External Parameter Calibration Method and Device, Equipment, and Storage Medium for Images

Abstract

The present disclosure relates to an external parameter calibration method for images, including: obtaining a camera image and a radar image corresponding to the external parameter to be calibrated. Converting the camera image into a preset standard image. The acquisition position of the camera image is consistent with the acquisition position of the standard image. Performing external parameter calibration based on the standard image and the radar image. The present disclosure is applicable to online external parameter calibration of the camera graphics collected by the camera and the radar images collected by the lidar, performing external parameter calibration processing on the converted standard image and the radar image, and obtaining calibrated external parameters that match the camera image and the radar image. Thus, in the external parameter calibration method of the present application, by introducing the standard image, when performing external parameter calibration by combining the standard image with the radar image, it is only necessary to convert the camera image collected by the camera into the standard image, and there is no need to update and adjust the external parameter calibration algorithm for the standard image and the radar image, thereby improving the applicable range of external parameter calibration.

Description

Technical Field

[0001] The present disclosure relates to the technical field of multi-sensor information fusion, and particularly to an external parameter calibration method, device, equipment and storage medium for images. Background Art

[0002] In the fields of robotics and autonomous driving, cameras and lidar are widely used sensors. In order to better perceive the surrounding environment or facilitate the annotation of the surrounding environment for machine learning, it is necessary to fuse the data of multiple sensors. Different sensors are deployed at different positions and orientations. In order to fuse the data, external parameters are needed to represent how to transform a certain sensor from one position and orientation to another position and orientation. For example, transforming the lidar sensor from its original position and orientation to the same position and orientation as the camera sensor so that the data collected by the lidar matches the data collected by the camera.

[0003] Currently, external parameter calibration algorithms are often used to process the image data collected by camera sensors and the point cloud data collected by lidar. The external parameter calibration algorithm is generally a neural network, and the calibrated external parameters are output for data fusion.

[0004] A system vehicle equipped with camera and lidar sensors is called an acquisition system. Existing external parameter calibration algorithms are related to the specifications of the cameras in the acquisition system. If the camera specifications are changed, the algorithm needs to be adjusted and the neural network needs to be retrained, which is rather cumbersome. Summary of the Invention

[0005] In view of this, the present disclosure proposes an external parameter calibration method for images. By converting the camera image into a preset standard image and then using an external parameter calibration algorithm that matches the standard image and the radar image, it is possible to avoid retraining the external parameter calibration algorithm after changing cameras of different specifications.

[0006] According to one aspect of the present disclosure, there is provided an external parameter calibration method for images, including:

[0007] Obtaining a camera image and a radar image corresponding to the external parameter to be calibrated;

[0008] Converting the camera image into a preset standard image;

[0009] Wherein, the acquisition position of the camera image is consistent with the acquisition position of the standard image;

[0010] Performing external parameter calibration based on the standard image and the radar image.

[0011] In a possible implementation, when converting the camera image into the preset standard image, it is performed according to a pre-constructed conversion mapping.

[0012] In a possible implementation, the conversion mapping is as follows:

[0013]

[0014] where (u1, v1) are the pixel coordinates of the standard image, and (u2, v2) are the pixel coordinates of the camera image. and are the focal lengths of the imaging devices corresponding to the standard image and the camera image in units of horizontal pixels respectively. and are the focal lengths of the imaging devices corresponding to the standard image and the camera image in units of vertical pixels respectively. and are the horizontal offsets of the principal points of the imaging devices corresponding to the standard image and the camera image on the acquired images respectively. and are the vertical offsets of the principal points of the imaging devices corresponding to the standard image and the camera image on the acquired images respectively.

[0015] In a possible implementation, when constructing the conversion mapping, it is based on the internal parameter matrix of the imaging device corresponding to the standard image and the internal parameter matrix of the imaging device corresponding to the camera image.

[0016] In a possible implementation, the camera models corresponding to the imaging device for acquiring the standard image and the imaging device for acquiring the camera image are consistent.

[0017] In a possible implementation, when constructing the conversion mapping based on the internal parameter matrix of the imaging device corresponding to the standard image and the internal parameter matrix of the imaging device corresponding to the camera image, it includes:

[0018] Based on the focal length in units of horizontal pixels and the focal length in units of vertical pixels of the imaging device for acquiring the standard image, as well as the horizontal offset and vertical offset of the principal point of the imaging device for acquiring the standard image on the standard image, the internal parameter matrix of the imaging device corresponding to the standard image is obtained;

[0019] Based on the focal length in units of horizontal pixels and the focal length in units of vertical pixels of the imaging device for acquiring the camera image, as well as the horizontal offset and vertical offset of the principal point of the imaging device for acquiring the camera image on the camera image, the internal parameter matrix of the imaging device corresponding to the camera image is obtained;

[0020] Construct the conversion mapping based on the internal parameter matrix of the camera device corresponding to the standard image and the internal parameter matrix of the camera device corresponding to the camera image.

[0021] According to another aspect of the present disclosure, there is provided an external parameter calibration device for images, including: an acquisition module, a conversion module, and an external parameter calibration module;

[0022] The acquisition module is configured to obtain the camera image and the radar image corresponding to the external parameter to be calibrated;

[0023] The conversion module is configured to convert the camera image into the preset standard image;

[0024] The external parameter calibration module is configured to perform external parameter calibration according to the standard image and the radar image.

[0025] In a possible implementation manner, when the conversion module is configured to convert the camera image into the preset standard image, it is performed according to a pre-constructed conversion mapping.

[0026] According to another aspect of the present disclosure, there is provided an external parameter calibration device for images, including:

[0027] A processor;

[0028] A memory for storing processor-executable instructions;

[0029] Wherein, when the processor is configured to execute the executable instructions, the method described in any one of the above is implemented.

[0030] According to another aspect of the present disclosure, there is provided a non-volatile computer-readable storage medium, on which computer program instructions are stored, and characterized in that when the computer program instructions are executed by a processor, the method described in any one of the above is implemented.

[0031] The present disclosure is applicable to online extrinsic parameter calibration for camera graphics collected by a camera and radar images collected by a lidar. Camera images and radar images corresponding to the extrinsic parameters to be calibrated are obtained, where the radar images are point cloud data. The camera images are converted into preset standard images, where the acquisition positions of the camera images are consistent with those of the standard images, that is, the camera device for collecting the camera images and the camera device for collecting the standard images are at the same position and are taken at the same angle. The converted standard images and radar images are processed using an extrinsic parameter calibration algorithm to obtain calibrated extrinsic parameters that match the camera images and radar images. Among them, the extrinsic parameter calibration algorithm is pre-trained to match the standard images and radar images. In this way, after replacing the camera, since the extrinsic parameter calibration algorithm used is based on the standard images and radar images, even if the camera specifications are changed, but the standard images are not changed, so there is no need to re-train the extrinsic parameter calibration algorithm according to the replaced camera. It is only necessary to convert the camera images collected by the camera into standard images, which is relatively simple compared to re-training the extrinsic parameter calibration algorithm.

[0032] Other features and aspects of the present disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The drawings included in and constituting a part of the specification illustrate exemplary embodiments, features, and aspects of the present disclosure and are used to explain the principles of the present disclosure together with the specification.

[0034] Figure 1 A flowchart of an extrinsic parameter calibration method for images according to an embodiment of the present disclosure is shown;

[0035] Figure 2 A schematic diagram of a planar image coordinate system according to an embodiment of the present disclosure is shown;

[0036] Figure 3 A schematic diagram of a camera device coordinate system according to an embodiment of the present disclosure is shown;

[0037] Figure 4 A schematic diagram of the positional relationship between the camera device coordinate system and the planar image coordinate system and the planar image coordinate system according to an embodiment of the present disclosure is shown;

[0038] Figure 5 A main structure diagram of an extrinsic parameter calibration device for images according to an embodiment of the present disclosure is shown;

[0039] Figure 6 A main structure diagram of an extrinsic parameter calibration device for images according to an embodiment of the present disclosure is shown;

[0040] Figure 7Shows the schematic diagram of the external parameter calibration algorithm according to the embodiments of the present disclosure. Detailed implementation manners

[0041] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings do not have to be drawn to scale unless otherwise specified.

[0042] The term "exemplary" used herein means "serving as an example, embodiment, or illustration". Any embodiment described herein as "exemplary" does not necessarily have to be construed as superior or better than other embodiments.

[0043] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following detailed implementation manners. Those skilled in the art should understand that the present disclosure can be implemented without some specific details. In some instances, methods, means, elements, and circuits well-known to those skilled in the art are not described in detail so as to highlight the gist of the present disclosure.

[0044] For the convenience of understanding the technical solutions of the present application, the terms in the present application are first explained accordingly. The external parameter refers to how to transform a certain sensor from one position and orientation to another position and orientation. For example, transforming the lidar sensor from its original position and orientation to the same position and orientation as the camera sensor. Usually, a matrix is used to represent this transformation, which includes a rotation transformation and a translation transformation. Among them, the external parameter can be estimated by measuring the position and angle of the sensor, but the accurate external parameter needs to be obtained through the external parameter calibration algorithm.

[0045] The internal parameter refers to the parameters included in the camera model obtained after modeling the camera for image acquisition, and these parameters are generally provided by the camera supplier.

[0046] Figure 1 Shows the flowchart of the external parameter calibration method for images according to the embodiments of the present disclosure. Figure 2 Shows the schematic diagram of the plane image coordinate system in the embodiments of the present disclosure; Figure 3 Shows the schematic diagram of the camera device coordinate system in the embodiments of the present disclosure; Figure 4 Shows the schematic diagram of the positional relationship between the camera device coordinate system and the plane image coordinate system and the plane image coordinate system in the embodiments of the present disclosure; Figure 5 Shows the main structure diagram of the external parameter calibration device for images according to the embodiments of the present disclosure; Figure 6 Shows the main structure diagram of the external parameter calibration device for images according to the embodiments of the present disclosure. Figure 7 Shows the schematic diagram of the external parameter calibration algorithm according to the embodiments of the present disclosure.

[0047] As shown Figure 1 in the figure, the external parameter calibration method for images includes: Step S100: Obtain the camera image and radar image corresponding to the external parameter to be calibrated. Step S200: Convert the camera image into a preset standard image. Wherein, the acquisition position of the camera image is consistent with the acquisition position of the standard image. Step S300: Perform external parameter calibration according to the standard image and the radar image.

[0048] This disclosure is applicable to online external parameter calibration of the camera graphics collected by the camera and the radar images collected by the lidar. Obtain the camera image and radar image corresponding to the external parameter to be calibrated, where the radar image is point cloud data. Convert the camera image into a preset standard image, where the acquisition position of the camera image is consistent with the acquisition position of the standard image, that is, the imaging device for collecting the camera image and the imaging device for collecting the standard image are in the same position and are taken at the same angle. Process the converted standard image and radar image using an external parameter calibration algorithm to obtain the calibrated external parameters that match the camera image and the radar image.

[0049] Among them, the external parameter calibration algorithm is pre-trained to match the standard image and the radar image. In this way, after replacing the camera, there is no need to re-train the external parameter calibration algorithm according to the replaced camera. It is only necessary to convert the camera image collected by the camera into a standard image. Compared with re-training the external parameter calibration algorithm, it greatly reduces the operation of updating the external parameter calibration algorithm every time after the camera is replaced, which effectively simplifies the external parameter calibration method.

[0050] Among them, when processing the converted standard image and radar image using an external parameter calibration algorithm, the external parameter calibration algorithm is a commonly used technical means in the art. For example, refer to Figure 7 , a depth convolutional neural network (CNN) can be used to train RegNet to infer the correspondence between the depth measurement of the lidar projected onto the imaging device for collecting the standard image and the RGB image. RegNet maps the three steps of calibration (feature extraction, feature matching, and global regression) into a single real-time CNN model. During training, the system is randomly readjusted so that RegNet can infer the correspondence between the depth measurement of the lidar projected onto the imaging device for collecting the standard image and the RGB image (standard image), and finally regress the calibrated external parameters. By iteratively executing multiple CNNs, training is performed on data with different degrees of decalibration.

[0051] As shown Figure 7 in the figure, use the initial external parameter H initProject the depth points onto the RGB image. In the first and second parts of the CNN network, NiN (Network-in-Network) blocks are used to extract rich features for matching, where the index shows the kernel size k of the first convolutional layer of the NiN block. The number of feature channels is shown in the upper right corner of each module. The last part of the CNN network regresses the misalignment by collecting global information using two fully connected layers. (Note: The NiN block consists of a k×k convolution followed by several 1×1 convolutions.) During training, the misalignment matrix is randomly permuted to form different projection data of the depth points. The projection of the depth points varies strongly with the given initial extrinsic parameter values. When the initial calibration deviates from the calibrated ground truth (GT) by 20° in rotation and 1.5 m in translation, most of the points projected from the lidar point cloud may be outside the image area, making it difficult to establish a correspondence with the RGB image. Even in these cases, the trained CNN network can still improve the calibration. Using the newly estimated calibration parameters, the depth points can be projected again, resulting in more depth points for relevant calculations. Then, this step is iterated multiple times to obtain the final quantity to be calibrated. That is, the extrinsic parameters after calibration.

[0052] In a possible implementation, when converting the camera image to a preset standard image, it is performed according to a pre-constructed conversion mapping. Through the pre-constructed conversion mapping, the camera image is converted into a standard image.

[0053] Specifically, since in the method of the present disclosure embodiment, when converting the camera picture to a standard picture through a pre-constructed conversion mapping, it includes processing operations for coordinate conversion of each pixel point in the camera picture and the standard picture. Therefore, a corresponding coordinate system needs to be constructed for coordinate conversion of the pixel points.

[0054] Among them, referring to Figure 2 , the planar image collected by the imaging device is a mapping of the data collected by the rectangular sensor of the imaging device. Therefore, taking the long side and the wide side of the planar image as the u-axis and the v-axis respectively, and taking the upper left vertex of the planar image as the origin, a planar image coordinate system is constructed so that the planar image is located in the positive direction of the u-axis and the positive direction of the v-axis.

[0055] Using the above method, a camera picture coordinate system and a standard picture coordinate system are constructed respectively according to the camera picture and the standard picture. According to the conversion mapping:

[0056]

[0057] The coordinates (u1, v1) of each pixel point of the standard image can be made to correspond one-to-one to the coordinates (u2, v2) of each pixel point in the camera image.

[0058] Wherein, and are the focal lengths of the imaging devices corresponding to the standard image and the camera image respectively, in horizontal pixels. and are the focal lengths of the imaging devices corresponding to the standard image and the camera image respectively, in vertical pixels. and are the horizontal offsets of the principal points of the imaging devices corresponding to the standard image and the camera image respectively on the acquired image. and are the vertical offsets of the principal points of the imaging devices corresponding to the standard image and the camera image respectively on the acquired image. That is, when converting the camera image to the standard image, first, the coordinates of each pixel point in the camera image are converted through the above conversion mapping formula. When mapping and converting the pixel point coordinates of the standard image for each pixel point, the mapping and conversion can be performed sequentially for each pixel point. Since the calculation of the mapping for each pixel point is a process without mutual dependence, the mapping conversion can also be performed synchronously for multiple pixel points. When performing the mapping conversion synchronously for multiple pixel points, the number of pixel points selected for synchronous mapping conversion in the camera image can be a preset number. The preset number is generally the number of available GPU computing units or the number of available CPU cores. For example, the value range of the preset number can be from 1 to 8. At the same time, after mapping the pixel point coordinates of the camera image to the standard image according to the above conversion mapping formula, it also includes the process of copying the color values of each pixel point, so that the finally converted standard image is consistent with the camera image. That is, obtain the color values of each pixel point coordinate (u2, v2) in the corresponding camera image, and copy the color values of each pixel point (u2, v2) in the camera image to the corresponding pixel point coordinates (u1, v1) of the standard image, so as to convert the camera picture into a standard picture.

[0059] Here, it should also be noted that the specifications of the camera for collecting the standard image can be flexibly set according to the actual situation. Generally speaking, the specifications of the camera for collecting the standard image usually select a moderate rule, neither too large nor too small.

[0060] In a possible implementation, but in order to achieve better results (to make the converted standard image retain as much information as possible and make each pixel point on the standard image have a mapped pixel point on the camera image as much as possible), the specifications of the imaging device for collecting the standard image generally select a moderate acquisition camera. The setting of the standard image can be: 720p image (1280x720), and the specifications of the imaging device for collecting the standard image can be: By setting the camera rules of the standard image to the above parameters, the converted standard image can retain as much image information as possible and can also reduce black edges.

[0061] Furthermore, and are all the parameters in the internal parameter matrix of the camera corresponding to the standard image and the internal parameter matrix of the camera corresponding to the camera image. A conversion mapping is constructed based on the internal parameter matrix of the camera corresponding to the standard image and the internal parameter matrix of the camera corresponding to the camera image.

[0062] Here, it should be noted that since in the method of constructing the conversion mapping, it includes using the internal parameter matrix under the camera model, and the camera model is used to simulate the imaging process of the camera. Therefore, the camera models corresponding to the camera device for collecting the standard image and the camera device for collecting the camera image are consistent.

[0063] For example, both the camera device for collecting the standard image and the camera device for collecting the camera image use ordinary cameras and are modeled using the pinhole camera model. Refer to Figure 2 Figure 3 and Figure 4 Taking the optical center of the lens of the camera device as the origin of the camera device coordinate system, taking the optical axis of the lens of the camera device as the z-axis of the camera device coordinate system, and taking the long side and the short side of the rectangular sensor of the camera device as the x-axis and y-axis of the camera device coordinate system respectively, a camera device coordinate system is constructed. As Figure 4 shown, the planar image collected by the camera device is a mapping of the data collected by the rectangular sensor of the camera device. Therefore, taking the long side and the short side of the planar image as the u-axis and v-axis respectively, and taking the upper left vertex of the planar image as the origin to construct a planar image coordinate system, where the positive direction of the u-axis is the same as the positive direction of the x-axis of the camera device coordinate system, and the positive direction of the v-axis is the same as the positive direction of the y-axis of the camera device coordinate system.

[0064] Therefore, a point (x, y, z) in the camera device coordinate system is projected onto the point (u, v) in the corresponding image coordinate system, and its expression is:

[0065]

[0066] where K is the internal parameter matrix composed of the camera internal parameters:

[0067]

[0068] where when constructing the conversion mapping based on the internal parameter matrix of the camera corresponding to the standard image and the internal parameter matrix of the camera corresponding to the camera image, it includes: based on the focal length of the camera device for collecting the standard image in horizontal pixels and the focal length in terms of vertical pixels and the horizontal offset of the principal point of the imaging device that captures the standard image on the standard image and the vertical offset The internal parameter matrix of the imaging device corresponding to the standard image is obtained. Based on the focal length of the imaging device that captures the camera image in terms of horizontal pixels and the focal length in terms of vertical pixels and the horizontal offset of the principal point of the imaging device that captures the camera image on the camera image and the vertical offset The internal parameter matrix of the imaging device corresponding to the camera image is obtained. A conversion mapping is constructed based on the internal parameter matrix of the imaging device corresponding to the standard image and the internal parameter matrix of the imaging device corresponding to the camera image.

[0069] Furthermore, two imaging devices with different specifications are used to capture the same scene at the same position and at the same angle, that is, the imaging device for capturing the camera image and the imaging device for capturing the standard image are in the same camera coordinate system. Then, for any point (x, y, z) in this camera coordinate system, the following relational expressions can be obtained:

[0070]

[0071] Therefore, there is:

[0072]

[0073]

[0074] That is:

[0075]

[0076] Furthermore, the conversion mapping is obtained for converting the camera image into the standard image.

[0077] For any pixel point (u1, v1) on the standard image, using the conversion mapping, each pixel point on the standard image is mapped to each pixel point (u2, v2) on the camera image, and the color values corresponding to each coordinate point on the corresponding camera image are obtained, thereby obtaining the converted standard image.

[0078] When mapping each pixel point (u1, v1) on the standard image to the camera image using the conversion mapping, the standard image may exceed the coordinate range of the camera picture. At this time, the pixel coordinates of the standard picture that exceed the camera picture are set to a preset color. Through such processing, the converted standard image may have preset color pixels at the edges. And an appropriate amount of single-color edges in the standard image have little impact on subsequent processing. Among them, the preset color can be set to black.

[0079] Furthermore, referring to Figure 5 , according to another aspect of the present disclosure, there is also provided an external parameter calibration device 100 for an image, including: an acquisition module 110, a conversion module 120, and an external parameter calibration module 130. The acquisition module 110 is configured to obtain a camera image and a radar image corresponding to the external parameter to be calibrated. The conversion module 120 is configured to convert the camera image into a preset standard image. The external parameter calibration module 130 is configured to perform external parameter calibration according to the standard image and the radar image.

[0080] Furthermore, according to another aspect of the present disclosure, there is also provided an external parameter calibration device 200 for an image. Referring to Figure 6 , the external parameter calibration device 200 for an image in the embodiments of the present disclosure includes a processor 210 and a memory 220 for storing executable instructions that can be executed by the processor 210. Among them, when the processor 210 is configured to execute the executable instructions, it implements the external parameter calibration method for an image described in any one of the foregoing.

[0081] Here, it should be noted that the number of processors 210 can be one or more. At the same time, in the external parameter calibration device 200 for an image in the embodiments of the present disclosure, an input device 230 and an output device 240 may also be included. Among them, the processor 210, the memory 220, the input device 230, and the output device 240 can be connected through a bus or in other ways, and specific limitations are not made here.

[0082] The memory 220, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and various modules, such as: the programs or modules corresponding to the external parameter calibration method for an image in the embodiments of the present disclosure. The processor 210 executes various functional applications and data processing of the external parameter calibration device 200 for an image by running the software programs or modules stored in the memory 220.

[0083] The input device 230 can be used to receive input numbers or signals. Among them, the signal can be a key signal related to the user settings and function control of the device / terminal / server. The output device 240 can include a display device such as a display screen.

[0084] According to another aspect of the present disclosure, there is also provided a non-volatile computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor 210, the external parameter calibration method for images described in any of the foregoing is implemented.

[0085] The embodiments of the present disclosure have been described above. The above description is exemplary and not exhaustive, and is also not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, the practical application, or the improvement of the technology in the market, or to enable other ordinary technicians in the art to understand the embodiments disclosed herein.

Claims

1. An external parameter calibration method for images, characterized in that, Including: Obtain a camera image and a radar image corresponding to the external parameters to be calibrated; Convert the camera image into a preset standard image; Wherein, the acquisition position of the camera image is consistent with the acquisition position of the standard image; Perform external parameter calibration based on the standard image and the radar image; When converting the camera image into the preset standard image, it is performed according to a pre-constructed conversion mapping, and the conversion mapping is: Among them, (u1, v1) are the pixel coordinates of the standard image, and (u2, v2) are the pixel coordinates of the camera image. and are the focal lengths of the imaging devices corresponding to the standard image and the camera image respectively, in units of horizontal pixels. and are the focal lengths of the imaging devices corresponding to the standard image and the camera image respectively, in units of vertical pixels. and are the horizontal offsets of the principal points of the imaging devices corresponding to the standard image and the camera image respectively on the acquired images. and are the vertical offsets of the principal points of the imaging devices corresponding to the standard image and the camera image respectively on the acquired images.

2. The method according to claim 1, characterized in that, When constructing the conversion mapping, it is based on the internal parameter matrix of the camera device corresponding to the standard image and the internal parameter matrix of the camera device corresponding to the camera image.

3. The method according to claim 2, wherein The camera model corresponding to the camera device for acquiring the standard image is consistent with the camera model corresponding to the camera device for acquiring the camera image.

4. The method according to claim 2, wherein When constructing the conversion mapping based on the internal parameter matrix of the camera device corresponding to the standard image and the internal parameter matrix of the camera device corresponding to the camera image, it includes: Based on the focal length in horizontal pixels and the focal length in vertical pixels of the camera device for acquiring the standard image, and the horizontal offset and vertical offset of the principal point of the camera device for acquiring the standard image on the standard image, obtain the internal parameter matrix of the camera device corresponding to the standard image; Based on the focal length in horizontal pixels and the focal length in vertical pixels of the camera device for acquiring the camera image, and the horizontal offset and vertical offset of the principal point of the camera device for acquiring the camera image on the camera image, obtain the internal parameter matrix of the camera device corresponding to the camera image; Construct the conversion mapping based on the internal parameter matrix of the camera device corresponding to the standard image and the internal parameter matrix of the camera device corresponding to the camera image.

5. An external parameter calibration device for images, characterized in that, Including: An acquisition module, a conversion module, and an external parameter calibration module; The acquisition module is configured to obtain a camera image and a radar image corresponding to the external parameters to be calibrated; The conversion module is configured to convert the camera image into a preset standard image; The external parameter calibration module is configured to perform external parameter calibration based on the standard image and the radar image; When converting the camera image into the preset standard image, it is performed according to a pre-constructed conversion mapping, and the conversion mapping is: Among them, (u1, v1) are the pixel coordinates of the standard image, and (u2, v2) are the pixel coordinates of the camera image. and are the focal lengths of the imaging devices corresponding to the standard image and the camera image in units of horizontal pixels, respectively. and are the focal lengths of the imaging devices corresponding to the standard image and the camera image in units of vertical pixels, respectively. and are the horizontal offsets of the principal points of the imaging devices corresponding to the standard image and the camera image on the acquired images, respectively. and are the vertical offsets of the principal points of the imaging devices corresponding to the standard image and the camera image on the acquired images, respectively.

6. An external parameter calibration device for images, characterized in that, Including: A processor; A memory for storing processor-executable instructions; Wherein, the processor is configured to implement the method described in any one of claims 1 to 4 when executing the executable instructions.

7. A non - volatile computer - readable storage medium having computer program instructions stored thereon, characterized in that, The computer program instructions implement the method described in any one of claims 1 to 4 when executed by the processor.

Images