Device positioning method, apparatus, computer equipment, readable storage medium and program product

Abstract

The present application relates to a device positioning method, apparatus, computer equipment, readable storage medium and program product. The method includes: obtaining the current actual position information of the characteristic hole, detecting the current image position information of the characteristic hole from the current characteristic hole image, determining the main shooting coordinate system, and detecting the first pose transformation parameter between any two of all observation coordinate systems, the second pose transformation parameter between each observation coordinate system and the characteristic hole coordinate system of the characteristic hole in the current time period, and the historical zero-bias pose transformation parameter between the main shooting coordinate system and the characteristic hole coordinate system; based on the historical zero-bias pose transformation parameter and the second pose transformation parameter corresponding to the main shooting coordinate system, generating the characteristic hole pose change; based on the above parameters, correcting the characteristic hole pose change; based on the corrected characteristic hole pose change, positioning the target vehicle. The method can accurately locate the target vehicle.

Description

Technical Field

[0001] The present application relates to the technical field of device positioning, and in particular to a device positioning method, apparatus, system, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the development of science and technology, equipment positioning technology is attracting the attention of many people. Taking vehicles as an example, in the welding, painting and other processes of automobile manufacturing, the stable operation of the vehicle depends on the accurate positioning of the vehicle.

[0003] At present, most automobile production lines use the traditional lifting roller bed positioning method to achieve vehicle positioning. Among them, the roller bed on the automobile welding line is a conveying equipment used to transport the car body that has completed welding at the previous station to the next station for further welding. The lifting roller bed positioning is a method of supporting the lifting roller bed to carry the vehicle, so that the vehicle falls into the preset target area along with the roller bed, thereby achieving vehicle positioning.

[0004] However, the current device positioning method is not accurate enough. Summary of the invention

[0005] Based on this, it is necessary to provide an accurate device positioning method, apparatus, system, computer equipment, computer-readable storage medium and computer program product to address the above technical issues.

[0006] In a first aspect, the present application provides a device positioning method, comprising:

[0007] Acquire current actual position information of the characteristic hole and current characteristic hole image, and detect current image position information of the characteristic hole from the current characteristic hole image, where the current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0008] Determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first pose transformation parameters between any two of all the observation coordinate systems, the second pose transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias pose transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position in the historical time period;

[0009] Based on the historical zero bias pose transformation parameters and the second pose transformation parameters corresponding to the main shooting coordinate system, a characteristic hole pose change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period is generated;

[0010] Based on the current actual position information, all current image position information, the first pose transformation parameters, all second pose transformation parameters and the historical zero bias pose transformation parameters, the characteristic hole pose change amount is corrected;

[0011] Based on the corrected characteristic hole posture change, the target vehicle is positioned.

[0012] In a second aspect, the present application also provides a device positioning apparatus, comprising:

[0013] A position acquisition module is used to acquire the current actual position information of the characteristic hole and the current characteristic hole image, and detect the current image position information of the characteristic hole from the current characteristic hole image. The current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0014] A posture parameter detection module is used to determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first posture transformation parameters between any two of all observation coordinate systems, the second posture transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias posture transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system. The historical zero-bias posture transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position in the historical time period;

[0015] A characteristic hole posture change detection module is used to generate a characteristic hole posture change between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period based on the historical zero bias posture transformation parameters and the second posture transformation parameters corresponding to the main shooting coordinate system;

[0016] A correction module is used to correct the change amount of the characteristic hole posture based on the current actual position information, all current image position information, the first posture transformation parameter, all second posture transformation parameters and the historical zero bias posture transformation parameter;

[0017] The positioning module is used to locate the target vehicle based on the corrected characteristic hole posture change.

[0018] In a third aspect, the present application further provides a device positioning system, the system comprising a shooting device, a second delivery device and a controller, the shooting device and the main shooting device are both used to shoot characteristic hole images on the target vehicle, and the controller is used to execute the following device positioning method:

[0019] Acquire current actual position information of the characteristic hole and current characteristic hole image, and detect current image position information of the characteristic hole from the current characteristic hole image, where the current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0020] Determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first pose transformation parameters between any two of all the observation coordinate systems, the second pose transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias pose transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position in the historical time period;

[0021] Based on the historical zero bias pose transformation parameters and the second pose transformation parameters corresponding to the main shooting coordinate system, a characteristic hole pose change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period is generated;

[0022] Based on the current actual position information, all current image position information, the first pose transformation parameters, all second pose transformation parameters and the historical zero bias pose transformation parameters, the characteristic hole pose change amount is corrected;

[0023] Based on the corrected characteristic hole posture change, the target vehicle is positioned.

[0024] In a fourth aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0025] Acquire current actual position information of the characteristic hole and current characteristic hole image, and detect current image position information of the characteristic hole from the current characteristic hole image, where the current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0026] Determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first pose transformation parameters between any two of all the observation coordinate systems, the second pose transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias pose transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position in the historical time period;

[0027] Based on the historical zero bias pose transformation parameters and the second pose transformation parameters corresponding to the main shooting coordinate system, a characteristic hole pose change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period is generated;

[0028] Based on the current actual position information, all current image position information, the first pose transformation parameters, all second pose transformation parameters and the historical zero bias pose transformation parameters, the characteristic hole pose change amount is corrected;

[0029] Based on the corrected characteristic hole posture change, the target vehicle is positioned.

[0030] In a fifth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the following steps are implemented:

[0031] Acquire current actual position information of the characteristic hole and current characteristic hole image, and detect current image position information of the characteristic hole from the current characteristic hole image, where the current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0032] Determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first pose transformation parameters between any two of all the observation coordinate systems, the second pose transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias pose transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position in the historical time period;

[0033] Based on the historical zero bias pose transformation parameters and the second pose transformation parameters corresponding to the main shooting coordinate system, a characteristic hole pose change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period is generated;

[0034] Based on the current actual position information, all current image position information, the first pose transformation parameters, all second pose transformation parameters and the historical zero bias pose transformation parameters, the characteristic hole pose change amount is corrected;

[0035] Based on the corrected characteristic hole posture change, the target vehicle is positioned.

[0036] In a sixth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0037] Acquire current actual position information of the characteristic hole and current characteristic hole image, and detect current image position information of the characteristic hole from the current characteristic hole image, where the current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0038] Determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first pose transformation parameters between any two of all the observation coordinate systems, the second pose transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias pose transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position in the historical time period;

[0039] Based on the historical zero bias pose transformation parameters and the second pose transformation parameters corresponding to the main shooting coordinate system, a characteristic hole pose change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period is generated;

[0040] Based on the current actual position information, all current image position information, the first pose transformation parameters, all second pose transformation parameters and the historical zero bias pose transformation parameters, the characteristic hole pose change amount is corrected;

[0041] Based on the corrected characteristic hole posture change, the target vehicle is positioned.

[0042] The above-mentioned equipment positioning method, device, system, computer equipment, computer-readable storage medium and computer program product, due to the large size of vehicles and other equipment, the process of positioning large vehicles and other equipment will be too complicated, and the complicated positioning process may lead to inaccurate positioning results. Therefore, this application takes vehicles as an example to provide a more accurate equipment positioning method, including: starting from the characteristic holes on the target vehicle, analyzing the characteristic hole images of the characteristic holes taken by all shooting devices in the current time period, and determining the main shooting coordinate system in the observation coordinate system of all shooting devices, combining the historical zero bias posture transformation parameters with the second The pose transformation parameters are used to generate the pose change of the feature hole coordinate system and the main shooting coordinate system between the historical time period and the current time period. The pose change of the feature hole obtained at this time is the initial change. It is necessary to continuously correct the pose change of the feature hole based on the current actual position information, all current image position information, the first pose transformation parameters, all second pose transformation parameters and the pose change between the historical zero bias pose transformation parameters, so that the pose change of the feature hole is closer to the actual pose change. At this time, based on the corrected optimal feature hole pose change, the result of pose positioning of the target vehicle is also more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments of the present application or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 A diagram of an application environment of a device positioning method in an embodiment;

[0045] Figure 2 A schematic diagram of a flow chart of a device positioning method in an embodiment;

[0046] Figure 3 A schematic diagram of the layout of a vehicle and a photographing device in one embodiment;

[0047] Figure 4 It is a flowchart of a device positioning method in another embodiment;

[0048] Figure 5 is a structural block diagram of a device positioning apparatus in one embodiment;

[0049] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are used to explain the present application and are not used to limit the present application.

[0051] The device positioning method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown in the figure, Figure 1 It is an application loop diagram taking four shooting devices as an example. The terminal 102 communicates with the controller 104 through the network, and the controller 104 can communicate with multiple shooting devices 106.

[0052] The user selects a device to be located on the terminal 102, for example, the device may be a target vehicle 108 in the current environment, and triggers a vehicle location operation, so that the terminal 102 sends a vehicle location request of the target vehicle 108 to the controller 104. In the current time period, the controller 104 receives the vehicle location request, and when the target vehicle 108 arrives at a preset area, controls all the shooting devices 106 to shoot the characteristic holes on the target vehicle 108 to generate a current characteristic hole image, and detects the current image position information of the characteristic holes from the current characteristic hole image. In addition, the controller 104 also obtains the current actual position information of the characteristic holes.

[0053] The controller 104 determines the main shooting coordinate system in the observation coordinate system of all shooting devices, detects the first pose transformation parameter between any two of all observation coordinate systems, the historical zero bias pose transformation parameter between the main shooting coordinate system and the feature hole coordinate system of the feature hole when the target vehicle is in the zero bias position in the historical time period, and the second pose transformation parameter between all observation coordinate systems and the feature hole coordinate system in the current time period; based on the historical zero bias pose transformation parameter and the second pose transformation parameter, generates the feature hole pose change amount between the feature hole coordinate system and the main shooting coordinate system between the historical time period and the current time period; based on the current actual position information, all current image position information, the first pose transformation parameter, all second pose transformation parameters and the historical zero bias pose transformation parameter, corrects the feature hole pose change amount; based on the corrected feature hole pose change amount, performs pose positioning on the target vehicle 108.

[0054] Furthermore, the controller 104 may also feed back the position and posture positioning result of the target vehicle 108 to the terminal 102 , which is then displayed to the user by the terminal 102 .

[0055] The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, projection devices, etc. Portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. Head-mounted devices may be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc.

[0056] In an exemplary embodiment, Figure 2 As shown, a device positioning method is provided, which is applied to Figure 1 The controller 104 in FIG. 1 is used as an example for explanation.

[0057] S100, obtaining current actual position information of the characteristic hole and the current characteristic hole image, and detecting current image position information of the characteristic hole from the current characteristic hole image.

[0058] The current characteristic hole image is obtained by all the photographing devices photographing the characteristic holes on the target vehicle in the current time period. The vehicle body image photographed by each photographing device should contain at least 4 characteristic holes, and these characteristic holes should contain at least 2 process holes.

[0059] Feature holes include but are not limited to body features in the shape of circles, rectangles, etc. Feature holes include process holes. Process holes are welding holes made during sheet metal welding to reduce weld nodules and achieve aesthetics or facilitate processing and assembly. In layman's terms, process holes are holes formed due to process needs, with higher processing accuracy and more reliable positioning.

[0060] Specifically, Figure 3 As shown, the present application places multiple shooting devices around the body of the target vehicle, such as 2D (Two-Dimensional) cameras, generally 4, for shooting the relevant characteristic holes of the body. Since the characteristic holes are fixed holes on the body, the position change of the characteristic holes is also the position change of the body. By detecting the position change of the characteristic holes, the position change of the body is determined to realize the positioning of the target vehicle.

[0061] More specifically, in the current time period, when the target vehicle reaches the preset area, all the shooting devices shoot the characteristic holes on the target vehicle to obtain the current characteristic hole image of the characteristic hole. The current image position information of the characteristic hole is detected from the current characteristic hole image, and the image position information is also the two-dimensional position information. Furthermore, the three-dimensional position information of the characteristic hole in the current time period can also be obtained, that is, the current actual position information.

[0062] In an exemplary embodiment, the method of detecting the current image position information of the characteristic hole from the current characteristic hole image includes but is not limited to methods such as Hough transform, template matching, and deep learning to extract the current image position information of the characteristic hole. The current image position information of the characteristic hole is actually the two-dimensional coordinates of the center of the characteristic hole. It should be understood that the center of the characteristic hole is generally a relatively stable geometric feature, such as a "centroid". Further, through methods such as Hough transform, template matching, and deep learning, not only the image position information of the characteristic hole can be detected, but also information such as the hole spacing between the characteristic holes can be obtained.

[0063] In an exemplary embodiment, the current actual position information of the characteristic hole is obtained, including: obtaining the actual three-dimensional position information of the characteristic hole in the current time period based on prior information such as the vehicle body digital model, and in addition, information such as the actual hole spacing between the characteristic holes can also be obtained.

[0064] S200, determining the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detecting the first pose transformation parameters between any two of all the observation coordinate systems, the second pose transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system.

[0065] Among them, the historical zero-bias pose transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position during the historical time period. This application uses more than one coordinate system, for example, the observation coordinate system where the shooting device is located and the characteristic hole coordinate system. When the shooting device is a camera, the observation coordinate system can also be called the camera coordinate system. The characteristic hole coordinate system is the vehicle coordinate system, also known as the three-dimensional world coordinate system.

[0066] Specifically, first, a shooting device is selected from all shooting devices as the main shooting device, and the observation coordinate system corresponding to the main shooting device is used as the main shooting coordinate system.

[0067] Secondly, detect the first pose transformation parameter between any two of all observation coordinate systems. The first pose transformation parameter can be regarded as the pose transformation parameter between the shooting devices, that is, the external calibration parameter of the shooting device. The first pose transformation parameter is a corrected parameter and is an accurate external calibration parameter.

[0068] Detect the second pose transformation parameters between the observation coordinate systems corresponding to all shooting devices in the current time period and the characteristic hole coordinate system.

[0069] In addition, when the vehicle reaches the zero bias position in the preset area during the historical time period, the historical pose transformation parameters between all observation coordinate systems and the characteristic hole coordinate system of the characteristic hole are also detected. Further, the historical pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system of the characteristic hole are obtained from all historical pose transformation parameters.

[0070] In an exemplary embodiment, when obtaining the second posture transformation parameters of the characteristic hole coordinate system relative to the observation coordinate system in the current time period, it can be achieved through the internal calibration parameters of the shooting device, the current image position information of the characteristic hole and the current actual position information, that is, through the internal calibration parameters of the shooting device, the current image position information of the characteristic hole and the current actual position information, the PnP (Perspective-3-Point, 3-point perspective method) problem is solved using methods including but not limited to P3P (Perspective-3-Point, 3-point perspective method), EPnP (Efficient Perspective-n-Point, a solution method for calculating the relationship between the camera's posture and the three-dimensional world coordinates) and other methods to obtain the second posture transformation parameters between the characteristic hole coordinate system and all the observation coordinate systems in the current time period. Furthermore, the process of obtaining the historical posture transformation parameters of the characteristic hole coordinate system relative to the observation coordinate system in the historical time period is the same.

[0071] S300, generating a characteristic hole posture change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period based on the historical zero bias posture transformation parameter and the second posture transformation parameter corresponding to the main shooting coordinate system.

[0072] Specifically, the historical zero-bias pose transformation parameters are the pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system when the target vehicle is at the zero-bias position in the historical time period, and the second pose transformation parameters are the pose transformation parameters between all observation coordinate systems and the characteristic hole coordinate system in the current time period.

[0073] From the second pose transformation parameters, obtain the current second pose transformation parameters between the main shooting coordinate system and the feature hole coordinate system in the current time period. Since the feature hole pose change is essentially the pose change when the main shooting device shoots the feature hole in two different time periods, that is, the feature hole pose change between the feature hole coordinate system and the main shooting coordinate system in two different time periods, in layman's terms, the feature hole pose change is the current second pose transformation parameters between the feature hole coordinate system and the main shooting coordinate system in the current time period, and the difference between the historical zero bias pose transformation parameters between the feature hole coordinate system and the main shooting coordinate system in the historical time period.

[0074] Based on the current second pose transformation parameters between the characteristic hole coordinate system and the main shooting coordinate system in the current time period and the difference between the historical zero bias pose transformation parameters between the characteristic hole coordinate system and the main shooting coordinate system in the historical time period, the characteristic hole pose change between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period is generated.

[0075] S400, based on the current actual position information, all current image position information, the first pose transformation parameter, all second pose transformation parameters and the historical zero bias pose transformation parameters, the characteristic hole pose change amount is corrected.

[0076] Specifically, based on the current actual position information, all current image position information, the first pose transformation parameter, all second pose transformation parameters, and historical zero-bias pose transformation parameters, a nonlinear least squares problem is constructed, and the characteristic hole pose change is accurately corrected by minimizing the positioning error. Furthermore, while correcting the characteristic hole pose change, multiple pose transformation parameters such as the first pose transformation parameter, all second pose transformation parameters, and historical zero-bias pose transformation parameters are actually corrected at the same time, so that the characteristic hole pose change finally obtained based on these parameters is more accurate. Furthermore, the positioning error includes but is not limited to reprojection error and pose error, and the characteristic hole pose change is accurately corrected by minimizing the reprojection error and pose error.

[0077] S500: Positioning the target vehicle based on the corrected characteristic hole position change.

[0078] Specifically, since the characteristic hole is a fixed characteristic hole on the target vehicle, the corrected characteristic hole posture change is also the vehicle posture change. Based on the vehicle posture change, the target vehicle is accurately positioned.

[0079] Furthermore, the change in posture is essentially the change in position and angle. Based on the change in vehicle posture, not only can the position of the target vehicle be located, but also the angle of the target vehicle can be determined to obtain the posture information of the target vehicle. Therefore, positioning the target vehicle based on the change in vehicle posture is actually positioning the target vehicle based on the change in vehicle position, and positioning the target vehicle based on the change in vehicle angle, so as to generate a posture positioning result based on the position positioning result and the angle positioning result.

[0080] In the above-mentioned equipment positioning method, since vehicles and other equipment are too large, the process of positioning large vehicles and other equipment will be too complicated, and the complicated positioning process may lead to inaccurate positioning results. For this reason, the present application takes vehicles as an example and provides a more accurate equipment positioning method, including: starting from the characteristic hole on the target vehicle, analyzing the characteristic hole images of the characteristic holes taken by all shooting devices in the current time period, and determining the main shooting coordinate system in the observation coordinate system of all shooting devices, combining the historical zero-bias posture transformation parameters and the second posture transformation parameters, generating the characteristic hole posture change between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period. The characteristic hole posture change obtained at this time is the initial change. It is necessary to continuously correct the characteristic hole posture change based on the current actual position information, all current image position information, the first posture transformation parameters, all second posture transformation parameters and the posture change between the historical zero-bias posture transformation parameters, so that the characteristic hole posture change is closer to the actual posture change. At this time, based on the corrected optimal characteristic hole posture change, the result of posture positioning of the target vehicle is also more accurate.

[0081] In an exemplary embodiment, based on the current actual position information, all current image position information, the first pose transformation parameter, all second pose transformation parameters and the historical zero bias pose transformation parameter, the characteristic hole pose change amount is corrected, including:

[0082] According to the current actual position information, all current image position information and the first pose transformation parameter, all the second pose transformation parameters under the first target reprojection error are corrected, and according to the historical zero bias pose transformation parameter and all the corrected second pose transformation parameters, the characteristic hole pose change under the first target pose error is corrected.

[0083] Among them, a first fusion error between the first target reprojection error and the first target posture error is less than a first preset error value.

[0084] Specifically, the first pose transformation parameter is an optimized and accurate external calibration parameter of the shooting device. The second pose transformation parameter is a pose transformation parameter between all observation coordinate systems and the characteristic hole coordinate system in the current time period.

[0085] In the process of correcting the characteristic hole posture change amount based on the current actual position information, all current image position information, the first pose transformation parameter, all second pose transformation parameters and the historical zero-bias pose transformation parameter, it is essentially based on the current actual position information, all current image position information, the first pose transformation parameter and the historical zero-bias pose transformation parameter, all the second pose transformation parameters and the characteristic hole posture change amount are corrected at the same time. That is to say, while continuously correcting all the second pose transformation parameters according to the current actual position information, all current image position information and the first pose transformation parameter, it is possible to continuously correct the characteristic hole posture change amount based on the optimized all second pose transformation parameters and the historical zero-bias pose transformation parameter, so that all the second pose transformation parameters and the characteristic hole posture change amount are finally at the optimal value, and the corrected characteristic hole posture change amount is obtained when all the second pose transformation parameters are at the optimal state.

[0086] Furthermore, based on the current actual position information, all current image position information and the first pose transformation parameters, the second pose transformation parameters under the first target reprojection error are detected. The bundle adjustment method is used, and the initial second pose transformation parameters are used as initial values. The second pose transformation parameters are iteratively optimized to obtain accurate corrected second pose transformation parameters.

[0087] The bundle adjustment method is a nonlinear least squares problem. It optimizes the position of the three-dimensional point and the posture of the shooting device by minimizing the reprojection error. That is, it corrects the second posture transformation parameters under the reprojection error of the first target and optimizes all the second posture transformation parameters between each observation coordinate system and the characteristic hole coordinate system.

[0088] Among them, reprojection error is an important concept in the field of computer vision, which is used to evaluate the accuracy of tasks such as camera calibration, 3D reconstruction and motion estimation. The calculation of reprojection error is based on the difference between the projection of a 3D space point on the image plane and the pixel point obtained by reprojecting it through the estimated camera parameters.

[0089] Furthermore, based on the historical zero-bias posture transformation parameters and all the corrected second posture transformation parameters, the characteristic hole posture change under the first target posture error is corrected. At this time, a nonlinear least squares problem is constructed through the historical zero-bias posture transformation parameters and all the corrected second posture transformation parameters, and the characteristic hole posture change under the first target posture error is corrected by minimizing the posture error. Furthermore, the above nonlinear least squares problem can be solved by using the Gauss-Newton method and the Levenberg-Marquardt method to obtain an accurate characteristic hole posture change.

[0090] Based on the above analysis, the correction of the characteristic hole posture change is actually achieved by minimizing the first positioning error, wherein the first positioning error includes the first target reprojection error and the first target posture error, and minimizing the first positioning error means that the second fusion error between the first target reprojection error and the first target posture error is less than the first preset error value. When the first fusion error between the first target reprojection error and the first target posture error is less than the first preset error value, it is considered that the first positioning error at this time is an appropriate error, and the characteristic hole posture change at this time is corrected to an accurate characteristic hole posture change.

[0091] In the above embodiment, by minimizing the first target reprojection error based on the current actual position information and the current image position information, all second posture transformation parameters are accurately corrected, and then based on minimizing the first target posture error and accurately correcting all second posture transformation parameters, the process of correcting the characteristic hole posture change is also more accurate, thereby improving the accuracy of posture positioning of the target vehicle.

[0092] In an exemplary embodiment, according to the current actual position information, all current image position information and the first pose transformation parameter, correcting all second pose transformation parameters under the first target reprojection error includes:

[0093] For any shooting device, internal calibration parameters are obtained, and based on the internal calibration parameters, current actual position information, current image position information and the first pose transformation parameters, the second pose transformation parameters under the first target reprojection error are corrected to obtain the corrected second pose transformation parameters.

[0094] According to the historical zero bias posture transformation parameters and all the corrected second posture transformation parameters, the hole posture change under the first target posture error is corrected, including:

[0095] A third posture transformation parameter is obtained from all corrected second posture transformation parameters, wherein the third posture transformation parameter is a posture transformation parameter between the main shooting coordinate system and the characteristic hole coordinate system; based on the historical zero bias posture transformation parameter and the third posture transformation parameter, the hole posture change under the first target posture error is corrected to obtain a corrected hole posture transformation.

[0096] The corrected second posture transformation parameter is the posture transformation parameter when the reprojection error of the first sub-target is less than the second preset error value. The corrected hole posture transformation amount is the hole posture change amount when the first target posture error is less than the third preset error value.

[0097] Specifically, for the first target reprojection error alone, the first target reprojection error represents the error between the current image position information of the feature hole and the current actual position information of the feature hole projected to the image coordinate system of the shooting device. Therefore, the first target reprojection error can be expressed as: i -π(T nk T k P i ).

[0098] Where n represents the type of camera, and the current actual position information of the i-th feature hole is P i , the current image position information of the i-th feature hole is p i , the second pose transformation parameter of the characteristic hole coordinate system relative to the nth observation coordinate system in the current time period is T n For example, the second pose transformation parameter of the vehicle coordinate system relative to the jth observation coordinate system in the current time period is T j , the second pose transformation parameter of the vehicle coordinate system relative to the mth observation coordinate system in the current time period is T m ,π() represents the internal calibration parameters of the nth camera, that is, the projection function, T nk is the first pose transformation parameter between the nth observation coordinate system and the kth observation coordinate system, where the kth observation coordinate system is the main shooting coordinate system.

[0099] The specific explanation of the first target reprojection error is: in the current time period, the current actual position information of the i-th feature hole in the feature hole coordinate system is projected into the main shooting coordinate system through the second pose transformation parameter of the feature hole coordinate system relative to the main shooting coordinate system, and then the current actual position information is projected into the n1-th observation coordinate system through the first pose transformation parameter between the main shooting coordinate system and the n1-th observation coordinate system, and then based on the internal calibration parameters of the n1-th shooting device, the current actual position information is projected into the pixel coordinate system of the n-th shooting device to obtain the current projection position information corresponding to the n1-th shooting device. The difference between the current image position information and the current projection position information in the current feature hole image captured by the n1-th shooting device is taken as the first target reprojection error of the n1-th shooting device.

[0100] Therefore, based on the internal calibration parameters, the current actual position information, the current image position information and the first pose transformation parameters, all second pose transformation parameters can be corrected based on the internal calibration parameters, the current actual position information, the current image position information and the first pose transformation parameters to obtain the corrected second pose transformation parameters corresponding to each second pose transformation parameter when the first target reprojection error is minimized.

[0101] Furthermore, after correcting the second pose transformation parameters between all observation coordinate systems and the characteristic hole coordinate system in the current time period, since the main shooting device is one of the shooting devices, that is, the main shooting coordinate system is one of the observation coordinate systems, the third pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system can also be obtained from all corrected second pose transformation parameters. Based on the historical zero bias pose transformation parameters and the third pose transformation parameters, the hole pose change under the first target pose error is corrected to obtain the corrected hole pose transformation.

[0102] Specifically, for the first target pose error, taking the main shooting coordinate system as the kth observation coordinate system as an example, the first target pose error can be expressed as log(T k1 -1 T k2 T car -1 ), where T k1 -1 is the historical zero bias pose transformation parameter of the characteristic hole coordinate system relative to the kth observation coordinate system in the historical time period, T k2 is the third pose transformation parameter of the characteristic hole coordinate system relative to the kth observation coordinate system in the current time period, T car It is the change of the characteristic hole position and posture between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period.

[0103] Through the above method, when the historical zero-bias posture transformation parameters and the third posture transformation parameters are obtained, the hole posture change under the first target posture error can be corrected to obtain the first posture transformation parameter when the first target posture error is the smallest. At this time, the corrected hole posture change is the hole posture change when the first target posture error is less than the third preset error value, which is more accurate.

[0104] In an exemplary embodiment, the first target reprojection error and the first target pose error are two least squares problems. The first target reprojection error and the first target pose error can be merged together, and all second pose transformation parameters and hole pose transformation amounts can be jointly corrected to obtain the optimal all second pose transformation parameters and hole pose transformation amounts.

[0105] That is to say, the first target reprojection error and the first target pose error are combined to obtain the first positioning error. The expression of the first positioning error in the least square form is:

[0106] Min{Σ(||p i -π(T nk T k P i )||2 2 +||log(T k1 -1 T k2 T car -1 )||2 2}.

[0107] Among them, the first term in the formula represents the reprojection error of the first target, the second term in the formula represents the pose error of the first target, and log() represents the conversion of the pose into a six-dimensional vector.

[0108] The norms of the first target reprojection error and the first target pose error are calculated respectively, and then the first positioning error is obtained based on the sum of the squares of the norms of the two, and all the second pose transformation parameters and the hole pose transformation amount when the first positioning error is minimized are calculated.

[0109] In the above embodiment, based on the internal calibration parameters, the current actual position information, the current image position information and the first pose transformation parameters, the technical means of minimizing the first target reprojection error is utilized to accurately correct all the second pose transformation parameters, and then based on minimizing the first target pose error and accurately all the corrected second pose transformation parameters, the process of correcting the hole pose transformation amount is also more accurate, thereby improving the accuracy of vehicle positioning.

[0110] In an exemplary embodiment, Figure 4 As shown, S200 includes:

[0111] S210: Determine a main shooting coordinate system in the observation coordinate systems of all shooting devices.

[0112] S220, obtaining historical actual position information of the characteristic hole and historical characteristic hole images, and detecting historical image position information of the characteristic hole from the historical characteristic hole images.

[0113] S230, detecting historical pose transformation parameters between each observation coordinate system and the characteristic hole coordinate system in the historical time period, and initial first pose transformation parameters between any two of all observation coordinate systems.

[0114] S240, based on the historical actual position information, all historical image position information and all historical posture transformation parameters, all initial first-position transformation parameters are corrected to obtain all first-position transformation parameters.

[0115] S250, detecting the second pose transformation parameters between all observation coordinate systems and the characteristic hole coordinate system in the current time period and the historical zero bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system.

[0116] The historical characteristic hole images are obtained by photographing the characteristic holes on the target vehicle by all photographing devices within a historical time period.

[0117] Specifically, in the historical time period, when the target vehicle reaches the zero offset position in the preset area, the characteristic hole on the target vehicle is photographed by all the photographing devices to obtain the historical characteristic hole image of the characteristic hole. The historical image position information of the characteristic hole is detected from the historical characteristic hole image. Furthermore, the three-dimensional position information of the characteristic hole in the historical time period, that is, the historical actual position information, can also be obtained.

[0118] Furthermore, the method of detecting the historical image position information of the characteristic hole from the historical characteristic hole image is similar to the method of detecting the current image position information of the characteristic hole in the current time period, including but not limited to Hough transform, template matching, deep learning and other methods to extract the historical image position information of the characteristic hole.

[0119] The historical actual position information of the characteristic holes is obtained, including: obtaining the historical three-dimensional position information of the characteristic holes in the historical time period according to the prior information such as the vehicle body digital model in the historical time period, and in addition, the historical actual hole spacing information between the characteristic holes can also be obtained.

[0120] Furthermore, the historical pose transformation parameters between each observation coordinate system and the characteristic hole coordinate system in the historical time period, as well as the initial first pose transformation parameters between any two of all observation coordinate systems are detected. Both the historical pose transformation parameters and the initial first pose transformation parameters are not very accurate pose transformation parameters, and they need to be optimized to obtain accurate pose transformation parameters.

[0121] For the first pose transformation parameter, the first pose transformation parameter is essentially the external calibration parameter of the camera. For the external calibration parameter calibration of multiple cameras, a laser tracker and a tracking ball can be used to indirectly calibrate the external calibration parameters between cameras, but this method is costly and cumbersome. Therefore, the technical solution does not require the use of external equipment, and only the body feature hole data captured by the camera can be used to calibrate the external calibration parameters of the camera.

[0122] The method for correcting the initial first pose transformation parameters can be to correct the initial first pose transformation parameters based on the historical actual position information, all historical image position information and all historical pose transformation parameters to obtain the first pose transformation parameters. In this process, the historical pose transformation parameters and the initial first pose transformation parameters are actually corrected at the same time to obtain the first pose transformation parameters when the historical pose transformation parameters are optimal. The first pose transformation parameters at this time are the corrected initial first pose transformation parameters.

[0123] Furthermore, since the initial first pose transformation parameters are the pose transformation parameters between any two of all observation coordinate systems, and the historical pose transformation parameters are the historical pose transformation parameters between each observation coordinate system and the characteristic hole coordinate system, when optimizing the initial first pose transformation parameters between any two of all observation coordinate systems, the historical pose transformation parameters of the corresponding two observation coordinate systems are selected from all historical pose transformation parameters to optimize the initial first pose transformation parameters. For example, when optimizing the initial first pose transformation parameters between the kth observation coordinate system and the jth observation coordinate system, the historical pose transformation parameters of the kth observation coordinate system and the historical pose transformation parameters of the jth observation coordinate system are selected from all historical pose transformation parameters to optimize the initial first pose transformation parameters.

[0124] In an exemplary embodiment, the method for obtaining the historical pose transformation parameters between the observation coordinate system and the characteristic hole coordinate system of a certain shooting device includes: obtaining the internal calibration parameters of the shooting device, solving the PnP problem through P3P, EPnP and other methods based on the internal calibration parameters corresponding to the shooting device, historical image position information and historical actual position information, and detecting the historical pose transformation parameters of the characteristic hole coordinate system relative to the observation coordinate system of the shooting device. At this time, the historical pose transformation parameters obtained are also the initial pose, which is an uncorrected pose and has a certain inaccuracy.

[0125] That is to say, based on the historical actual position information, all historical image position information and all historical pose transformation parameters, the initial first pose transformation parameters are corrected. When the first pose transformation parameters are obtained, all historical pose transformation parameters and the initial first pose transformation parameters are corrected simultaneously. While continuously correcting all historical pose transformation parameters, the initial first pose transformation parameters can be continuously corrected based on all optimized historical pose transformation parameters. Finally, all historical pose transformation parameters and the initial first pose transformation parameters are at optimal values, and the first pose transformation parameters are obtained when all historical pose transformation parameters are at the optimal value.

[0126] In an exemplary embodiment, the calibration of multiple shooting devices includes the internal parameter calibration of each shooting device and the external parameter calibration between the shooting devices. The calibration of the shooting devices can be carried out after the installation position of the shooting devices is determined and the working distance and field of view of the shooting devices are adjusted.

[0127] The external parameter calibration between the shooting devices has been mentioned above, and the calibration of the internal calibration parameters of the shooting devices can be calibrated using calibration plates including but not limited to Chessboard, ChArUco, CCTag, etc. In order to ensure that the working distance of the camera during calibration is as consistent as possible with the working distance in actual application, the calibration plate is set near the body process hole photographed by the shooting device. The calibration plate constantly changes its posture. At the same time, the shooting device collects the calibration plate image, and uses an open source calibration algorithm to complete the calibration of the internal calibration parameters of each shooting device. Among them, the calibration plate is a flat plate with a fixed-pitch pattern array. By shooting the flat plate with a fixed-pitch pattern array by the shooting device and calculating it with the calibration algorithm, the internal parameter geometric model of the shooting device can be obtained, thereby obtaining high-precision measurement and reconstruction results.

[0128] In the above embodiment, by analyzing the historical characteristic hole images taken by the shooting device within the historical time period, combined with the historical posture transformation parameters between each observation coordinate system and the characteristic hole coordinate system within the historical time period, all historical posture transformation parameters and the initial first posture transformation parameters are continuously corrected, so that the first posture transformation parameters are obtained when all historical posture transformation parameters are optimal, that is, the posture transformation parameters between the two shooting devices are closer to the actual relative posture, and then based on the first posture transformation parameters as the accurate external calibration parameters of the shooting device, the subsequent posture positioning results of the target vehicle are also more accurate.

[0129] In an exemplary embodiment, based on the historical actual position information, all historical image position information and all historical pose transformation parameters, all initial first pose transformation parameters are modified, including:

[0130] Based on the historical actual position information and all historical image position information, all historical pose transformation parameters under the second target reprojection error are corrected, and based on all corrected historical pose transformation parameters, all initial first pose transformation parameters under the second target pose error are corrected.

[0131] Among them, the second fusion error between the second target reprojection error and the second target posture error is less than the fourth preset error value.

[0132] Specifically, based on the internal calibration parameters of the shooting equipment, the historical actual position information and historical image position information of the characteristic hole, the bundle adjustment method is used, and the historical pose transformation parameters are taken as initial values. The historical pose transformation parameters are iteratively optimized to obtain accurate historical pose transformation parameters.

[0133] That is, through the bundle adjustment method, all historical pose transformation parameters under the second target reprojection error are corrected to optimize the historical pose transformation parameters between the characteristic hole coordinate system and the observation coordinate system.

[0134] Furthermore, based on all corrected historical pose transformation parameters, the first pose transformation parameter under the second target pose error is corrected. At this time, a nonlinear least squares problem is constructed through all corrected historical pose transformation parameters, and the first pose transformation parameter under the second target pose error is corrected by minimizing the pose error. Furthermore, the Gauss-Newton method and the Levenberg-Marquardt method can be used to solve the above nonlinear least squares problem to obtain accurate first pose transformation parameters.

[0135] Based on the above analysis, the correction of the initial first pose transformation parameter is actually achieved by minimizing the second positioning error, wherein the second positioning error includes the second target reprojection error and the second target pose error, and minimizing the second positioning error means that the second fusion error between the second target reprojection error and the second target pose error is less than the fourth preset error value. When the second fusion error between the second target reprojection error and the second target pose error is less than the fourth preset error value, it is considered that the second positioning error at this time is an appropriate error, and the initial first pose transformation parameter at this time is corrected to an accurate first pose transformation parameter.

[0136] In an exemplary embodiment, for the second target reprojection error alone, the second target reprojection error represents the error between the historical image position information of the characteristic hole and the image coordinate system of the projection of the historical actual position information of the characteristic hole to the shooting device. Therefore, the second target reprojection error can be expressed as: i ’ -π(T n ’ P i ’ ).

[0137] Where n represents the type of shooting equipment, and the historical actual position information of the i-th feature hole is P i ’ , the historical image position information of the i-th feature hole is p i ’ , the historical pose transformation parameter of the characteristic hole coordinate system relative to the nth observation coordinate system in the historical time period is T n ’ For example, the historical pose transformation parameter of the vehicle coordinate system relative to the jth observation coordinate system in the historical time period is T j ’ , the historical pose transformation parameter of the vehicle coordinate system relative to the kth observation coordinate system in the historical time period is T k ’ ,π() represents the internal calibration parameters of the nth camera device, that is, the projection function.

[0138] The specific explanation of the second target reprojection error is: in the historical time period, the historical actual position information of the i-th feature hole in the feature hole coordinate system is projected into the n-th observation coordinate system through the historical pose transformation parameters of the feature hole coordinate system relative to the n-th observation coordinate system, and then the historical actual position information of the i-th feature hole in the feature hole coordinate system is projected into the pixel coordinate system of the n-th shooting device through the internal calibration parameters of the n-th observation coordinate system to obtain the projection position information corresponding to the n-th shooting device. Then the difference between the historical image position information and the projection position information in the historical feature hole image shot by the n-th shooting device is taken as the second target reprojection error of the n-th shooting device.

[0139] Therefore, based on obtaining the historical actual position information, all historical image position information and internal calibration parameters, all historical pose transformation parameters can be corrected based on the historical actual position information, all historical image position information and internal calibration parameters to obtain all historical pose transformation parameters when the reprojection error of the second target is minimized.

[0140] In an exemplary embodiment, for the second target pose error alone, taking the initial first pose transformation parameters as the kth observation coordinate system and the jth observation coordinate system as an example, the second target pose error can be represented as log(T j ’ T k -1’ T jk -1 ), where T j ’ is the historical pose transformation parameter of the characteristic hole coordinate system relative to the jth observation coordinate system in the historical time period, T k ’ is the historical pose transformation parameter of the characteristic hole coordinate system relative to the kth observation coordinate system in the historical time period, T jk is the first pose transformation parameter between the j-th coordinate system and the k-th coordinate system.

[0141] When all corrected historical pose transformation parameters are obtained, the initial first pose transformation parameters under the second target pose error are corrected based on all corrected historical pose transformation parameters to obtain the first pose transformation parameters when the second target pose error is the smallest. At this time, the first pose transformation parameters are a more accurate pose.

[0142] In an exemplary embodiment, the second target reprojection error and the second target pose error are two least squares problems. The second target reprojection error and the second target pose error can be merged together, and all historical pose transformation parameters and the initial first pose transformation parameters can be jointly corrected to obtain the optimal all historical pose transformation parameters and the initial first pose transformation parameters.

[0143] That is to say, the second target reprojection error and the second target pose error are combined to obtain the second positioning error. Taking the first pose transformation parameter between the kth observation coordinate system and the jth observation coordinate system as an example, the expression of the second positioning error represented by the least squares form is:

[0144] Min{Σ||p i ’ -π(T n ’ P i ’ )||2 2 +||log(T j ’ T k -1’ T jk -1 )||2 2}, n=j, k.

[0145] Among them, the first term in the formula represents the reprojection error of the second target, the second term in the formula represents the pose error of the second target, and log() represents the conversion of the pose into a six-dimensional vector.

[0146] The norms of the second target reprojection error and the second target pose error are calculated respectively, and then the second positioning error is obtained based on the sum of the squares of the norms of the two, and all historical pose transformation parameters and the initial first pose transformation parameters when the second positioning error is minimum are calculated.

[0147] In the above embodiment, by minimizing the second target reprojection error based on the historical actual position information and the historical image position information, all historical pose transformation parameters are accurately corrected, and then based on minimizing the second target pose error and accurately determining all historical pose transformation parameters, the process of correcting the initial first pose transformation parameters is also more accurate, thereby improving the accuracy of obtaining the first pose transformation parameters between the observation coordinate systems.

[0148] In an exemplary embodiment, positioning the target vehicle based on the corrected characteristic hole position change includes:

[0149] The corrected characteristic hole posture change is determined as the vehicle posture change of the target vehicle, the standard vehicle posture information when the target vehicle is in the zero bias position is obtained, and the posture of the target vehicle is positioned based on the standard vehicle posture information and the vehicle posture change.

[0150] Specifically, since the characteristic hole is a fixed hole on the vehicle, the corrected characteristic hole position change is also the vehicle position change of the target vehicle.

[0151] Furthermore, the vehicle posture change includes the vehicle position change and the vehicle angle change. Therefore, in the process of obtaining the standard vehicle posture information when the target vehicle is in the zero bias position and positioning the target vehicle based on the standard vehicle posture information and the vehicle posture change, firstly, the target vehicle is positioned through the standard vehicle position information and the vehicle position change. Secondly, the standard vehicle angle information can also be obtained, and the target vehicle is positioned based on the standard vehicle position information and the vehicle angle change. The target posture information of the target vehicle is generated based on the results of the position positioning and angle judgment of the target vehicle.

[0152] Among them, the standard vehicle angle information and the standard vehicle position information are standard posture information obtained when a vehicle of a certain type reaches the zero bias position in a preset area for the first time, also known as zero bias posture information.

[0153] In the above embodiment, by combining the standard posture information obtained when the vehicle first reaches the zero offset position in the preset area with the vehicle posture change, the position and angle information of the target vehicle can be accurately determined when it subsequently reaches the preset area, thereby obtaining accurate posture information of the target vehicle.

[0154] In some embodiments, taking the device as a vehicle and the shooting device as a camera as an example, the device positioning method of the present application places multiple 2D cameras (generally 4) around the vehicle body to shoot the relevant characteristic holes of the vehicle body, and calculates the change of the vehicle body relative to the zero offset position by comparing the position change of the characteristic holes, thereby realizing the positioning of the target vehicle. The result can be used to monitor whether the arrival of the target vehicle exceeds the tolerance, and can also be used for the operation trajectory compensation of the relevant robot.

[0155] The following will illustrate the above positioning method through a detailed application example. Specifically, the implementation of this method is divided into two steps: first calibrate multiple cameras and then locate the target vehicle.

[0156] 1. Calibration of multiple cameras: The calibration of multiple cameras includes the internal parameter calibration of each camera and the external parameter calibration between cameras. The camera calibration work can be carried out after determining the installation position of the camera and adjusting the working distance and field of view of the camera.

[0157] 1.1. Intrinsic calibration: The intrinsic calibration of the camera can use calibration boards including but not limited to Chessboard, ChArUco, CCTag, etc. The calibration board constantly changes its posture near the characteristic holes of the car body photographed by the camera. At the same time, the camera collects the image of the calibration board. The intrinsic calibration of each camera can be completed using an open source calibration algorithm.

[0158] 1.2. External parameter calibration: For the external parameter calibration of multiple cameras, laser trackers and tracking balls can be used to indirectly calibrate the external parameters between cameras, but this method is costly and cumbersome. This technical solution does not require the use of external equipment, and only requires the body feature hole data captured by the camera to achieve external parameter calibration between cameras.

[0159] Specifically, during the historical time period, when the vehicle body is in place and stops, the camera completes the acquisition of the characteristic hole image on the vehicle body, and the above steps are repeated until a sufficient number of characteristic hole image data on the vehicle body are acquired.

[0160] The specific process is as follows:

[0161] 1. According to the prior information such as the digital model of the vehicle body, the three-dimensional historical actual position information, hole spacing and other information of the characteristic holes are obtained.

[0162] 2. All cameras shoot the characteristic holes on the target vehicle to obtain historical characteristic hole images. The historical image position information of the characteristic holes in the historical characteristic hole images is extracted using methods including but not limited to Hough transform, template matching, deep learning, etc., that is, the two-dimensional hole center coordinates. It should be understood that the center of the characteristic hole is generally a relatively stable geometric feature, such as the "centroid".

[0163] 3. Based on the camera intrinsic parameters, the two-dimensional coordinates and the three-dimensional coordinates of the above-mentioned cameras, the PnP problem is solved using methods including but not limited to P3P and EPnP to obtain the initial historical pose transformation parameters between each camera coordinate system and the characteristic hole coordinate system.

[0164] 4. Based on the camera intrinsic parameters, the two-dimensional coordinates and three-dimensional coordinates of the above-mentioned cameras, the bundle adjustment method is used to iterate with the initial historical pose transformation parameters obtained above as the initial values ​​to obtain the accurate historical pose transformation parameters between each camera coordinate system and the characteristic hole coordinate system. It should be understood that the bundle adjustment method is a nonlinear least squares problem that optimizes the position of the three-dimensional point and the posture of the camera by minimizing the reprojection error.

[0165] 5. Perform steps 1 to 4 for all images taken by each camera to obtain the accurate pose between each camera coordinate system and the characteristic hole coordinate system on the vehicle body at each observation. Using these pose relationships, a nonlinear least squares problem can be constructed. By minimizing the pose error, the first pose transformation parameter between two cameras (i.e., the external parameter between two cameras) can be obtained.

[0166] 6. The Gauss-Newton method and the Levenberg-Marquardt method can be used to solve the above nonlinear least squares problem, thereby completing the external parameter calibration between the two cameras.

[0167] Taking the calibration of the external parameters between the kth camera and the jth camera as an example, the constrained optimization problem can be expressed as:

[0168] Min{Σ||p i ’ -π(T n ’ P i ’ )||2 2 +||log(T j ’ T k -1’ T jk -1 )||2 2}, n=j, k.

[0169] Where log() represents the conversion of the pose into a six-dimensional vector. n represents the type of shooting equipment, and the historical actual position information of the i-th feature hole is P i ’ , the historical image position information of the i-th feature hole is p i ’ , the historical pose transformation parameter of the characteristic hole coordinate system relative to the nth observation coordinate system in the historical time period is T n ’ , π() represents the internal calibration parameters of the nth camera, that is, the projection function. j ’ is the historical pose transformation parameter of the characteristic hole coordinate system relative to the jth observation coordinate system in the historical time period, T k ’ is the historical pose transformation parameter of the characteristic hole coordinate system relative to the kth observation coordinate system in the historical time period, T jk is the first pose transformation parameter between the j-th coordinate system and the k-th coordinate system.

[0170] It should be explained that the first term in the formula is the second target reprojection error, and the second term in the formula is the second target pose error. The second target reprojection error and the second target pose error are two least squares problems. Combine them together and solve T together. n (historical pose transformation parameters between the vehicle coordinate system and the camera coordinate system) and T jk (the first pose transformation parameter between the camera coordinate system and the camera coordinate system) so that the final T n and T jk They are all optimal solutions.

[0171] 2. Positioning of the target vehicle: After completing the calibration of the camera's internal and external parameters, the positioning of the target vehicle can be carried out. For a certain vehicle type, the state when it first arrives at the preset area is recorded as the zero-bias posture, and the camera collects the characteristic hole image at the same time. In addition, the historical zero-bias characteristic hole image when the target vehicle is in the zero-bias position can also be selected from the historical characteristic hole images. When the vehicle body reaches the preset area again, the camera collects the characteristic hole image again to realize the positioning calculation of the target vehicle.

[0172] The specific steps are as follows:

[0173] 1. According to the prior information such as the digital model of the vehicle body when reaching the preset area, the three-dimensional current actual position information and hole spacing of the characteristic hole are obtained.

[0174] 2. All cameras shoot the characteristic holes on the target vehicle to obtain the current characteristic hole image. The two-dimensional hole center coordinates of the characteristic holes in the current characteristic hole image are extracted using methods including but not limited to Hough transform, template matching, deep learning, etc., that is, the current image position information.

[0175] 3. Based on the camera intrinsic parameters, the two-dimensional coordinates and the three-dimensional coordinates of the above-mentioned cameras, the PnP problem is solved using methods including but not limited to P3P and EPnP to obtain the initial second pose transformation parameters between each camera coordinate system and the characteristic hole coordinate system in the current time period.

[0176] 4. Based on the camera intrinsic parameters of each camera, the accurate extrinsic parameters obtained in the first step, the two-dimensional coordinates and three-dimensional coordinates of the characteristic hole, the bundle adjustment method is used to iterate with the initial second pose transformation parameters obtained above as the initial values ​​to obtain the accurate second pose transformation parameters between each camera coordinate system and the characteristic hole coordinate system.

[0177] 5. Perform steps 1 to 4 for all images taken by each camera to obtain the accurate second pose transformation parameters between each camera and the vehicle body at each observation. Using these pose relationships and combining other parameters, a nonlinear least squares problem can be constructed. By minimizing the pose error, the pose change of the characteristic hole can be obtained.

[0178] 6. The Gauss-Newton method and the Levenberg-Marquardt method can be used to solve the above nonlinear least squares problem, so as to obtain the change in the body posture. For example, the constrained optimization problem can be expressed as: Min{Σ(||p i -π(T nk T k P i )||2 2 +||log(T k1 -1 T k2 T car-1 )||2 2}.

[0179] Where n represents the type of camera, and the current actual position information of the i-th feature hole is P i , the current image position information of the i-th feature hole is p i , the second pose transformation parameter of the characteristic hole coordinate system relative to the nth observation coordinate system in the current time period is T n For example, the second pose transformation parameter of the vehicle coordinate system relative to the jth observation coordinate system in the current time period is T j , the second pose transformation parameter of the vehicle coordinate system relative to the mth observation coordinate system in the current time period is T m ,π() represents the internal calibration parameters of the nth camera, that is, the projection function, T nk is the first position transformation parameter between the nth observation coordinate system and the kth observation coordinate system, where the kth observation coordinate system is the main shooting coordinate system. k1 -1 is the historical zero bias pose transformation parameter of the characteristic hole coordinate system relative to the kth observation coordinate system in the historical time period, T k2 is the third pose transformation parameter of the characteristic hole coordinate system relative to the kth observation coordinate system in the current time period, T car is the change in the characteristic hole pose between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period, π() represents the projection function of the camera, and log() represents the conversion of the pose into a six-dimensional vector.

[0180] The first term in the formula is the first target reprojection error, and the second term in the formula is the first target pose error. Through the constrained optimization problem constructed, all second pose transformation parameters are corrected by the first target reprojection error to obtain all second pose transformation parameters when the first target reprojection error constraint is minimized. Then, based on all the corrected second pose transformation parameters and the historical zero-bias pose transformation parameters, the pose change of the characteristic hole is optimized to obtain the pose change of the characteristic hole that meets the constraint conditions, that is, the pose change of the characteristic hole when the second pose error constraint is minimized. The pose change of the characteristic hole of the process hole is determined as the pose change of the vehicle, so as to combine the pose change with the zero-bias pose to locate the pose of the target vehicle.

[0181] In an exemplary embodiment, the nonlinear least squares problem constructed by integrating the camera calibration step and the target vehicle positioning step can also be expressed as:

[0182] Min{Σ(||p i -π(T nk T k P i)||2 2 +Σ||log(T n ’ T k -1’ T nk -1 )||2 2 +||log(T k1 -1 T k2 T car -1 )||2 2}.

[0183] Among them, the first term is the reprojection error, the second term is the external parameter error, and the third term is the pose error.

[0184] Compared with the current technology, the beneficial effects of this application are:

[0185] 1. When calibrating the camera's external parameters, the existing technology usually requires expensive equipment such as laser trackers, and the calibration process is relatively cumbersome. This technical solution uses the characteristic holes of the body to calibrate the camera's external parameters, without the need for other equipment, and the calibration process does not affect the production line; when the number of samples used for calibration is large enough, the calibration accuracy can meet the manufacturing process of the vehicle model. When adding cameras to be compatible with new models, the calibration process is more advantageous than the existing technology.

[0186] 2. Based on the above calibration results, the target vehicle positioning result calculated by this technical solution also conforms to the manufacturing process of the vehicle model. This result can be used to monitor the target vehicle's in-place deviation, and the positioning result can also be converted into the robot's operating compensation to correct the robot's operating trajectory.

[0187] In some embodiments, the present application also has some alternative solutions, such as:

[0188] 1. When calibrating the camera's intrinsic parameters, you can use calibration boards including but not limited to Chessboard, ChArUco, CCTag, etc.

[0189] 2. When performing 2D detection on characteristic holes, methods including but not limited to Hough transform, template matching, deep learning, etc. may be used.

[0190] 3. The three-dimensional information of the characteristic holes may include but is not limited to the actual position information of the characteristic holes, the hole spacing between the characteristic holes, the characteristic hole diameter and other information.

[0191] 4. The process features used are not limited to feature holes, rectangular shaped process features are also acceptable.

[0192] 5. When constructing a constrained optimization problem for solution, methods including but not limited to P3P and bundle adjustment can be used for solution.

[0193] 6. The number of cameras is usually 4, but more cameras can be added to be compatible with new models.

[0194] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0195] Based on the same inventive concept, the embodiment of the present application also provides a device positioning device for implementing the device positioning method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more device positioning device embodiments provided below can refer to the limitations on the device positioning method above, and will not be repeated here.

[0196] In an exemplary embodiment, Figure 5 As shown, a device positioning device is provided, comprising: a position acquisition module 100, a posture parameter detection module 200, a characteristic hole posture change detection module 300, a correction module 400 and a positioning module 500, wherein:

[0197] The position acquisition module 100 is used to acquire the current actual position information of the characteristic hole and the current characteristic hole image, and detect the current image position information of the characteristic hole from the current characteristic hole image, where the current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0198] The pose parameter detection module 200 is used to determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first pose transformation parameter between any two of all the observation coordinate systems, the second pose transformation parameter between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias pose transformation parameter between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias pose transformation parameter is the transformation parameter when the target vehicle is at the zero-bias position in the historical time period;

[0199] The characteristic hole posture change detection module 300 is used to generate the characteristic hole posture change between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period based on the historical zero bias posture transformation parameters and the second posture transformation parameters corresponding to the main shooting coordinate system;

[0200] A correction module 400, for correcting the characteristic hole posture change amount based on the current actual position information, all current image position information, the first posture transformation parameter, all second posture transformation parameters and the historical zero bias posture transformation parameter;

[0201] The positioning module 500 is used to position the target vehicle based on the corrected characteristic hole position change.

[0202] In one embodiment, the correction module 400 is also used to correct all second pose transformation parameters under the first target reprojection error based on the current actual position information, all current image position information and the first pose transformation parameter; and correct the characteristic hole pose change under the first target pose error based on the historical zero bias pose transformation parameter and all corrected second pose transformation parameters, wherein the first fusion error between the first target reprojection error and the first target pose error is less than the first preset error value.

[0203] In one embodiment, the correction module 400 is further used to obtain internal calibration parameters for any shooting device; based on the internal calibration parameters, the current actual position information, the current image position information and the first pose transformation parameter, the second pose transformation parameter under the first target reprojection error is corrected to obtain the corrected second pose transformation parameter, wherein the corrected second pose transformation parameter is the pose transformation parameter when the reprojection error of the first sub-target is less than the second preset error value;

[0204] The correction module 400 is also used to obtain a third posture transformation parameter from all corrected second posture transformation parameters, wherein the third posture transformation parameter is a posture transformation parameter between the main shooting coordinate system and the characteristic hole coordinate system; based on the historical zero bias posture transformation parameter and the third posture transformation parameter, the hole posture change under the first target posture error is corrected to obtain a corrected hole posture transformation, wherein the corrected hole posture transformation is the hole posture change when the first target posture error is less than the third preset error value.

[0205] In one of the embodiments, the posture parameter detection module 200 is also used to obtain the historical actual position information and historical characteristic hole images of the feature hole, and detect the historical image position information of the feature hole from the historical characteristic hole image, where the historical characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the historical time period; detect the historical posture transformation parameters between each observation coordinate system and the characteristic hole coordinate system within the historical time period, as well as the initial first position transformation parameters between any two of all observation coordinate systems; based on the historical actual position information, all historical image position information and all historical posture transformation parameters, correct all initial first position transformation parameters to obtain all first position transformation parameters.

[0206] In one embodiment, the posture parameter detection module 200 is also used to correct all historical posture transformation parameters under the second target reprojection error based on the historical actual position information and all historical image position information; based on all corrected historical posture transformation parameters, all initial first posture transformation parameters under the second target posture error are corrected, wherein the second fusion error between the second target reprojection error and the second target posture error is less than a fourth preset error value.

[0207] In one of the embodiments, the positioning module 400 is also used to determine the corrected characteristic hole posture change as the vehicle posture change of the target vehicle; obtain the standard vehicle posture information when the target vehicle is in the zero bias position, and perform posture positioning of the target vehicle based on the standard vehicle posture information and the vehicle posture change.

[0208] Each module in the above-mentioned device positioning apparatus can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module above.

[0209] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 6As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as the actual position information of the characteristic hole, the first characteristic hole image and the second characteristic hole image. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a device positioning method is implemented.

[0210] Those skilled in the art will understand that Figure 6 The structure shown in the figure is a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0211] In one embodiment, a device positioning system is also provided, the system includes a shooting device, a second delivery device and a controller, the shooting device and the main shooting device are both used to shoot the characteristic hole image on the target vehicle, and the controller is used to execute the following device positioning method:

[0212] Acquire the current actual position information of the characteristic hole and the current characteristic hole image, and detect the current image position information of the characteristic hole from the current characteristic hole image, where the current characteristic hole image is obtained by all shooting devices shooting the characteristic holes on the target vehicle within the current time period;

[0213] Determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first pose transformation parameters between any two of all the observation coordinate systems, the second pose transformation parameters between each observation coordinate system in the current time period and the characteristic hole coordinate system of the characteristic hole, and the historical zero-bias pose transformation parameters between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias pose transformation parameters are the transformation parameters when the target vehicle is at the zero-bias position in the historical time period;

[0214] Based on the historical zero bias pose transformation parameters and the second pose transformation parameters corresponding to the main shooting coordinate system, a characteristic hole pose change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period is generated;

[0215] Based on the current actual position information, all current image position information, the first pose transformation parameters, all second pose transformation parameters and the historical zero bias pose transformation parameters, the characteristic hole pose change amount is corrected;

[0216] Based on the corrected characteristic hole posture change, the target vehicle is positioned.

[0217] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0218] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0219] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0220] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0221] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0222] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A device positioning method, characterized in that: The method comprises: Acquire current actual position information and current characteristic hole image of the characteristic hole, and detect current image position information of the characteristic hole from the current characteristic hole image, wherein the current characteristic hole image is obtained by all shooting devices shooting the characteristic hole on the target vehicle within the current time period; Determine the main shooting coordinate system among the observation coordinate systems of all shooting devices, and detect the first pose transformation parameter between any two of all the observation coordinate systems, the second pose transformation parameter between each of the observation coordinate systems and the characteristic hole coordinate system of the characteristic hole in the current time period, and the historical zero bias pose transformation parameter between the main shooting coordinate system and the characteristic hole coordinate system, wherein the historical zero bias pose transformation parameter is the transformation parameter when the target vehicle is at the zero bias position in the historical time period, and the first pose transformation parameter is the initial first pose transformation parameter when the historical pose transformation parameter is optimal, wherein the initial first pose transformation parameter is the pose transformation parameter between any two of all the observation coordinate systems, and the historical pose transformation parameter is the pose transformation parameter between each observation coordinate system and the characteristic hole coordinate system in the historical time period; Based on the historical zero bias pose transformation parameter and the second pose transformation parameter corresponding to the main shooting coordinate system, generating a characteristic hole pose change amount between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period; Based on the current actual position information, all current image position information, the first posture transformation parameters, all second posture transformation parameters and the historical zero-bias posture transformation parameters, the characteristic hole posture change amount is corrected, wherein the correction of the characteristic hole posture change amount based on the current actual position information, all current image position information, the first posture transformation parameters, all second posture transformation parameters and the historical zero-bias posture transformation parameters includes: according to the current actual position information, all current image position information and the first posture transformation parameters, respectively correcting all second posture transformation parameters under the first target reprojection error; according to the historical zero-bias posture transformation parameters and all corrected second posture transformation parameters, correcting the characteristic hole posture change amount under the first target posture error, wherein the first fusion error between the first target reprojection error and the first target posture error is less than the first preset error value; Based on the corrected characteristic hole posture change, the target vehicle is positioned.

2. The method according to claim 1, characterized in that The method of correcting all second pose transformation parameters under the first target reprojection error according to the current actual position information, all current image position information and the first pose transformation parameter comprises: For any of the above-mentioned photographing devices, obtaining internal calibration parameters; Based on the internal calibration parameters, the current actual position information, the current image position information and the first posture transformation parameters, the second posture transformation parameters under the first target reprojection error are corrected to obtain corrected second posture transformation parameters, wherein the corrected second posture transformation parameters are posture transformation parameters when the first sub-target reprojection error is less than a second preset error value; The method of correcting the hole posture change under the first target posture error according to the historical zero bias posture transformation parameter and all the corrected second posture transformation parameters includes: Acquire a third posture transformation parameter from all corrected second posture transformation parameters, wherein the third posture transformation parameter is a posture transformation parameter between the main shooting coordinate system and the characteristic hole coordinate system; Based on the historical zero bias posture transformation parameters and the third posture transformation parameters, the hole posture change under the first target posture error is corrected to obtain a corrected hole posture change, wherein the corrected hole posture change is the hole posture change when the first target posture error is less than the third preset error value.

3. The method according to claim 1, characterized in that The detecting of the first pose transformation parameter between any two of all observation coordinate systems comprises: Acquire historical actual position information and historical characteristic hole images of the characteristic hole, and detect historical image position information of the characteristic hole from the historical characteristic hole images, wherein the historical characteristic hole images are obtained by all photographing devices photographing the characteristic holes on the target vehicle within a historical time period; Detecting historical pose transformation parameters between each observation coordinate system and the characteristic hole coordinate system in a historical time period, and initial first pose transformation parameters between any two of all observation coordinate systems; Based on the historical actual position information, all historical image position information and all historical posture transformation parameters, all the initial first posture transformation parameters are corrected to obtain all the first posture transformation parameters.

4. The method according to claim 3, characterized in that The method of correcting all the initial first pose transformation parameters based on the historical actual position information, all the historical image position information and all the historical pose transformation parameters includes: Based on the historical actual position information and all historical image position information, all historical posture transformation parameters under the second target reprojection error are corrected; Based on all corrected historical pose transformation parameters, all initial first pose transformation parameters under the second target pose error are corrected, wherein a second fusion error between the second target reprojection error and the second target pose error is less than a fourth preset error value.

5. The method according to claim 1, characterized in that The positioning of the target vehicle based on the corrected characteristic hole position change comprises: Determine the corrected characteristic hole position change as the vehicle position change of the target vehicle; The standard vehicle posture information when the target vehicle is in the zero bias position is obtained, and the posture positioning of the target vehicle is performed based on the standard vehicle posture information and the vehicle posture change amount.

6. A device positioning device, characterized in that: The device comprises: A position acquisition module, used to acquire current actual position information and current characteristic hole image of the characteristic hole, and detect current image position information of the characteristic hole from the current characteristic hole image, wherein the current characteristic hole image is obtained by all shooting devices shooting the characteristic hole on the target vehicle within the current time period; A posture parameter detection module is used to determine the main shooting coordinate system in the observation coordinate systems of all shooting devices, and detect the first posture transformation parameter between any two of all the observation coordinate systems, the second posture transformation parameter between each of the observation coordinate systems and the characteristic hole coordinate system of the characteristic hole in the current time period, and the historical zero-bias posture transformation parameter between the main shooting coordinate system and the characteristic hole coordinate system, the historical zero-bias posture transformation parameter is the transformation parameter when the target vehicle is in the zero-bias position in the historical time period, the first posture transformation parameter is the initial first posture transformation parameter when the historical posture transformation parameter is optimal, wherein the initial first posture transformation parameter is the posture transformation parameter between any two of all the observation coordinate systems, and the historical posture transformation parameter is the posture transformation parameter between each observation coordinate system and the characteristic hole coordinate system in the historical time period; A characteristic hole posture change detection module is used to generate a characteristic hole posture change between the characteristic hole coordinate system and the main shooting coordinate system between the historical time period and the current time period based on the historical zero bias posture transformation parameter and the second posture transformation parameter corresponding to the main shooting coordinate system; A correction module, for correcting the characteristic hole posture change amount based on the current actual position information, all current image position information, the first posture transformation parameter, all second posture transformation parameters and the historical zero bias posture transformation parameter, wherein the correction of the characteristic hole posture change amount based on the current actual position information, all current image position information, the first posture transformation parameter, all second posture transformation parameters and the historical zero bias posture transformation parameter comprises: respectively correcting all second posture transformation parameters under the first target reprojection error according to the current actual position information, all current image position information and the first posture transformation parameter; correcting the characteristic hole posture change amount under the first target posture error according to the historical zero bias posture transformation parameter and all the corrected second posture transformation parameters, wherein the first fusion error between the first target reprojection error and the first target posture error is less than the first preset error value; The positioning module is used to position the target vehicle based on the corrected characteristic hole position change.

7. The device according to claim 6, characterized in that The positioning module is also used to determine the corrected characteristic hole posture change as the vehicle posture change of the target vehicle; obtain standard vehicle posture information when the target vehicle is in a zero bias position, and perform posture positioning on the target vehicle based on the standard vehicle posture information and the vehicle posture change.

8. A device positioning system, characterized in that: The system includes a plurality of photographing devices and a controller, all of the photographing devices are used to photograph characteristic hole images on a target vehicle, and the controller is used to execute the device positioning method according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images