A joint mapping method based on coal pile height field matching and space-time global optimization

Abstract

The application discloses a kind of based on coal pile height field matching and space-time global optimization combined mapping method, it is related to the positioning technical field of ship unloader, to solve the problem that point cloud obtained when ship unloader scans in high altitude is relatively sparse, and the data of ship unloader is difficult to directly match;Method includes: S1, using the multiscale height field correlation matching algorithm of ship unloader and ship unloader coal pile point cloud, using coal pile as common feature body, by matching its overall height distribution structure to find common area by gridding;S2, point cloud registration is carried out to common area, and the initial transformation relationship of ship unloader and ship unloader is determined;S3, construct global constraint factor graph optimization model, input the key frame point cloud and map data of ship unloader and the point cloud data of ship unloader, output the combined map of ship unloader and ship unloader.The method of the application can significantly improve the robustness of initial registration, and realize cross-time and space error compensation, improve map global consistency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of positioning of a ship unloader, and particularly relates to a joint mapping method based on coal pile height field matching and spatio-temporal global optimization. BACKGROUND

[0002] In the process of port bulk coal transfer, first, the ship unloader gathers the bulk coal to the center of the ship cabin, and then the ship unloader grabs the bulk coal and places it on the belt conveyor for transportation. At present, the ship unloading operation mainly relies on manual operation, and there are problems of low work efficiency and poor working environment. The automation of ship unloading is an important means to improve work efficiency and ensure personnel safety. Among them, the joint mapping of the ship unloader and the ship unloader is the basis for realizing the automatic joint operation. However, the existing equipment has defects: the laser radar of the ship unloader is easily blocked by the coal pile, resulting in a limited field of view in the cabin, and error accumulation is prone to occur during pose estimation; the point cloud obtained by the ship unloader when scanning at a high altitude is relatively sparse, and the data of the ship unloader cannot be directly matched.

[0003] For example, the Chinese patent with publication number CN118363032A relates to a pusher machine positioning method based on ship unloader laser cooperative positioning, which realizes cooperative positioning through point cloud downsampling and index marking, and the specific technical solution is different from the present application. SUMMARY

[0004] The present application solves the problem that the point cloud obtained by the ship unloader when scanning at a high altitude is relatively sparse, and the data of the ship unloader cannot be directly matched. A joint mapping method based on coal pile height field matching and spatio-temporal global optimization is proposed, which significantly improves the initial registration robustness.

[0005] In order to achieve the above purpose, the present application adopts the following technical solution: a joint mapping method based on coal pile height field matching and spatio-temporal global optimization, comprising the following steps:

[0006] S1, using a multi-scale height field correlation matching algorithm of the ship unloader and the ship unloader coal pile point cloud, using the coal pile as a common feature body, and matching the overall height distribution structure of the coal pile through gridding to find the common area;

[0007] S2, performing point cloud registration on the common area to determine the initial transformation relationship between the ship unloader and the ship unloader;

[0008] S3, constructing a global constraint factor graph optimization model, inputting the key frame point cloud and map data of the ship unloader and the point cloud data of the ship unloader, and outputting the joint map of the ship unloader and the ship unloader.

[0009] In the technical solution, firstly, an initial transformation relationship between the ship unloader and the ship unloader is constructed based on the coal pile height field matching, then pose optimization is performed based on the obtained initial transformation relationship, and after the optimized relationship is obtained, the whole joint mapping method based on global constraint factor graph optimization can be obtained.

[0010] The application is further provided with that the step S1 comprises:

[0011] S11, correcting the point cloud data of the ship unloader, and performing branch and bound on the coal pile data of the ship unloader according to the size of the point cloud data of the ship unloader;

[0012] S12, performing gridding processing on the coal pile data of the ship unloader and the ship unloader respectively to construct a coal pile height field, including a high-resolution height grid map Hc and a low-resolution height grid map Hu;

[0013] S13, calculating the similarity of Hc and Hu, and determining the common area of the coal pile according to the similarity.

[0014] The multi-scale height field correlation matching algorithm is further limited to finally obtain the common area of the coal pile.

[0015] The application is further provided with that the step S3 comprises:

[0016] S31, superimposing the key frame pose data and the point cloud data of the ship unloader to generate map data of the ship unloader;

[0017] S32, calculating the common area of the map data of the ship unloader and the point cloud data of the ship unloader according to step S1;

[0018] S33, registering the common area to obtain the relative observation data between the ship unloader and the ship unloader;

[0019] S34, generating an optimization function in combination with the historical relative observation data, and obtaining the optimized relative transformation relationship according to the optimization function.

[0020] Due to the sparsity of the point cloud of the ship unloader and the ship unloader, the initial transformation relationship data is not accurate enough, which negatively affects the mapping accuracy. To solve this problem, the transformation relationship is optimized through the above steps.

[0021] The application is further provided with that the point cloud data of the ship unloader is corrected, including:

[0022] The input point cloud of the ship unloader is subjected to PCA analysis, the main direction of the ship unloader is perpendicular to the main direction of the ship unloader, and the point cloud of the ship unloader is rotated.

[0023] In the technical solution, the point cloud data of the ship unloader is corrected to lay a foundation for subsequent processing.

[0024] The application is further provided that the constructed coal pile height field specifically comprises:

[0025] The point cloud data of the stripping machine is subjected to PCA alignment, and then a low-resolution height grid map Hu is constructed.

[0026] The point cloud data of the stripping machine is subjected to PCA alignment, and then a low-resolution height grid map Hu is constructed.

[0027] The application is further provided that the similarity of Hc and Hu is specifically calculated by using a normalized cross-correlation algorithm.

[0028] The application is further provided that after the similarity of Hc and Hu is calculated, the grid resolution and the search position are iteratively optimized until the similarity is maximized, and at this time, the common area of the stripping machine and the unloading machine is found out according to the maximized similarity.

[0029] The corresponding common area is found by maximizing the similarity.

[0030] The application is further provided that the left side of the optimization function equation is the optimized relative observation, and the right side of the optimization function equation is the minimum value of the cumulative sum of the product of the transposition of the first parameter and the first parameter.

[0031] The relative observation data is optimized by the optimization function, and the optimized relative transformation relationship is obtained by the optimized relative observation data.

[0032] The application is further provided that the first parameter specifically includes the reciprocal of the state of the stripping machine at the i-th moment, the reciprocal of the relative observation data, and the continuous product of the state of the unloading machine at the j-th moment.

[0033] The application is further provided that the map data of the stripping machine is the cumulative sum of the product of the pose data of the i-th key frame of the stripping machine and the point cloud data of the i-th key frame of the stripping machine.

[0034] The pose data and the point cloud data of the stripping machine of the key frame are read, and the data of multiple key frames are superimposed.

[0035] The application can bring the following beneficial effects:

[0036] 1. The joint mapping method based on coal pile height field matching and spatio-temporal global optimization provided in the application can effectively overcome the difficulty of direct matching of sparse-dense point clouds, provide reliable initial values for subsequent optimization, and significantly improve the robustness of initial registration; 2. The spatio-temporal global constraint factor graph optimization mapping method is used to realize cross-spatio-temporal error compensation and improve the global consistency of the map;

[0037] 3、The method of the application is significantly better than the traditional scheme in terms of unstructured scene registration success rate, long-term consistency and calculation efficiency, and can meet the high-precision mapping needs of the whole operation cycle. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is a schematic diagram of a joint mapping framework of the coal pile height field matching and spatio-temporal global optimization based joint mapping method of the application.

[0039] Figure 2 is a schematic diagram of the coordinate conversion relationship of the coal pile height field matching and spatio-temporal global optimization based joint mapping method of the application.

[0040] Figure 3 is a schematic diagram of the multi-scale height field correlation matching algorithm of the coal pile height field matching and spatio-temporal global optimization based joint mapping method of the application.

[0041] Figure 4 is a schematic diagram of the spatio-temporal global constraint factor based graph optimization framework of the coal pile height field matching and spatio-temporal global optimization based joint mapping method of the application. DETAILED DESCRIPTION

[0042] Embodiment 1

[0043] The embodiment proposes a joint mapping method based on coal pile height field matching and spatio-temporal global optimization, referring to Figure 1 , Figure 2 , Figure 3 and Figure 4 , which includes the following steps.

[0044] Step S1, first, determine the common area according to the multi-scale height field correlation matching algorithm of the coal pile point cloud of the ship unloader and the ship unloader, specifically take the coal pile as the common feature body, and match the overall height distribution structure by gridding to find the corresponding common area.

[0045] In the ship unloading operation, the ship unloader needs to pass through the common area of the coal pile to determine the conversion relationship; Figure 2 the coordinate relationship of the ship unloader and the ship unloader, the ship unloader is in the ship cabin, and the ship unloader is above the ship unloader. Since most of the coal pile area is in the ship cabin, the data captured by the radar of the ship unloader is less than that of the ship unloader. If the iterative nearest point method is directly used, it is difficult to find an effective conversion relationship, so the common area of the ship unloader and the ship unloader coal pile needs to be detected to ensure that the matching relationship is correctly found. In Figure 2 , the prior art uses the conversion matrix A coordinate transformation is performed, which is usually composed of a 3*3 rotation matrix R and a 3*1 translation vector t; and the present application adopts multi-scale height field matching and point cloud registration to obtain the initial transformation relationship T between the ship unloader and the ship unloader init , that is, the optimal transformation initial value T init .

[0046] As Figure 3 The multi-scale height field correlation matching algorithm of the ship unloader and the ship unloader coal pile point cloud is shown, which uses coal piles as unstructured but common feature bodies, and matches the overall height distribution structure (not local feature points) of the grid to find the common area.

[0047] Referring to Figure 1 , Figure 2 and Figure 3 , for step S1, it mainly includes the following several sub-steps.

[0048] Step S11, the obtained ship unloader point cloud data is corrected, and then the ship unloader point cloud size is used to branch and bound the ship unloader coal pile data.

[0049] For the above process, more specifically, the input ship unloader point cloud is subjected to PCA analysis, PCA (principal components analysis) is a principal component analysis technique, which can convert multiple indicators into a few comprehensive indicators by dimension reduction. Specifically, the main direction of the ship unloader is perpendicular to the main direction of the ship unloader, that is, the ship unloader point cloud is rotated, so as to lay a foundation for subsequent data processing.

[0050] Then, according to the size of the ship unloader point cloud, the ship unloader coal pile data is subjected to branch and bound, further clarifying the data range, and then extracting appropriate coal pile data from both sides of the ship unloader and the ship unloader point cloud data.

[0051] Step S12, after completing the above step S11, the ship unloader and the ship unloader coal pile data are subjected to grid processing to construct a coal pile height field, wherein the constructed coal pile height field mainly includes a high-resolution height grid Hc and a low-resolution height grid Hu.

[0052] For the high-resolution height grid Hc: it is constructed by down-sampling the ship unloader point cloud.

[0053] For the low-resolution height grid Hu: after PCA analysis of the ship unloader point cloud, a low-resolution height grid Hu is constructed (the resolution is adaptively adjusted).

[0054] Step S13, after obtaining the high-resolution height grid map Hc and the low-resolution height grid map Hu in step S13, the similarity of the high-resolution height grid map Hc and the low-resolution height grid map Hu is calculated, and the common area of the coal pile is determined according to the similarity.

[0055] In the embodiment, the correlation search is performed on the multi-scale height of the coal pile, and the similarity of the high-resolution height grid map Hc and the low-resolution height grid map Hu is calculated by using the normalized cross correlation (NCC).

[0056] The above scheme can solve the problem of sparse-dense point cloud registration, and significantly improve the robustness of initial registration. Specifically, the dependence on the easily failed local geometric features is avoided, the overall height distribution structure of the coal pile is directly matched, the volume difference is calculated through gridding, and the difficulty of direct matching of sparse-dense point clouds is effectively overcome, thereby providing a reliable initial value for subsequent optimization.

[0057] Step S2, subsequently, the point cloud registration is performed on the common area, and finally the initial transformation relationship between the ship unloader and the ship unloader is obtained.

[0058] By iteratively optimizing the grid resolution and the search position until the NCC value is maximized, the common area of the ship unloader and the ship unloader is obtained, and the point cloud registration is performed on the common area, and finally the optimal transformation initial value T is determined. init .

[0059] Step S3, a global constraint factor graph optimization model is constructed, the key frame point cloud data of the ship unloader, the map data and the point cloud data of the ship unloader are input, and finally the joint map of the ship unloader and the ship unloader is output.

[0060] Step S3 includes the following sub-steps.

[0061] Step S31, the key frame pose data and the point cloud data of the ship unloader are superimposed to generate the map data of the ship unloader; first, the pose data and the point cloud data of the key frame are read from the ship unloader, and then the position data and the point cloud data of multiple key frames are superimposed.

[0062] The map data of the ship unloader can be obtained by the above data superposition, that is, the cumulative data of the ship unloader from the i-th to the i+m-th key frame, and the map data of the ship unloader is equal to the product of the position data of the ship unloader at the i-th key frame and the point cloud data of the ship unloader at the i-th key frame, and the cumulative sum in the interval from the i-th to the i+m-th key frame, wherein i and m are both integers greater than 0.

[0063] The pose data and the point cloud data of the key frame of the ship unloader are read, and the data of multiple key frames are superimposed to obtain the cumulative data.

[0064] Step S32, the common area of the map data of the ship unloader and the point cloud data of the ship unloader is calculated according to step S1; in this process, the point cloud data of the ship unloader can be directly read from the radar of the ship unloader. After inputting the above-mentioned ship unloader point cloud data, the common area of the map data of the ship unloader and the point cloud data of the ship unloader is calculated by using the multi-scale height field correlation matching algorithm in step S1.

[0065] Step S33, the common area is registered to obtain the relative observation data between the ship unloader and the ship unloader; specifically, the common area of the ship unloader and the ship unloader is registered to obtain the relative observation data between the ship unloader and the ship unloader.

[0066] Step S34, an optimization function is generated in combination with the historical relative observation data, and an optimized relative transformation relationship is obtained according to the optimization function.

[0067] The historical relative observation data between the ship unloader and the ship unloader is obtained, and a corresponding observation data set is generated, and the optimization function is specifically: the optimized relative observation is equal to the transpose of the first parameter and the cumulative sum of the first parameter product taking the minimum value; wherein the first parameter is equal to the inverse of the state of the ship unloader at i time, the inverse of the relative observation data and the continuous product of the state of the ship unloader at j time.

[0068] The above-mentioned optimization function is used to optimize the relative observation data between the ship unloader and the ship unloader.

[0069] After optimization by the optimization function, the optimized relative transformation relationship, i.e. the above-mentioned optimized relative observation, can be obtained, and after obtaining the optimized relative transformation relationship, the whole joint mapping method based on global constraint factor graph optimization can be obtained.

[0070] The application realizes cross-time and space error compensation by constructing a factor graph optimization model that fuses multi-time key frames, including fusing ship unloader pose, ship unloader pose, relative observation constraint, historical observation constraint and other data.

[0071] Through the above-mentioned steps, cross-time and space error compensation can be realized, and the global consistency of the map is improved: the factor graph optimization framework fuses the data of the ship unloader-ship unloader equipment and the historical observation data at multiple times, constructs a cross-device and cross-time pose constraint network, effectively smooths single observation noise, significantly suppresses pose drift and error accumulation, and finally generates a globally consistent high-precision map.

[0072] The core steps of the present application are highly field matching, multi-time sequence factor graph and global map output, while the Chinese patent with publication number CN118363032A mainly reduces sampling, index marking, single frame GICP registration and pose output, and the technical problems solved (positioning vs. mapping) and the implementation paths of the two are essentially different. And the Chinese patent with publication number CN118363032A has the following limitations: dependent on local features: using generalized iterative closest point (GICP) to register the down-sampled point cloud, but the coal yard scene lacks stable local features, and the registration is prone to failure; no historical optimization: only processing current frame data, not constructing cross-time sequence constraints, leading to map drift over time; data homogenization assumption: requiring the ship unloader and the pusher point cloud to have "similar data volume", but the density difference between the stripping machine (dense) and the ship unloader (sparse) point cloud is too large to meet this condition.

[0073] The main improvements of the present application are as follows: abandoning structured feature matching: first creating "coal pile overall height distribution" as a non-structured matching feature to avoid relying on easily failed local geometric features; integrating multi-time sequence data: integrating historical key frame observations through factor graph optimization to suppress long-term drift; compatible heterogeneous point cloud: directly designing a registration process that takes into account the differences between sparse and dense point clouds.

[0074] The present application also includes a stripping machine and ship unloader joint mapping experiment: in the stripping machine and ship unloader joint mapping experiment, the stripping machine is deployed below the ship unloader and starts running. In the initial stage of the experiment, due to the pose deviation during independent mapping of the two, the maps generated by the stripping machine and the ship unloader respectively have a spatial misalignment problem. The point cloud data of the two are offset, which essentially reflects the difference in spatial positioning reference and data collection angle of the equipment. The ship unloader is installed higher than the stripping machine, and its coordinate system center is spatially misaligned with the stripping machine. After joint mapping, the pose deviation of single machine mapping is effectively eliminated, and a global unified spatial coordinate system is constructed. At the same time, the collaborative mapping mechanism of the two compensates for the single machine blind area, forming an integrated perception space covering the operating area, which provides accurate global maps for subsequent collaborative operation in complex scenarios.

[0075] Embodiment 2

[0076] The present embodiment proposes a joint mapping method based on coal pile height field matching and spatio-temporal global optimization, including the following steps.

[0077] Step S1, first, determine the common area according to the multi-scale height field correlation matching algorithm of the stripping machine and ship unloader coal pile point cloud, specifically taking the coal pile as a common feature body, and matching its overall height distribution structure by gridding to find the corresponding common area.

[0078] In the process of cleaning, the cleaning machine needs to pass through the common area of the coal pile to determine the conversion relationship. Figure 2 For the coordinate relationship between the cleaning machine and the ship unloader, the cleaning machine is in the cabin, and the ship unloader is above the cleaning machine. Since most of the coal pile area is in the cabin, the data captured by the radar of the ship unloader is less than that of the cleaning machine. If the iterative nearest point method is directly used, it is difficult to find an effective conversion relationship, so the common area of the coal pile of the cleaning machine and the ship unloader needs to be detected to ensure that the matching relationship is correctly found.

[0079] As Figure 3 shown is a multi-scale height field correlation matching algorithm using the coal pile point cloud of the cleaning machine and the ship unloader. The coal pile is used as an unstructured but common feature body, and the overall height distribution structure (rather than local feature points) is matched by gridding to find the common area.

[0080] Referring to Figure 1 , Figure 2 and Figure 3 , step S1 mainly includes the following sub-steps.

[0081] Step S11, the obtained point cloud data of the ship unloader is corrected, and then the size of the ship unloader point cloud is used to branch and bound the coal pile data of the cleaning machine.

[0082] For the above process, more specifically, the input ship unloader point cloud is subjected to PCA analysis. PCA (principal component analysis) is a principal component analysis technique that can convert multiple indicators into a few comprehensive indicators using dimensionality reduction. Specifically, the main direction of the ship unloader is perpendicular to the main direction of the cleaning machine, i.e. the ship unloader point cloud is rotated, to lay a foundation for subsequent data processing.

[0083] Then, according to the size of the ship unloader point cloud, the coal pile data of the cleaning machine is subjected to branch and bound to further clarify the data range, and then appropriate coal pile data is extracted from both sides of the ship unloader and cleaning machine point cloud data.

[0084] Step S12, after completing the above step S11, the coal pile data of the ship unloader and the cleaning machine are subjected to gridding processing to construct a coal pile height field, wherein the constructed coal pile height field mainly includes a high-resolution height grid Hc and a low-resolution height grid Hu.

[0085] For the high-resolution height grid Hc: it is constructed by down-sampling the cleaning machine point cloud.

[0086] For the low-resolution height grid Hu: after PCA analysis of the ship unloader point cloud, a low-resolution height grid Hu (resolution adaptively adjusted) is constructed.

[0087] Step S13, after obtaining the high-resolution height grid map Hc and the low-resolution height grid map Hu in step S13, the similarity of the high-resolution height grid map Hc and the low-resolution height grid map Hu is calculated, and the common area of the coal pile is determined according to the similarity.

[0088] In the embodiment, the correlation search is performed on the multi-scale height of the coal pile, and the similarity of the high-resolution height grid map Hc and the low-resolution height grid map Hu is calculated by using the normalized cross correlation (NCC) calculation.

[0089] The above scheme can solve the problem of sparse-dense point cloud registration, and significantly improve the robustness of initial registration. Specifically, the dependence on the easily failed local geometric features is avoided, the overall height distribution structure of the coal pile is directly matched, the difficulty of direct matching of sparse-dense point clouds is overcome through the grid and volume difference calculation, and a reliable initial value is provided for subsequent optimization.

[0090] Step S2, then, the point cloud registration is performed on the common area, and finally the initial transformation relationship between the ship unloader and the ship unloader is obtained.

[0091] By iteratively optimizing the grid resolution and the search position until the NCC value is maximized, the common area of the ship unloader and the ship unloader is obtained, and the point cloud registration is performed on the common area, and finally the optimal transformation initial value T is determined. init .

[0092] Step S3, a global constraint factor graph optimization model is constructed, the key frame point cloud data of the ship unloader, the map data and the point cloud data of the ship unloader are input, and finally the joint map of the ship unloader and the ship unloader is output.

[0093] Step S3 includes the following sub-steps.

[0094] Step S31, the key frame pose data and the point cloud data of the ship unloader are superimposed to generate the map data of the ship unloader; first, the pose data and the point cloud data of the key frame are read from the ship unloader, and then the position data and the point cloud data of the multiple key frames are superimposed.

[0095] The map data of the ship unloader can be obtained by the above data superposition, that is, the cumulative data of the ship unloader from the i-th to the i+m-th key frame, and the map data of the ship unloader is equal to the product of the position data of the ship unloader at the i-th key frame and the point cloud data of the ship unloader at the i-th key frame, and the cumulative sum in the interval from the i-th to the i+m-th key frame, wherein i and m are both integers greater than 0.

[0096] The pose data and the point cloud data of the key frames are superimposed to obtain accumulated data.

[0097] In step S32, a common area of the map data of the ship unloader and the point cloud data of the ship unloader is calculated according to step S1, and in this process, the point cloud data of the ship unloader can be directly read from the radar of the ship unloader.

[0098] In step S33, the common area is registered to obtain relative observation data between the ship unloader and the ship unloader.

[0099] In step S34, an optimization function is generated in combination with historical relative observation data, and an optimized relative transformation relationship is obtained according to the optimization function.

[0100] The historical relative observation data between the ship unloader and the ship unloader is obtained, and a corresponding observation data set is generated, and the optimization function is specifically: the optimized relative observation is equal to the minimum value of the cumulative sum of the transpose of the first parameter and the product of the first parameter; wherein the first parameter is equal to the inverse of the state of the ship unloader at i time, the inverse of the relative observation data and the continuous product of the state of the ship unloader at j time.

[0101] The optimization function is used to optimize the relative observation data between the ship unloader and the ship unloader.

[0102] After optimization by the optimization function, the optimized relative transformation relationship, i.e., the optimized relative observation, can be obtained, and after obtaining the optimized relative transformation relationship, the entire joint mapping method based on global constraint factor graph optimization can be obtained.

[0103] The present application realizes cross-time and space error compensation by constructing a factor graph optimization model that fuses multiple time sequence key frames, including data such as fusion of ship unloader pose, ship unloader pose, relative observation constraint, historical observation constraint, etc.

[0104] Through the above steps, cross-time and space error compensation can be realized, and the global consistency of the map can be improved: the factor graph optimization framework fuses the data of the ship unloader-ship unloader equipment and the historical observation data at multiple times, constructs a cross-device and cross-time pose constraint network, effectively smooths single observation noise, significantly suppresses pose drift and error accumulation, and finally generates a globally consistent high-precision map.

[0105] On the basis of embodiment 1, the embodiment also proposes a joint mapping system based on coal pile height field matching and space-time global optimization, which mainly includes an initial relationship construction module and an optimization module. The initial relationship construction module can execute the processes of steps S1 and S2. The optimization module can execute the process of step S3 based on the space-time global constraint factor graph optimization mapping method. The optimization module can be connected with the initial relationship construction module, and the joint mapping task can be completed through the cooperative work of the optimization module and the initial relationship construction module.

[0106] Embodiment 3

[0107] On the basis of embodiment 1, the factor graph optimization used can also be replaced by pose graph optimization, but the factor graph optimization is more flexible and supports multiple constraint types. The grid height difference calculation process used can also be replaced by voxel density matching, but the coal pile height difference in this embodiment is more distinguishable.

Claims

1. A joint mapping method based on coal pile height field matching and spatio-temporal global optimization, The method comprises the steps of: S1, using a multi-scale height field correlation matching algorithm of the coal pile point cloud of the ship unloader and the ship unloader, using the coal pile as a common feature body, and finding a common area by matching the overall height distribution structure of the coal pile through gridding; comprising: correcting the point cloud data of the ship unloader, and performing branch and bound on the coal pile data of the ship unloader according to the size of the point cloud data of the ship unloader; performing gridding processing on the coal pile data of the ship unloader and the ship unloader respectively to construct a coal pile height field, including a high-resolution height grid Hc and a low-resolution height grid Hu; calculating the similarity of Hc and Hu, and determining the common area of the coal pile according to the similarity; S2, performing point cloud registration on the common area to determine the initial transformation relationship between the ship unloader and the ship unloader; S3, constructing a global constraint factor graph optimization model, inputting the key frame point cloud and map data of the ship unloader and the point cloud data of the ship unloader, and outputting the joint map of the ship unloader and the ship unloader; comprising: superimposing the key frame pose data and the point cloud data of the ship unloader to generate the map data of the ship unloader; calculating the common area of the ship unloader map data and the ship unloader point cloud data from S1; registering the common area to obtain the relative observation data between the ship unloader and the ship unloader; generating an optimization function in combination with historical relative observation data, and obtaining an optimized relative transformation relationship from the optimization function; The left side of the optimization function equation is the optimized relative observation, and the right side of the optimization function equation is the minimum value of the cumulative sum of the transpose of the first parameter and the product of the first parameter; the first parameter is the inverse of the state of the ship unloader at time i, the inverse of the relative observation data, and the continuous product of the state of the ship unloader at time j.

2. The method according to claim 1, wherein, After optimization by the optimization function, the optimized relative transformation relationship, i.e. the optimized relative observation, can be obtained, and after obtaining the optimized relative transformation relationship, the entire joint mapping method based on global constraint factor graph optimization can be obtained.

3. The method according to claim 1, wherein, The point cloud data of the ship unloader is corrected, comprising: performing PCA analysis on the input point cloud of the ship unloader, making the main direction of the ship unloader perpendicular to the main direction of the ship unloader, and rotating the point cloud of the ship unloader.

4. The method according to claim 1, wherein, The coal pile height field is constructed, specifically comprising: down-sampling the point cloud data of the ship unloader, and then constructing a high-resolution height grid Hc; after PCA alignment of the point cloud data of the ship unloader, constructing a low-resolution height grid Hu.

5. The method according to claim 1, wherein, The similarity of Hc and Hu is calculated by using a normalized cross-correlation algorithm.

6. The method according to claim 1 or 5, wherein, After the similarity of Hc and Hu is calculated, the grid resolution and the search position are iteratively optimized until the similarity is maximized, at which time the common area of the ship unloader and the ship unloader is found according to the maximized similarity.

7. The method according to claim 1, wherein, The pose data and the point cloud data of the ship unloader are read to perform superposition processing on the data of multiple key frames to obtain cumulative data.

8. The method according to claim 2, wherein, The map data of the ship unloader is the cumulative sum of the product of the pose data of the i-th key frame of the ship unloader and the point cloud data of the i-th key frame of the ship unloader.

Images