A method for constructing a reference of a multi-motion platform based on structured light

Abstract

The application provides a multi-moving-platform active reference self-construction method based on structured light, a projector and a camera are mounted on a reference moving platform, the projector actively projects structured light to a target, and the spatial coordinates of the light spots in a reference coordinate system are obtained through the structured light method. Cameras are mounted on multiple observation moving platforms to obtain structured light image data, and the pose information of the multiple observation moving platforms in the reference coordinate system is obtained through PnP pose estimation, so that the multi-moving-platform active reference self-construction is realized. The method provided by the application compensates for the defects of the traditional feature-based pose estimation method in weak texture scenes, such as accuracy decline or failure, and can be applied to large structure topography reconstruction and measurement in a weak texture information scene, and is a multi-moving-platform reference self-construction method with flexibility and applicability.

Description

Technical Field

[0001] This invention mainly relates to the field of dynamic platform reference self-construction technology, and in particular to a multi-motion platform active reference self-construction method based on structured light. Background Technology

[0002] In outdoor inspection and surveying, mobile platforms such as drones and unmanned vehicles are needed to collect image information of targets within a designated area, and then obtain observation results through data processing. To ensure observation accuracy, the mobile platform needs to approach the target to collect local information, and then obtain the overall information of the target through data fusion. Since the mobile platform itself is in motion, its spatial pose is constantly changing, and data fusion has a reference alignment problem, requiring the construction of a unified observation reference. On the other hand, for mobile platforms used in indoor operations such as industrial robots and robotic arms, they often need to obtain their own position and attitude information and establish a relative relationship with the scene or target when performing tasks, and perform motion planning under a unified reference.

[0003] A unified reference system is fundamental for information fusion and task planning on most moving platforms. Essentially, it involves pose estimation of the moving platform. First, a reference coordinate system is constructed within the scene. Then, pose estimation methods are used to obtain the pose information of the moving platform relative to the reference system at each moment. Traditional positioning technologies, such as GPS, are limited by wireless signals and difficult to use indoors. Furthermore, GPS only outputs the position information of the moving platform, lacking attitude information. Epipolar geometric constraint pose estimation methods based on monocular cameras lack depth information, resulting in pose estimations that lack scale information. Pose estimation methods based on binocular cameras or depth cameras suffer from poor accuracy and stability due to the influence of observation distance.

[0004] Visual odometry, which combines visual and inertial navigation information, integrates a camera and an inertial sensor to achieve stable pose estimation, but its accuracy is low and it is not suitable for scenarios with high accuracy requirements.

[0005] In summary, benchmark unification is a common requirement faced by most dynamic platforms, but existing technologies struggle to construct stable and accurate unified benchmarks. Therefore, it is essential to research a dynamic platform benchmark self-construction method that is widely applicable, highly stable, and provides high solution accuracy. Summary of the Invention

[0006] To address the technical problems existing in the prior art, this invention proposes a multi-motion platform active benchmark self-construction method based on structured light. This method actively projects light information onto weakly textured scenes or targets and realizes the self-construction of multiple motion platform benchmarks based on structured light visual calculation.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] On one hand, this invention provides a method for active self-construction of a multi-motion platform reference based on structured light, comprising:

[0009] A projector on a reference moving platform projects structured light onto the object's surface, and cameras on multiple moving platforms, including the reference moving platform and the observation moving platform, capture structured light images of the object's surface.

[0010] Structured light feature points are extracted from the structured light images of the object surface acquired by each moving platform, and the feature descriptor corresponding to each structured light feature point is calculated.

[0011] Based on the two-dimensional image coordinates of the structured light feature points extracted from the structured light feature points in the structured light image on the object surface, and combined with the calibrated relative pose of the projector and the camera, the three-dimensional spatial coordinates of the structured light feature points are obtained through structured light triangulation.

[0012] The pose of the observation platform is estimated by using the three-dimensional spatial coordinates of the structured light feature points and the feature descriptor corresponding to each structured light feature point, so as to obtain the relative pose of the observation platform and the reference platform and realize the spatial alignment of the observation platform to the reference platform.

[0013] A scene point cloud is formed based on the fixed and invariant feature points in the scene obtained from the reference motion platform at the initial time. The pose of the reference motion platform is estimated using the scene point cloud, thereby obtaining the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time, and realizing the time alignment of the reference motion platform.

[0014] Furthermore, the observation platform described in this invention comprises one or more.

[0015] Furthermore, the present invention controls the acquisition time of cameras on all moving platforms through a unified signal. Initially, all cameras start data acquisition synchronously and at the same frame rate. When the end signal is received, all cameras stop data acquisition.

[0016] Furthermore, this invention employs an accelerated robust feature extraction algorithm to extract feature points from the structured light image of the object surface, obtaining the two-dimensional image coordinates of the structured light feature points in the structured light image of the object surface. Further, the feature descriptor corresponding to each structured light feature point is calculated, including:

[0017] An integral image is created, and a second-order Gaussian differential convolution is performed on the input structured light image of the object surface to calculate the Hessian matrix.

[0018] The integral image is convolved in different directions using filters of different sizes to construct a scale space. Local maxima are extracted using the Hessian matrix discriminant and used as structured light feature points to obtain the two-dimensional image coordinates of the structured light feature points in the structured light image on the object surface.

[0019] The corresponding feature descriptor is obtained by calculating the principal direction of the neighborhood window of each structured light feature point using the Harr wavelet response.

[0020] Furthermore, the methods for obtaining feature descriptors include:

[0021] Create a 60° sector window and perform a rotational scan on the circular neighborhood centered on the structured light feature point. Statistically analyze the Harr wavelet response of each scanned region and take the main direction corresponding to the sector region with the largest statistical value as the main direction of the feature point.

[0022] Centered on the structured light feature point, the coordinate axis is rotated to the main direction of the feature point, and the neighborhood of the structured light feature point is divided into 16 positive direction sub-regions;

[0023] Create a 60° sector window and perform a rotational scan on the circular neighborhood centered on the structured light feature point. Statistically analyze the Harr wavelet response results of each positive direction sub-region and normalize them to obtain the feature descriptor corresponding to the structured light feature point.

[0024] Furthermore, the present invention provides a preferred method for achieving spatial alignment of the observation moving platform to the reference moving platform, comprising: using feature descriptors to match the structured light feature points on the object surface structured light image acquired by the camera on the observation moving platform at any given moment with the structured light feature points on the object surface structured light image acquired by the camera on the reference moving platform, to obtain the two-dimensional image coordinates of the structured light feature points under the observation moving platform; combining structured light triangulation to obtain the three-dimensional spatial coordinates of the structured light feature points, and using PnP pose estimation to calculate the pose information of the observation moving platform in the reference coordinate system at that moment, thereby achieving spatial alignment of the observation moving platform to the reference moving platform.

[0025] Furthermore, the present invention provides a preferred method for achieving time alignment of a reference motion platform, comprising: at an initial moment, initializing the image data obtained by the reference motion platform, obtaining fixed and invariant feature points in the scene through feature extraction and structured light triangulation, forming a fixed and invariant scene point cloud, and obtaining the spatial three-dimensional coordinates of each fixed and invariant feature point in the scene point cloud;

[0026] For image data obtained by the reference motion platform at any time after the initial time, the two-dimensional coordinates of each fixed feature point are obtained by feature matching. Then, the transformation relationship between the reference motion platform and the initial pose is calculated by PnP pose estimation. In this way, the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time is calculated, and the time alignment of the reference motion platform is achieved.

[0027] On the other hand, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0028] A projector on a reference moving platform projects structured light onto the object's surface, and cameras on multiple moving platforms, including the reference moving platform and the observation moving platform, capture structured light images of the object's surface.

[0029] Structured light feature points are extracted from the structured light images of the object surface acquired by each moving platform, and the feature descriptor corresponding to each structured light feature point is calculated.

[0030] Based on the two-dimensional image coordinates of the structured light feature points extracted from the structured light feature points in the structured light image on the object surface, and combined with the calibrated relative pose of the projector and the camera, the three-dimensional spatial coordinates of the structured light feature points are obtained through structured light triangulation.

[0031] The pose of the observation platform is estimated by using the three-dimensional spatial coordinates of the structured light feature points and the feature descriptor corresponding to each structured light feature point, so as to obtain the relative pose of the observation platform and the reference platform and realize the spatial alignment of the observation platform to the reference platform.

[0032] A scene point cloud is formed based on the fixed and invariant feature points in the scene obtained from the reference motion platform at the initial time. The pose of the reference motion platform is estimated using the scene point cloud, thereby obtaining the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time, and realizing the time alignment of the reference motion platform.

[0033] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, performs the following steps:

[0034] A projector on a reference moving platform projects structured light onto the object's surface, and cameras on multiple moving platforms, including the reference moving platform and the observation moving platform, capture structured light images of the object's surface.

[0035] Structured light feature points are extracted from the structured light images of the object surface acquired by each moving platform, and the feature descriptor corresponding to each structured light feature point is calculated.

[0036] Based on the two-dimensional image coordinates of the structured light feature points extracted from the structured light feature points in the structured light image on the object surface, and combined with the calibrated relative pose of the projector and the camera, the three-dimensional spatial coordinates of the structured light feature points are obtained through structured light triangulation.

[0037] The pose of the observation platform is estimated by using the three-dimensional spatial coordinates of the structured light feature points and the feature descriptor corresponding to each structured light feature point, so as to obtain the relative pose of the observation platform and the reference platform and realize the spatial alignment of the observation platform to the reference platform.

[0038] A scene point cloud is formed based on the fixed and invariant feature points in the scene obtained from the reference motion platform at the initial time. The pose of the reference motion platform is estimated using the scene point cloud, thereby obtaining the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time, and realizing the time alignment of the reference motion platform.

[0039] Compared with the prior art, the technical effects of the present invention are as follows:

[0040] This invention requires the collaborative implementation of multiple moving platforms and is a method for self-constructing a benchmark across multiple moving platforms based on structured light visual computation, actively projecting light information onto weakly textured scenes or targets. Specifically, a projector and camera are mounted on a benchmark moving platform. Structured light is actively projected onto the target using the projector, and the spatial coordinates of the light spot in the benchmark coordinate system are obtained using structured light methods. Cameras are mounted on multiple observation moving platforms to acquire structured light image data, and then the pose information in the benchmark coordinate system is obtained through PnP pose estimation, realizing active self-construction of the benchmark across multiple moving platforms. The method proposed in this invention overcomes the shortcomings of traditional feature-based pose estimation methods in terms of accuracy degradation or failure in weakly textured scenes by actively projecting structured light. It can be applied to the reconstruction and measurement of large structural shapes with weak texture information, and is a flexible and applicable method for self-constructing a benchmark across multiple moving platforms. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0042] Figure 1 This is a flowchart of an embodiment of the present invention;

[0043] Figure 2 Schematic diagram of structured light projection and data acquisition;

[0044] Figure 3 Schematic diagram of structured light feature triangulation measurement principle. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0046] Reference Figure 1 One embodiment provides a method for active benchmark self-construction of a multi-motion platform based on structured light, including:

[0047] (S1) Structured light projection and data acquisition;

[0048] A projector on a reference moving platform projects structured light onto the object's surface, and cameras on multiple moving platforms, including the reference moving platform and the observation moving platform, capture structured light images of the object's surface.

[0049] (S2) Structured light feature extraction;

[0050] Structured light feature points are extracted from the structured light images of the object surface acquired by each moving platform, and the feature descriptor corresponding to each structured light feature point is calculated.

[0051] (S3) Structured light feature triangulation measurement;

[0052] Based on the two-dimensional image coordinates of the structured light feature points extracted from the structured light feature points in the structured light image on the object surface, and combined with the calibrated relative pose of the projector and the camera, the three-dimensional spatial coordinates of the structured light feature points are obtained through structured light triangulation.

[0053] (S4) Spatial alignment of the observation platform;

[0054] The pose of the observation platform is estimated by using the three-dimensional spatial coordinates of the structured light feature points and the feature descriptor corresponding to each structured light feature point, so as to obtain the relative pose of the observation platform and the reference platform and realize the spatial alignment of the observation platform to the reference platform.

[0055] (S5) Reference dynamic platform time alignment;

[0056] A scene point cloud is formed based on the fixed and invariant feature points in the scene obtained from the reference motion platform at the initial time. The pose of the reference motion platform is estimated using the scene point cloud, thereby obtaining the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time, and realizing the time alignment of the reference motion platform.

[0057] In one embodiment, such as Figure 2As shown, in a weakly textured scene, a reference moving platform carrying a projector and camera hovers in the air, actively projecting structured light onto the object's surface. Simultaneously, multiple observation moving platforms, each carrying a camera, observe the structured light on the object's surface. The acquisition time of all cameras on all moving platforms is uniformly controlled by a signal. Upon receiving the start signal, all cameras synchronize and begin data acquisition at the same frame rate initially; upon receiving the end signal, all cameras stop data acquisition. By setting the same image acquisition frame rate and synchronous data acquisition control, for the temporal images obtained by all moving platforms, it can be ensured that the spatial coordinates corresponding to the structured light feature points extracted from images with the same frame number are consistent.

[0058] This invention does not limit the method of feature extraction, and those skilled in the art can choose from existing feature extraction methods according to the circumstances and needs.

[0059] In one embodiment, a Speed-Up Robust Feature Extraction (SURF) algorithm is proposed to extract feature points from structured light images of an object surface, obtaining the two-dimensional image coordinates of the structured light feature points in the object surface structured light image. The SURF algorithm extracts feature points with rotation and scale invariance by establishing a scale space. Specifically, it includes:

[0060] An integral image is created, and a second-order Gaussian differential convolution is performed on the input structured light image of the object surface to calculate the Hessian matrix.

[0061]

[0062] In the formula, H represents the Hessian matrix, σ is the image scale, and L... xx (x,y,σ),L xy (x,y,σ),L yy (x,y,σ) is the second derivative of the Gaussian filter.

[0063] Based on the integral image, convolution calculations are performed on the integral image in different directions using filters of different sizes to construct a scale space. Local maxima are extracted using the Hessian matrix discriminant, and these are used as structured light feature points to obtain the two-dimensional image coordinates of the structured light feature points in the structured light image of the object surface. The Hessian matrix discriminant is as follows:

[0064] Det(H) = D xx D yy -(ωD xy ) 2 >K

[0065] In the formula, D xx D xy D yThese are the results of convolving the image with filters of different sizes, where ω is the compensation coefficient and K is the given threshold.

[0066] The corresponding feature descriptors are obtained by calculating the principal direction of the neighborhood window of each structured light feature point using the HAR wavelet response, including:

[0067] Create a 60° sector window and perform a rotational scan on the circular neighborhood centered on the structured light feature point. Statistically analyze the Harr wavelet response of each scanned region and take the main direction corresponding to the sector region with the largest statistical value as the main direction of the feature point.

[0068] Centered on the structured light feature point, the coordinate axis is rotated to the main direction of the feature point, and the neighborhood of the structured light feature point is divided into 16 positive direction sub-regions;

[0069] Create a 60° sector window and perform a rotational scan on the circular neighborhood centered on the structured light feature point. Statistically analyze the Harr wavelet response results of each positive direction sub-region and normalize them to obtain the feature descriptor corresponding to the structured light feature point.

[0070] The triangulation principle of the structured light feature points described in this invention is as follows: Figure 3 In three-dimensional space, the pinhole imaging model yields the following:

[0071]

[0072] According to the principles of triangulation:

[0073]

[0074] The two equations combined yield:

[0075]

[0076]

[0077] The final three-dimensional coordinates of the spatial feature point P are:

[0078]

[0079]

[0080]

[0081] Where (x, y) are the two-dimensional image coordinates of the structured light feature points, obtained through feature point extraction; the focal length f is obtained through camera calibration; and α and b are obtained by calibrating the relative relationship between the camera and the projector. Through structured light feature triangulation, the spatial three-dimensional coordinates of the structured light feature points in the reference coordinate system can be obtained.

[0082] One embodiment provides a preferred method for spatial alignment of an observation moving platform with a reference moving platform, comprising: matching structured light feature points on the object surface structured light image captured by a camera on the observation moving platform at any given moment with structured light feature points on the object surface structured light image captured by a camera on the reference moving platform using feature descriptors, to obtain the two-dimensional image coordinates of the structured light feature points on the observation moving platform; combining structured light triangulation to obtain the three-dimensional spatial coordinates of the structured light feature points, and calculating the pose information of the observation moving platform in the reference coordinate system at that moment through PnP pose estimation, thereby achieving spatial alignment of the observation moving platform with the reference moving platform. The reference coordinate system is a coordinate system constructed based on the reference moving platform. Specifically, those skilled in the art can use coordinate system construction methods known or commonly used in the field to establish the reference system. Without loss of generality, the reference moving platform can be constructed using structured light methods.

[0083] One embodiment provides a preferred method for time alignment of a reference moving platform, comprising:

[0084] At the initial moment, the image data obtained by the reference motion platform is initialized. Through feature extraction and structured light triangulation, fixed feature points in the scene are obtained to form a fixed scene point cloud, and the spatial three-dimensional coordinates of each fixed feature point in the scene point cloud are obtained.

[0085] For image data obtained by the reference motion platform at any time after the initial time, the two-dimensional coordinates of each fixed feature point are obtained by feature matching. Then, the transformation relationship between the reference motion platform and the initial pose is calculated by PnP pose estimation. In this way, the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time is calculated, and the time alignment of the reference motion platform is achieved.

[0086] This completes the unification of the spatiotemporal reference for the multi-motion platform and enables the self-construction of the multi-motion platform reference.

[0087] On the other hand, the present invention provides a computer device including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the active benchmark self-construction method for a multi-motion platform based on structured light provided in any of the above embodiments. The computer device can be a server. The computer device includes a processor, a memory, a network interface, and a database connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and 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 computer program in the non-volatile storage medium. The database of the computer device is used to store sample data. The network interface of the computer device is used for communication with external terminals via a network connection.

[0088] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of the active reference self-construction method for a multi-motion platform based on structured light provided in any of the above embodiments.

[0089] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0090] Matters not covered in this invention are common knowledge.

[0091] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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 specification.

[0092] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

[0093] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for active self-construction of a reference for a multi-motion platform based on structured light, characterized in that, include: A projector on a reference moving platform projects structured light onto the object's surface, and cameras on multiple moving platforms, including the reference moving platform and the observation moving platform, capture structured light images of the object's surface. Structured light feature points are extracted from the structured light images of the object surface acquired by each moving platform, and the feature descriptor corresponding to each structured light feature point is calculated. Based on the two-dimensional image coordinates of the structured light feature points extracted from the structured light feature points in the structured light image on the object surface, and combined with the calibrated relative pose of the projector and the camera, the three-dimensional spatial coordinates of the structured light feature points are obtained through structured light triangulation. The pose of the observation platform is estimated by using the three-dimensional spatial coordinates of the structured light feature points and the feature descriptor corresponding to each structured light feature point, so as to obtain the relative pose of the observation platform and the reference platform and realize the spatial alignment of the observation platform to the reference platform. A scene point cloud is formed based on the fixed and invariant feature points in the scene obtained by the reference motion platform at the initial time. The pose of the reference motion platform is estimated using the scene point cloud, thereby obtaining the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time, and realizing the time alignment of the reference motion platform. Obtain feature descriptors, including: Create a 60° sector window and perform a rotational scan on the circular neighborhood centered on the structured light feature point. Statistically analyze the Harr wavelet response of each scanned region and take the main direction corresponding to the sector region with the largest statistical value as the main direction of the feature point. Centered on the structured light feature point, the coordinate axis is rotated to the main direction of the feature point, and the neighborhood of the structured light feature point is divided into 16 positive direction sub-regions; Create a 60° sector window and perform a rotational scan on the circular neighborhood centered on the structured light feature point. Statistically analyze the Harr wavelet response results of each positive direction sub-region and normalize them to obtain the feature descriptor corresponding to the structured light feature point.

2. The active reference self-construction method for a multi-motion platform based on structured light according to claim 1, characterized in that, There is one or more observation platforms.

3. The active reference self-construction method for a multi-motion platform based on structured light according to claim 2, characterized in that, The timing of data acquisition by cameras on all moving platforms is controlled uniformly by a signal. Initially, all cameras start data acquisition synchronously at the same frame rate. When the end signal is received, all cameras stop data acquisition.

4. The active reference self-construction method for a multi-motion platform based on structured light according to claim 3, characterized in that, An accelerated robust feature extraction algorithm is used to extract feature points from the structured light image of the object surface, and the two-dimensional image coordinates of the structured light feature points in the structured light image of the object surface are obtained.

5. The active reference self-construction method for a multi-motion platform based on structured light according to claim 4, characterized in that, Calculate the feature descriptor corresponding to each structured light feature point, including: An integral image is created, and a second-order Gaussian differential convolution is performed on the input structured light image of the object surface to calculate the Hessian matrix. The integral image is convolved in different directions using filters of different sizes to construct a scale space. Local maxima are extracted using the Hessian matrix discriminant and used as structured light feature points to obtain the two-dimensional image coordinates of the structured light feature points in the structured light image on the object surface. The corresponding feature descriptor is obtained by calculating the principal direction of the neighborhood window of each structured light feature point using the Harr wavelet response.

6. A method for active self-construction of a multi-motion platform reference based on structured light according to any one of claims 4 to 5, characterized in that, The method for achieving spatial alignment between the observation moving platform and the reference moving platform includes: using feature descriptors to match the structured light feature points on the object surface structured light image acquired by the camera on the observation moving platform at any given moment with the structured light feature points on the object surface structured light image acquired by the camera on the reference moving platform, thereby obtaining the two-dimensional image coordinates of the structured light feature points under the observation moving platform; combining structured light triangulation to obtain the three-dimensional spatial coordinates of the structured light feature points; and using PnP pose estimation to calculate the pose information of the observation moving platform in the reference coordinate system at that moment, thereby achieving spatial alignment between the observation moving platform and the reference moving platform.

7. A method for active self-construction of a multi-motion platform reference based on structured light according to any one of claims 4 to 5, characterized in that, The method for achieving time alignment of the reference motion platform includes: at the initial moment, initializing the image data obtained by the reference motion platform, obtaining fixed feature points in the scene through feature extraction and structured light triangulation, forming a fixed scene point cloud, and obtaining the spatial three-dimensional coordinates of each fixed feature point in the scene point cloud. For image data obtained by the reference motion platform at any time after the initial time, the two-dimensional coordinates of each fixed feature point are obtained by feature matching. Then, the transformation relationship between the reference motion platform and the initial pose is calculated by PnP pose estimation. In this way, the relative pose relationship between the reference motion platform at any time and the reference motion platform at the initial time is calculated, and the time alignment of the reference motion platform is achieved.

8. A computer device, characterized in that: It includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the active reference self-construction method for a multi-motion platform based on structured light as described in claim 1.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the active benchmark self-construction method for a multi-motion platform based on structured light as described in claim 1.

Images