coordinate set of the target vertex in the calibration pattern, the first pixel coordinate and the second pixel coordinate of the target vertex in the first image and the second image in the first image in the preset pixel coordinate system, and the camera intrinsic parameter of the camera to be evaluated, including:

[0014] The target vertex includes a first target vertex and a second target vertex.

[0015] generate the first camera extrinsic parameter at the first camera optical center position by using the conversion relationship between the world coordinate system and the pixel coordinate system according to the first world coordinate of the first target vertex in the world coordinate set, the first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameter.

[0016] generate the second camera extrinsic parameter at the second camera optical center position by using the conversion relationship between the world coordinate system and the pixel coordinate system according to the second world coordinate of the second target vertex in the world coordinate set, the second pixel coordinate of the second target vertex in the preset pixel coordinate system, and the camera intrinsic parameter.

[0017] Preferably, the calibration pattern is a pattern composed of a preset number of two-dimensional code patterns.

[0018] Before the first camera extrinsic parameter at the first camera optical center position is generated by using the conversion relationship between the world coordinate system and the pixel coordinate system according to the first world coordinate of the first target vertex in the world coordinate set, the first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameter, it further includes:

[0019] the first image is processed by using a two-dimensional code recognition algorithm to obtain the first position information of the first two-dimensional code pattern in the first image.

[0020] The first pixel coordinate of the first target vertex in the first two-dimensional code pattern is determined according to the first position information of the first two-dimensional code pattern and the preset pixel coordinate system.

[0021] The second camera extrinsic parameter at the second camera optical center position is generated according to the second world coordinate of the second target vertex in the world coordinate set, the second pixel coordinate of the second target vertex in the preset pixel coordinate system, and the camera intrinsic parameter, by using the conversion relationship between the world coordinate system and the pixel coordinate system.

[0022] The second image is processed by using a two-dimensional code recognition algorithm to obtain second position information of a second two-dimensional code pattern in the second image.

[0023] The second pixel coordinate of the second target vertex in the second two-dimensional code pattern is determined according to the second position information of the second two-dimensional code pattern and the preset pixel coordinate system.

[0024] Preferably, the calibration pattern is a checkerboard pattern.

[0025] Before the first camera extrinsic parameter at the first camera optical center position is generated according to the first world coordinate of the first target vertex in the world coordinate set, the first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameter, by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method further comprises:

[0026] First position information of a first checkerboard grid in the first image is determined.

[0027] The first pixel coordinate of the first target vertex in the first checkerboard grid is determined according to the first position information of the first checkerboard grid and the preset pixel coordinate system.

[0028] Before the second camera extrinsic parameter at the second camera optical center position is generated according to the second world coordinate of the second target vertex in the world coordinate set, the second pixel coordinate of the second target vertex in the preset pixel coordinate system, and the camera intrinsic parameter, by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method further comprises:

[0029] Second position information of a second checkerboard grid in the second image is determined.

[0030] The second pixel coordinate of the second target vertex in the second checkerboard grid is determined according to the second position information of the second checkerboard grid and the preset pixel coordinate system.

[0031] Preferably, the first estimated world coordinate of the first camera optical center position and the second estimated world coordinate of the second camera optical center position are generated according to the world coordinate set and the camera extrinsic parameter set, by using the conversion relationship between the world coordinate system and the camera coordinate system, comprising:

[0032] According to the first world coordinate and the first camera extrinsic parameter, a conversion relationship between a world coordinate system and a camera coordinate system is used to generate the first estimated world coordinate.

[0033] According to the second world coordinate and the second camera extrinsic parameter, a conversion relationship between a world coordinate system and a camera coordinate system is used to generate the second estimated world coordinate.

[0034] Preferably, the generating the evaluation result of the camera intrinsic parameter of the camera to be evaluated according to the first distance and the second distance comprises:

[0035] A difference between the first distance and the second distance is calculated.

[0036] It is judged whether a ratio between an absolute value of the difference and the second distance is less than a preset threshold value, to obtain a judgment result.

[0037] If the judgment result indicates that the ratio is less than the preset threshold value, it is determined that the evaluation result of the camera intrinsic parameter of the camera to be evaluated is qualified.

[0038] If the judgment result indicates that the ratio is not less than the preset threshold value, it is determined that the evaluation result of the camera intrinsic parameter of the camera to be evaluated is unqualified.

[0039] Preferably, the second distance is any one of a Euclidean distance, a Manhattan distance, a Chebyshev distance, and a Minkowski distance.

[0040] A camera intrinsic parameter evaluation device comprises,

[0041] An acquisition module is configured to acquire a first distance between a first optical center position when a camera to be evaluated collects a first image for a calibration pattern and a second optical center position when the camera to be evaluated collects a second image.

[0042] A first generation module is configured to, according to a world coordinate set of a target vertex in the calibration pattern in a preset world coordinate system, first pixel coordinates and second pixel coordinates of the target vertex in the first image and the second image in a preset pixel coordinate system, and a camera intrinsic parameter of the camera to be evaluated, generate a camera extrinsic parameter set by using a conversion relationship between a world coordinate system and a pixel coordinate system.

[0043] A second generation module is configured to, according to the world coordinate set and the camera extrinsic parameter set, generate a first estimated world coordinate of the first camera optical center position and a second estimated world coordinate of the second camera optical center position by using a conversion relationship between a world coordinate system and a camera coordinate system.

[0044] generate a second distance between the first camera optical center position and the second camera optical center position according to the first estimated world coordinate and the second estimated world coordinate.

[0045] generate a camera intrinsic evaluation result of the camera to be evaluated by using the first distance and the second distance.

[0046] A camera intrinsic evaluation device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the above method.

[0047] A computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement the above method.

[0048] At least one of the embodiments provided in the specification can achieve the following beneficial effects:

[0049] In the scheme, the camera intrinsic is used to generate the camera extrinsic under the condition that the distance between the camera and the calibration object is determined, so that the influencing factors of the camera extrinsic itself are eliminated, and then the influencing factors of the camera intrinsic are used to replace the influencing factors of the camera extrinsic. The first distance used in the scheme is the distance between the first camera optical center position and the second camera optical center position obtained by measuring or other means. The second distance calculated and generated by the camera extrinsic, the first pixel coordinate and the second pixel coordinate is mainly affected by the camera intrinsic. Therefore, when the camera intrinsic is evaluated according to the first distance and the second distance, only the camera intrinsic has one influencing factor, that is, the closer the positional relationship between the first distance and the second distance, the only effect that can be proved is caused by the camera intrinsic. Therefore, the closer the first distance and the second distance, the better the evaluation result of the camera intrinsic, and the evaluation result obtained is indeed better, that is, the accuracy of the evaluation result is higher. Therefore, the camera evaluation result obtained by using the scheme can mainly reflect the advantages and disadvantages of the camera intrinsic, so as to improve the accuracy of the camera intrinsic evaluation result. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the specification or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0051] Figure 1 A flowchart of a camera intrinsic evaluation method provided by an embodiment of the specification;

[0052] Figure 2 A schematic diagram of a position relationship between a camera to be evaluated and a calibration pattern provided for an embodiment of the present specification;

[0053] Figure 3 A structural schematic diagram of an evaluation device corresponding to a camera intrinsic provided for an embodiment of the present specification; Figure 1

[0054] Figure 4 A structural schematic diagram of an evaluation device corresponding to a camera intrinsic provided for an embodiment of the present specification. Figure 1 DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of one or more embodiments of the present specification clearer, the technical solutions of one or more embodiments of the present specification will be described clearly and completely below in combination with specific embodiments of the present specification and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present specification, rather than all the embodiments. Based on the embodiments in the present specification, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of one or more embodiments of the present specification.

[0056] The technical solutions provided by the embodiments of the present application will be described in detail below in combination with the drawings.

[0057] Figure 1 A flowchart of a camera intrinsic evaluation method provided for an embodiment of the present specification. From the perspective of the program, the execution subject of the flowchart can be a device for evaluating the camera intrinsic, or an application program carried on the device for evaluating the camera intrinsic. As shown in Figure 1 the flowchart can include the following steps:

[0058] Step 102: Obtain a first distance between a first optical center position when a first image is collected by a camera to be evaluated for a calibration pattern and a second optical center position when a second image is collected.

[0059] ​​In the embodiments of this specification, the optical center positions of the first and second cameras can be different optical center positions of the same camera when capturing the first and second images of the calibration object. The optical center positions of the first and second cameras can be symmetrically located to the left and right front of the calibration object, respectively, with the extension of the centerline of the calibration object as the axis of symmetry. Alternatively, they can be located to the left and right front of the calibration object in other ways. This solution does not specifically limit the exact positions of the optical center positions of the first and second cameras in front of the calibration object. The distance between the optical center positions of the first and second cameras, the distance between the camera and the calibration object when the camera is at the first optical center position, and the distance between the camera and the calibration object when the camera is at the second optical center position are not specifically limited; appropriate distances can be selected according to the actual scenario.

[0060] Figure 2 This is a schematic diagram illustrating the positional relationship between the camera and the calibration object proposed in this invention. Figure 2 As shown, the first camera can be positioned to the right front of the calibration object, and the second camera can be positioned to the left front of the calibration object, with the first and second cameras aligned. The distance between the first and second camera positions can be 160 cm, and the perpendicular distance from the calibration object to the line connecting the first and second camera positions can be 60 cm. Using existing technology, the optical position of the first camera can be calculated from its position (the optical center position of the first camera), and the optical position of the second camera can be calculated from its position (the optical center position of the second camera). It should be noted that... Figure 2 The diagram is merely a schematic representation of the positional relationship between the camera and the calibration object; it does not represent a specific limitation on the positional relationship between the camera and the calibration object.

[0061] In the embodiments of this specification, the first distance between the optical center positions of the first camera and the second camera can be obtained by existing technical means, and this solution does not specifically limit the method of obtaining the first distance. For ease of understanding, this solution proposes an implementation method for obtaining the first distance: firstly, the distance L1 between the camera housing and the camera body at the optical center positions of the first and second cameras is obtained using existing measurement methods; then, the distance L2 between the camera optical center and the camera housing is obtained using internal structural data of the camera; finally, the distance calculated using the formula S=L1+2L2 is the first distance between the optical center positions of the first and second cameras. It should be noted that the method for obtaining the first distance proposed in this solution is only for the convenience of those skilled in the art and does not constitute a limitation on the method of obtaining the first distance.

[0062] Step 104: generating a camera extrinsic parameter set by using a conversion relationship between a world coordinate system and a pixel coordinate system according to the world coordinate set of the target vertex in the calibration pattern in the preset world coordinate system, the first pixel coordinate of the target vertex in the first image and the second pixel coordinate in the second image in the preset pixel coordinate system, and the camera intrinsic parameter of the camera to be evaluated.

[0063] In the embodiments of the present specification, the calibration pattern can be a calibration pattern composed of some micro patterns, and the specifications of the micro patterns are not specifically limited, and any pattern that can be used to calibrate the camera intrinsic parameter in practice can be used. The target vertex can represent any one micro pattern, and the target vertex can also represent any one point in any one micro pattern. The preset world coordinate system can be a world coordinate system with the upper left corner of the calibration board where the calibration pattern is located as the origin, the X axis horizontally to the left, the Y axis horizontally downward, and the Z axis as 0. The preset world coordinate system can also be set in other ways, and the present solution does not specifically limit the setting of the preset world coordinate system. In the preset world coordinate system, the world coordinate set of the target vertex on the calibration pattern is determined. It should be noted that the target vertex can only contain one vertex, and at this time the world coordinate set only contains one world coordinate. The target vertex can also contain two different vertices, and at this time the world coordinate set contains two world coordinates. That is, one vertex can be used for evaluating the camera intrinsic parameter only once, or two different vertices can be used. In order to improve the accuracy of the evaluation, when the camera intrinsic parameter needs to be evaluated multiple times, multiple target vertices can be selected as needed.

[0064] In the embodiments of the present specification, the camera to be evaluated is a camera whose camera intrinsic parameter has been calibrated, so the camera intrinsic parameter of the camera to be evaluated is a known parameter. Therefore, on the premise that the world coordinates, the first pixel coordinates, and the second pixel coordinates of the target vertex are obtained, the camera extrinsic parameter can be generated based on the conversion relationship between the world coordinate system and the camera coordinate system by using the known camera intrinsic parameter, and the camera extrinsic parameter set is obtained.

[0065] When the camera intrinsic parameter is evaluated only once, the camera extrinsic parameter set can contain two camera extrinsic parameters. When the target vertex contains only one vertex, the two camera extrinsic parameters in the camera extrinsic parameter set are camera extrinsic parameters generated according to the same vertex. When the target vertex contains two different vertices, the two camera extrinsic parameters in the camera extrinsic parameter set are camera extrinsic parameters generated according to different target vertices.

[0066] Step 106: generating the first estimated world coordinate of the first camera optical center position and the second estimated world coordinate of the second camera optical center position by using the conversion relationship between the world coordinate system and the camera coordinate system according to the world coordinate set and the camera extrinsic parameter set.

[0067] In the embodiments of the present specification, the camera extrinsic parameter can represent how a real world point (world coordinates) is rotated and translated, and then falls on another real world point (camera coordinates). Therefore, after learning the world coordinates of the target vertex on the calibration pattern in the real world, the estimated world coordinates at the camera optical center position can be obtained according to the camera extrinsic parameter and the conversion relationship between the world coordinate system and the camera coordinate system, wherein the obtained estimated world coordinates are world coordinates in the preset world coordinate system.

[0068] Step 108: generating a second distance between the first camera optical center position and the second camera optical center position according to the first estimated world coordinates and the second estimated world coordinates.

[0069] Step 110: generating an evaluation result of the camera intrinsic parameter of the camera to be evaluated by using the first distance and the second distance.

[0070] In the embodiments of the present specification, after obtaining the first estimated world coordinates at the first camera optical center position and the second estimated world coordinates at the second camera optical center position, the second distance between the first estimated world coordinates and the second estimated world coordinates is calculated by using a calculation method between coordinate points. The evaluation of the camera intrinsic parameter is completed according to the similarity between the first distance and the second distance. The smaller the difference between the first distance and the second distance is, the better the calibration result of the camera intrinsic parameter of the camera to be evaluated is, and the larger the difference between the first distance and the second distance is, the worse the calibration result of the camera intrinsic parameter of the camera to be evaluated is.

[0071] In the embodiments of the present specification, the camera extrinsic parameter is generated by using the camera intrinsic parameter of the camera to be evaluated under the condition that the distance between the camera and the calibration object is determined, so as to eliminate the influencing factors of the camera extrinsic parameter itself, and then replace the influencing factors of the camera extrinsic parameter with the influencing factors of the camera intrinsic parameter. The first distance used in the present solution is the distance between the first camera optical center position and the second camera optical center position obtained by measuring or other means. The second distance calculated and generated by using the camera extrinsic parameter, the first pixel coordinates and the second pixel coordinates is mainly affected by the camera intrinsic parameter. Therefore, when the camera intrinsic parameter is evaluated according to the first distance and the second distance, there is only one influencing factor of the camera intrinsic parameter, that is, the closer the positional relationship between the first distance and the second distance is, the better the evaluation result of the camera intrinsic parameter is. Therefore, the closer the first distance and the second distance are, the better the evaluation result of the camera intrinsic parameter is, and the evaluation result obtained is indeed better, that is, the accuracy of the evaluation result is higher. Therefore, the camera evaluation result obtained by using the present solution can mainly reflect the advantages and disadvantages of the camera intrinsic parameter, so as to improve the accuracy of the evaluation result of the camera intrinsic parameter.

[0072] Based on Figure 1The method in the embodiment of the present specification further provides some specific embodiments of the method, which are described below.

[0073] The step 104 of generating the set of camera extrinsics according to the set of world coordinates of the target vertex in the calibration pattern in a preset world coordinate system, the first pixel coordinates and the second pixel coordinates of the target vertex in the first image and the second image in a preset pixel coordinate system, and the camera intrinsic of the camera to be evaluated, using the conversion relationship between the world coordinate system and the pixel coordinate system, can include:

[0074] The target vertex includes a first target vertex and a second target vertex.

[0075] According to the first world coordinate of the first target vertex, the first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic in the set of world coordinates, the first camera extrinsic at the first camera optical center position is generated using the conversion relationship between the world coordinate system and the pixel coordinate system.

[0076] According to the second world coordinate of the second target vertex, the second pixel coordinate of the second target vertex in the preset pixel coordinate system, and the camera intrinsic in the set of world coordinates, the second camera extrinsic at the second camera optical center position is generated using the conversion relationship between the world coordinate system and the pixel coordinate system.

[0077] In the embodiment of the present specification, the first image and the second image are two different images of the same camera to be evaluated taken on the same calibration object. When the target vertex contains only one vertex, the first target vertex and the second target vertex can be the same vertex, and when the target vertex contains two different vertices, the first target vertex and the second target vertex can be two different vertices. The first pixel coordinate of the first target vertex in the calibration pattern in the pixel coordinate system of the first image is obtained from the first image. Similarly, the second pixel coordinate of the second target vertex in the calibration pattern in the pixel coordinate system of the second image is obtained from the second image.

[0078] In the embodiment of the present specification, when the first target vertex and the second target vertex are the same vertex, the first world coordinate and the second world coordinate can be the same world coordinate; when the first target vertex and the second target vertex are different vertices, the first world coordinate and the second world coordinate can be two different world coordinates. The conversion relationship between the world coordinate system and the pixel coordinate system can be: uv =KTP W , wherein: P uv represents the pixel coordinate, P Wwherein, X represents a world coordinate, Z represents a scale factor, K represents a camera intrinsic parameter, and T represents a camera extrinsic parameter. Therefore, after obtaining the first world coordinate of the first target vertex, the first pixel coordinate of the first target vertex, and the camera intrinsic parameter, the first camera extrinsic parameter can be calculated by using the above formula. Similarly, after obtaining the second world coordinate of the second target vertex, the second pixel coordinate of the second target vertex, and the camera intrinsic parameter, the second camera extrinsic parameter can be calculated by using the above formula.

[0079] The calibration pattern can be any pattern that can be used for camera calibration. Based on this, the calibration pattern is a pattern composed of a preset number of two-dimensional code patterns.

[0080] Before the first camera extrinsic parameter at the first camera optical center position is generated according to the first world coordinate of the first target vertex in the world coordinate set, the first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameter, the conversion relationship between the world coordinate system and the pixel coordinate system can also be used.

[0081] The first image is processed by using a two-dimensional code recognition algorithm to obtain first position information of a first two-dimensional code pattern in the first image.

[0082] According to the first position information of the first two-dimensional code pattern and the preset pixel coordinate system, the first pixel coordinate of the first target vertex in the first two-dimensional code pattern is determined.

[0083] Before the second camera extrinsic parameter at the second camera optical center position is generated according to the second world coordinate of the second target vertex in the world coordinate set, the second pixel coordinate of the second target vertex in the preset pixel coordinate system, and the camera intrinsic parameter, the conversion relationship between the world coordinate system and the pixel coordinate system can also be used.

[0084] The second image is processed by using a two-dimensional code recognition algorithm to obtain second position information of a second two-dimensional code pattern in the second image.

[0085] According to the second position information of the second two-dimensional code pattern and the preset pixel coordinate system, the second pixel coordinate of the second target vertex in the second two-dimensional code pattern is determined.

[0086] In the embodiments of the present specification, the calibration pattern can be a pattern composed of a preset number of two-dimensional code patterns according to a preset rule. The preset number of two-dimensional code patterns can all be the same two-dimensional code pattern, all be different two-dimensional code patterns, or part of the two-dimensional code patterns be the same and part of the two-dimensional code patterns be different. The two-dimensional code pattern can be a tag36h11 type two-dimensional code or other types of two-dimensional codes, and the present scheme does not make specific limitations on the type of two-dimensional code. The labels of each two-dimensional code in the two-dimensional code pattern are different. The preset number of two-dimensional code patterns included in the calibration pattern can be arranged in n rows and n columns, or can be arranged in other ways, and the present scheme does not make specific limitations on the arrangement manner of the preset number of two-dimensional code patterns. The first target vertex can be any one of the four corner points in the range of the first two-dimensional code pattern, or can be other vertices that can be recognized in the range. The second target vertex has the same property as the first target vertex. In the embodiments of the present specification, the first position information of the first two-dimensional code pattern is obtained from the first image, and any applicable two-dimensional code recognition algorithm can be used as long as the algorithm can recognize the position of the two-dimensional code pattern. For example, the vertices of the two-dimensional code can be detected by calling the extractTags function in the opencv open source library, and the present scheme does not make specific limitations on the applicable algorithm and implementation manner.

[0087] For the convenience of understanding, the embodiment can determine the first position information. First, all the two-dimensional code patterns in the calibration pattern are parsed, and the parsing order can be from top to bottom and from left to right, that is, the first row of the first two-dimensional code pattern is parsed from left to right, and then all the subsequent two-dimensional code patterns in the first row are parsed in turn. When the parsing module parses the last two-dimensional code in the first row, it means that there is no two-dimensional code pattern in the first row that needs to be parsed. At this time, the parsing module will automatically start identifying and parsing the second row, which is consistent with the parsing process of the first row. The parsing module will parse all the two-dimensional code patterns in the second row. By analogy, all the two-dimensional code patterns in the calibration pattern are parsed. After all the two-dimensional code patterns are parsed, the server will determine the position information of each two-dimensional code pattern according to the parsing information of the parsing module in the parsing process of all the two-dimensional code patterns, that is, the time when the two-dimensional code pattern is parsed, the order of the parsed two-dimensional code pattern, and the manner information (whether the first time when the recognition module recognizes the two-dimensional code pattern is a line feed operation) when the parsing module recognizes the parsed two-dimensional code pattern. In turn, the position information of each two-dimensional code pattern is determined, such as the position information of the two-dimensional code pattern a, which is represented as (6, 5), indicating that the two-dimensional code pattern a is located in the 6th row and the 5th position from left to right. After determining the position of the first two-dimensional code pattern in the first image, the pixel position of the first target vertex in the first two-dimensional code pattern can be determined in combination with the preset pixel coordinate system, and the first pixel coordinate of the first target vertex is obtained.

[0088] It should be noted that the first target vertex is a two-dimensional code pattern or a point in the two-dimensional code pattern, which is mainly related to the pixel of the first image. If the first image pixel is low, a two-dimensional code pattern is only contained in the position range of a certain image pixel. At this time, the first target vertex can be a two-dimensional code pattern. After determining the position of the first two-dimensional code pattern, the corresponding pixel position in the first image can be found according to the position, and the pixel coordinate of the pixel position is determined as the first pixel coordinate of the first target vertex. If the image pixel is high, a two-dimensional code pattern contains multiple image pixel positions. At this time, the first target vertex can be a point in the two-dimensional code pattern. The pixel coordinate of the pixel position with the largest overlapping range with the first target vertex is determined as the first pixel coordinate of the first target vertex by combining the position range of the two-dimensional code pattern into the pixel coordinate system of the first image.

[0089] Similarly, the second pixel coordinates of the second target vertex on the second image can also be obtained according to the manner of obtaining the first pixel coordinates of the first target vertex on the first image. The specific embodiments provided in the present solution are only for the convenience of understanding of those skilled in the art, and do not limit the manner of obtaining the pixel coordinates of the target vertex in the two-dimensional code pattern.

[0090] The calibration pattern can be a two-dimensional code pattern, and can also be other patterns that can be applied to camera calibration. Based on this, the calibration pattern is a chessboard pattern.

[0091] Before the first camera extrinsic parameter at the first camera optical center position is generated according to the first world coordinate of the first target vertex in the world coordinate set, the first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameter by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method further includes:

[0092] The first position information of the first chessboard grid in the first image is determined.

[0093] According to the first position information of the first chessboard grid and the preset pixel coordinate system, the first pixel coordinate of the first target vertex in the first chessboard grid is determined.

[0094] Before the second camera extrinsic parameter at the second camera optical center position is generated according to the second world coordinate of the second target vertex in the world coordinate set, the second pixel coordinate of the second target vertex in the preset pixel coordinate system, and the camera intrinsic parameter by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method can further include:

[0095] The second position information of the second chessboard grid in the second image is determined.

[0096] According to the second position information of the second chessboard grid and the preset pixel coordinate system, the second pixel coordinate of the second target vertex in the second chessboard grid is determined.

[0097] In the embodiments of the present specification, the calibration pattern can be a pattern composed of a preset number of checkerboard grids. The first position information of the first checkerboard grid obtained from the first image can be obtained using any applicable algorithm as long as the algorithm can identify the position of the checkerboard grid, and the present solution does not make specific limitations on the applicable algorithm and implementation. In order to facilitate understanding, an implementation for determining the above-mentioned first position information is proposed. Since adjacent checkerboards in the checkerboard pattern are grids with large changes in color gray scale, when the gray scale suddenly changes from small to large, or the gray scale changes from large to small, it indicates that one checkerboard enters another checkerboard. Therefore, based on the characteristics of large gray scale changes of adjacent checkerboards, the number of checkerboard grids contained in the first image can be determined in turn, and then the position of each checkerboard grid is determined. For example, the position of the first checkerboard grid can be represented as (5, 3), which means the 5th row and the 3rd column. After determining the position of the first checkerboard grid in the first image, in combination with the pixel coordinate system of the first image, the pixel position of the first target vertex in the first checkerboard grid can be determined, and the first pixel coordinate of the first target vertex is obtained.

[0098] It should be noted that whether the first target vertex is a checkerboard grid or a point in the checkerboard grid is mainly related to the pixel of the first image. The specific distinguishing process and the determination process of the pixel coordinate are consistent with the distinguishing method and the determination process of the pixel coordinate in the two-dimensional code pattern, so they will not be described here.

[0099] Similarly, according to the way of obtaining the first pixel coordinate of the first target vertex on the first image, the second pixel coordinate of the second target vertex on the second image can also be obtained. It should be noted that the specific embodiments proposed in the present solution are only for the convenience of those skilled in the art to understand, and do not limit the way of obtaining the pixel coordinate of the target vertex in the checkerboard grid.

[0100] Step 106: generating the first estimated world coordinate of the first camera optical center position and the second estimated world coordinate of the second camera optical center position according to the set of world coordinates and the set of camera extrinsic parameters, using the conversion relationship between the world coordinate system and the camera coordinate system, can include:

[0101] Generating the first estimated world coordinate according to the first world coordinate and the first camera extrinsic parameter, using the conversion relationship between the world coordinate system and the camera coordinate system.

[0102] Generating the second estimated world coordinate according to the second world coordinate and the second camera extrinsic parameter, using the conversion relationship between the world coordinate system and the camera coordinate system.

[0103] In the embodiments of the present specification, the first camera extrinsic parameter calculated in the foregoing can be a camera extrinsic parameter of a first camera coordinate system at the position of the optical center of the first camera to the preset world coordinate system, and the second camera extrinsic parameter can be a camera extrinsic parameter of a second camera coordinate system at the position of the optical center of the second camera to the preset world coordinate system. At this time, it is necessary to transform the first camera coordinate system and the second camera coordinate system to the same preset world coordinate system, and this transformation is a process of inverting the camera extrinsic parameter T. For example, if the camera extrinsic parameter T represents a rotation from left to right, then inverting T represents a transformation from right to left. After the above transformation operation, the first camera coordinate system and the second camera coordinate system are converted to the same preset world coordinate system. Based on the above principle, the first world coordinate can be converted to the first estimated world coordinate at the position of the optical center of the first camera in the preset world coordinate system after inverting the first camera extrinsic parameter, and the second world coordinate can also be converted to the second estimated world coordinate at the position of the optical center of the second camera in the preset world coordinate system after inverting the second camera extrinsic parameter.

[0104] After obtaining the first estimated world coordinate at the position of the optical center of the first camera and the second estimated world coordinate at the position of the optical center of the second camera, a second distance between the first estimated world coordinate and the second estimated world coordinate can be obtained. According to the correlation between the first distance and the second distance, a test result of the camera intrinsic parameter of the to-be-tested camera is obtained. Based on this, the second distance is any one of the Euclidean distance, the Manhattan distance, the Chebyshev distance, and the Minkowski distance.

[0105] The use of the first distance and the second distance to generate the test result of the camera intrinsic parameter of the to-be-tested camera can include:

[0106] The difference between the first distance and the second distance is calculated.

[0107] It is judged whether the ratio between the absolute value of the difference and the second distance is less than a preset threshold value, and a judgment result is obtained.

[0108] If the judgment result indicates that the ratio is less than the preset threshold value, it is determined that the test result of the camera intrinsic parameter of the to-be-tested camera is qualified.

[0109] If the judgment result indicates that the ratio is not less than the preset threshold value, it is determined that the test result of the camera intrinsic parameter of the to-be-tested camera is unqualified.

[0110] In the embodiments of the present specification, the second distance can be any one of Euclidean distance, Manhattan distance, Chebyshev distance, and Minkowski distance, and the present solution does not make specific limitation on the attribute of the second distance. The first distance can represent the actual distance between the first camera optical center position and the second camera optical center position, and the second distance can represent the virtual distance generated between the first camera optical center position and the second camera optical center position in a manner of calculation and derivation. The preset ratio relationship between the first distance and the second distance is determined, and it is judged whether the preset ratio relationship is less than a preset threshold value. If the preset ratio relationship is less than the preset threshold value, it is proved that the camera internal parameter of the camera to be evaluated is qualified, and if the preset ratio relationship is not less than the preset threshold value, it is proved that the camera internal parameter of the camera to be evaluated is unqualified. The preset ratio relationship can be the ratio of the absolute value of the difference between the first distance and the second distance to the second distance, or can also be other ratio relationships, and the present solution does not make specific limitation on the preset ratio relationship.

[0111] In the embodiments of the present specification, the first distance used in the present solution is the distance between the first camera optical center position and the second camera optical center position obtained by measurement and the like, and the second distance generated by calculation from the camera external parameter, the first pixel coordinate and the second pixel coordinate is mainly affected by the camera internal parameter. Therefore, when the camera internal parameter is evaluated according to the first distance and the second distance, there is only one influencing factor of the camera internal parameter, that is, the closer the positional relationship between the first distance and the second distance, the better the evaluation result of the camera internal parameter. Therefore, the closer the first distance and the second distance, the better the evaluation result of the camera internal parameter, and the evaluation result obtained is indeed better, that is, the accuracy of the evaluation result is higher. Therefore, the camera evaluation result obtained by the present solution can mainly reflect the advantages and disadvantages of the camera internal parameter, so as to improve the accuracy of the camera internal parameter evaluation result.

[0112] Based on the same idea, the present specification also provides a device corresponding to the above-mentioned method. Figure 3 The present specification provides a camera internal parameter evaluation device corresponding to the above-mentioned method. Figure 1 The structure schematic diagram of the camera internal parameter evaluation device is shown in FIG. 4. Figure 3 As shown in FIG. 4, the device can include:

[0113] The acquisition module 402 is configured to acquire the first distance between the first optical center position when the camera to be evaluated collects the first image of the calibration pattern and the second optical center position when the second image is collected.

[0114] The first generating module 404 is configured to generate a camera extrinsic parameter set according to a world coordinate set of target vertexes in the calibration pattern in a preset world coordinate system, first pixel coordinates and second pixel coordinates of the target vertexes in the first image and the second image in a preset pixel coordinate system, and camera intrinsic parameters of the camera to be evaluated, by using a conversion relationship between the world coordinate system and the pixel coordinate system.

[0115] The second generating module 406 is configured to generate a first estimated world coordinate of the first camera optical center position and a second estimated world coordinate of the second camera optical center position according to the world coordinate set and the camera extrinsic parameter set, by using a conversion relationship between the world coordinate system and a camera coordinate system.

[0116] The third generating module 408 is configured to generate a second distance between the first camera optical center position and the second camera optical center position according to the first estimated world coordinate and the second estimated world coordinate.

[0117] The fourth generating module 410 is configured to generate an evaluation result of the camera intrinsic parameters of the camera to be evaluated by using the first distance and the second distance.

[0118] Optionally, the first generating module 404 can include:

[0119] The target vertexes include a first target vertex and a second target vertex.

[0120] The first generating unit is configured to generate a first camera extrinsic parameter at the first camera optical center position according to a first world coordinate of the first target vertex in the world coordinate set, first pixel coordinates of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameters, by using the conversion relationship between the world coordinate system and the pixel coordinate system.

[0121] The second generating unit is configured to generate a second camera extrinsic parameter at the second camera optical center position according to a second world coordinate of the second target vertex in the world coordinate set, second pixel coordinates of the second target vertex in the preset pixel coordinate system, and the camera intrinsic parameters, by using the conversion relationship between the world coordinate system and the pixel coordinate system.

[0122] Optionally, the calibration pattern is a pattern composed of a preset number of two-dimensional code patterns.

[0123] Before the first generating unit, the method can further include:

[0124] The first processing subunit is configured to process the first image by using a two-dimensional code recognition algorithm to obtain first position information of a first two-dimensional code pattern in the first image.

[0125] The first determining sub-unit is configured to determine a first pixel coordinate of the first target vertex in the first two-dimensional code pattern according to the first position information of the first two-dimensional code pattern and the preset pixel coordinate system.

[0126] Before the second generating unit, the method further can include:

[0127] The second processing sub-unit is configured to process the second image by using a two-dimensional code recognition algorithm to obtain second position information of a second two-dimensional code pattern in the second image.

[0128] The second determining sub-unit is configured to determine a second pixel coordinate of the second target vertex in the second two-dimensional code pattern according to the second position information of the second two-dimensional code pattern and the preset pixel coordinate system.

[0129] Optionally, the calibration pattern is a checkerboard pattern.

[0130] Before the first generating unit, the method further can include:

[0131] The third determining sub-unit is configured to determine first position information of a first checkerboard grid in the first image.

[0132] The fourth determining sub-unit is configured to determine a first pixel coordinate of the first target vertex in the first checkerboard grid according to the first position information of the first checkerboard grid and the preset pixel coordinate system.

[0133] Before the second generating unit, the method further can include:

[0134] The fifth determining sub-unit is configured to determine second position information of a second checkerboard grid in the second image.

[0135] The sixth determining sub-unit is configured to determine a second pixel coordinate of the second target vertex in the second checkerboard grid according to the second position information of the second checkerboard grid and the preset pixel coordinate system.

[0136] Optionally, the second generating module 406 can include:

[0137] The third generating unit is configured to generate the first estimated world coordinate by using a conversion relationship between a world coordinate system and a camera coordinate system according to the first world coordinate and the first camera extrinsic parameter.

[0138] The fourth generating unit is configured to generate the second estimated world coordinate by using a conversion relationship between a world coordinate system and a camera coordinate system according to the second world coordinate and the second camera extrinsic parameter.

[0139] The fourth generating module 410 can include:

[0140] The first computing unit is configured to calculate a difference between the first distance and the second distance.

[0141] The first judging unit is configured to judge whether a ratio between an absolute value of the difference and the second distance is less than a preset threshold, to obtain a judging result.

[0142] The first determining unit is configured to determine that the evaluation result of the camera intrinsic parameter of the camera to be evaluated is qualified if the judging result indicates that the ratio is less than the preset threshold.

[0143] The second determining unit is configured to determine that the evaluation result of the camera intrinsic parameter of the camera to be evaluated is unqualified if the judging result indicates that the ratio is not less than the preset threshold.

[0144] Optionally, the second distance is any one of a Euclidean distance, a Manhattan distance, a Chebyshev distance, and a Minkowski distance.

[0145] Based on the same idea, the present specification also provides a device corresponding to the above method.

[0146] Figure 4 A structural schematic diagram of a camera intrinsic parameter evaluation device corresponding to Figure 1 provided by an embodiment of the present specification is shown in FIG. 5. As shown in FIG. 5, the device 500 can include: Figure 4

[0147] at least one processor 510; and

[0148] a memory 530 in communication with the at least one processor; wherein

[0149] the memory 530 stores instructions 520 executable by the at least one processor 510, and the instructions are executed by the at least one processor 510 to enable the at least one processor 510 to execute the instructions to implement the above method.

[0150] The present application also provides a computer readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the above method.

[0151] It should be understood that in the method described in one or more embodiments of the present specification, the order of some steps can be adjusted according to actual needs, or some steps can be omitted.

[0152] Each of the embodiments in the present specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment mainly describes the difference from other embodiments. In particular, for Figure 4 ​The devices, computer-readable storage media described are, due to their substantial similarity to the method embodiments, described in relatively simple terms, with reference to the relevant parts of the description of the method embodiments.

[0153] The above only describes the embodiments of the present application and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A method for evaluating camera intrinsic parameters, characterized in that, The method comprises: acquiring a first distance between a first camera optical center position of a camera to be evaluated when a first image is collected for a calibration pattern and a second camera optical center position when a second image is collected; generating a camera extrinsic parameter set according to a world coordinate set of target vertices in the calibration pattern in a preset world coordinate system, first pixel coordinates and second pixel coordinates of the target vertices in the first image and the second image in a preset pixel coordinate system, and camera intrinsic parameters of the camera to be evaluated, by using a conversion relationship between the world coordinate system and the pixel coordinate system; generating a first estimated world coordinate of the first camera optical center position and a second estimated world coordinate of the second camera optical center position according to the world coordinate set and the camera extrinsic parameter set, by using a conversion relationship between the world coordinate system and a camera coordinate system; generating a second distance between the first camera optical center position and the second camera optical center position according to the first estimated world coordinate and the second estimated world coordinate; generating an evaluation result of the camera intrinsic parameters of the camera to be evaluated by using the first distance and the second distance.

2. The method of claim 1, wherein, The method comprises: The target vertices comprise a first target vertex and a second target vertex. A first camera extrinsic parameter at the first camera optical center position is generated according to a first world coordinate of the first target vertex in the world coordinate set, a first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameters, by using the conversion relationship between the world coordinate system and the pixel coordinate system. A second camera extrinsic parameter at the second camera optical center position is generated according to a second world coordinate of the second target vertex in the world coordinate set, a second pixel coordinate of the second target vertex in the preset pixel coordinate system, and the camera intrinsic parameters, by using the conversion relationship between the world coordinate system and the pixel coordinate system.

3. The method of claim 2, wherein, The calibration pattern is a pattern composed of a preset number of two-dimensional code patterns. Before the first camera extrinsic parameter at the first camera optical center position is generated according to the first world coordinate of the first target vertex in the world coordinate set, the first pixel coordinate of the first target vertex in the preset pixel coordinate system, and the camera intrinsic parameters, by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method further comprises: processing the first image by using a two-dimensional code recognition algorithm to obtain first position information of a first two-dimensional code pattern in the first image; determining the first pixel coordinate of the first target vertex in the first two-dimensional code pattern according to the first position information of the first two-dimensional code pattern and the preset pixel coordinate system. Before the generating the second camera extrinsic parameter at the second camera optical center position according to the second world coordinate of the second target vertex in the world coordinate set, the second pixel coordinate of the second target vertex under the preset pixel coordinate system, and the camera intrinsic parameter by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method further comprises: processing the second image by using a two-dimensional code recognition algorithm to obtain second position information of a second two-dimensional code pattern in the second image; determining the second pixel coordinate of the second target vertex in the second two-dimensional code pattern according to the second position information of the second two-dimensional code pattern and the preset pixel coordinate system.

4. The method of claim 2, wherein, The calibration pattern is a chessboard pattern. Before the generating the first camera extrinsic parameter at the first camera optical center position according to the first world coordinate of the first target vertex in the world coordinate set, the first pixel coordinate of the first target vertex under the preset pixel coordinate system, and the camera intrinsic parameter by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method further comprises: determining first position information of a first chessboard grid in the first image; determining the first pixel coordinate of the first target vertex in the first chessboard grid according to the first position information of the first chessboard grid and the preset pixel coordinate system. Before the generating the second camera extrinsic parameter at the second camera optical center position according to the second world coordinate of the second target vertex in the world coordinate set, the second pixel coordinate of the second target vertex under the preset pixel coordinate system, and the camera intrinsic parameter by using the conversion relationship between the world coordinate system and the pixel coordinate system, the method further comprises: determining second position information of a second chessboard grid in the second image; determining the second pixel coordinate of the second target vertex in the second chessboard grid according to the second position information of the second chessboard grid and the preset pixel coordinate system.

5. The method of claim 2, wherein, The generating the first estimated world coordinate at the first camera optical center position and the second estimated world coordinate at the second camera optical center position by using the conversion relationship between the world coordinate system and the camera coordinate system according to the world coordinate set and the camera extrinsic parameter set comprises: generating the first estimated world coordinate by using the conversion relationship between the world coordinate system and the camera coordinate system according to the first world coordinate and the first camera extrinsic parameter; generating the second estimated world coordinate by using the conversion relationship between the world coordinate system and the camera coordinate system according to the second world coordinate and the second camera extrinsic parameter.

6. The method of claim 1, wherein, The generating the evaluation result of the camera intrinsic parameter of the camera to be evaluated by using the first distance and the second distance comprises: calculating a difference value between the first distance and the second distance; judging whether a ratio between an absolute value of the difference value and the second distance is less than a preset threshold value to obtain a judgment result; if the judgment result indicates that the ratio is less than the preset threshold value, determining that the evaluation result of the camera intrinsic parameter of the camera to be evaluated is qualified. If the judging result indicates that the ratio is not less than the preset threshold, it is determined that the evaluation result of the camera intrinsic parameter of the camera to be evaluated is unqualified.

7. The method of claim 6, wherein, The second distance is any one of a Euclidean distance, a Manhattan distance, a Chebyshev distance, and a Minkowski distance.

8. A device for evaluating camera intrinsic parameters, characterized in that, The method comprises the steps of: The acquisition module is configured to acquire a first distance between a first camera optical center position when the camera to be evaluated collects a first image for a calibration pattern and a second camera optical center position when the camera to be evaluated collects a second image. The first generation module is configured to generate a camera extrinsic parameter set by using a conversion relationship between a world coordinate system and a pixel coordinate system according to a world coordinate set of a target vertex in the calibration pattern in the world coordinate system, first pixel coordinates and second pixel coordinates of the target vertex in the first image and the second image, and a camera intrinsic parameter of the camera to be evaluated. The second generation module is configured to generate a first estimated world coordinate of the first camera optical center position and a second estimated world coordinate of the second camera optical center position by using a conversion relationship between the world coordinate system and a camera coordinate system according to the world coordinate set and the camera extrinsic parameter set. The third generation module is configured to generate a second distance between the first camera optical center position and the second camera optical center position according to the first estimated world coordinate and the second estimated world coordinate. The fourth generation module is configured to generate an evaluation result of the camera intrinsic parameter of the camera to be evaluated by using the first distance and the second distance.

9. An apparatus for evaluating camera intrinsic parameters, comprising a memory, a processor and a computer program stored on the memory, characterized in that: The processor executes the computer program to implement the steps of the method in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 7.

Images