Material inventory method based on 3D model construction

Abstract

The present application provides a material inventory method based on three-dimensional model construction, including: obtaining point cloud data of a material warehouse; constructing a reconstruction model of the internal space of the material warehouse based on the point cloud data, wherein the reconstruction model includes materials piled up in the material warehouse; determining a material accumulation model based on the reconstruction model and a reference model of the internal space of the material warehouse, wherein the reference model is a model when there is no material in the internal space of the material warehouse; and counting the material remainder based on the material accumulation model. The reconstruction model of the internal space of the material warehouse is aligned with the reference model of the internal space of the material warehouse to determine the material accumulation model, thereby reducing errors and improving the accuracy of material inventory.

Description

Technical Field

[0001] The present application relates to the field of material storage and management, and in particular, to a material inventory method based on three-dimensional model construction. Background Art

[0002] At present, the traditional silo inventory method still relies on the traditional manual estimation method. The workload of the inventory department is huge, and the accuracy cannot be guaranteed, which cannot meet the requirements of the company's refined production. With the improvement of industrial automation, automatic inventory in silos will gradually replace traditional manual operations. Using the automatic inventory system, a real-time 3D model of the entire material yard is established to truly and dynamically reflect the working conditions of the material yard and the material inventory conditions (goods types and quantities) of the material yard, and to realize the real-time information display of the yard during stacking and unloading of materials, helping enterprise management to gradually transform from traditional extensive to modern intensive in a digital way. It can not only improve the level of enterprise management, but also gradually become one of the main platforms for decision-making analysis and production scheduling and command of management departments.

[0003] There are also some relevant literatures that propose some ideas for using 3D modeling to realize material inventory. However, in the field of coal storage and inventory, there are high errors when the 3D modeling solutions from other fields are directly migrated to this field (because when taking inventory of materials, the stacking area is usually very large, and it is impossible to scan the point cloud data of the bottom of the stacking, and the solution of scanning from the top down to obtain the point cloud is usually adopted). When such traditional solutions are applied to this scenario, they usually only use denoising to improve data accuracy without considering error correction, resulting in low accuracy of material balance estimation. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a material inventory method based on three-dimensional model construction, so as to reduce modeling errors through correction and improve the accuracy of material inventory.

[0005] In order to achieve the above purpose, the embodiments of the present application are implemented in the following manner:

[0006] In a first aspect, an embodiment of the present application provides a material inventory method based on a three-dimensional model, comprising: acquiring point cloud data of a material warehouse; based on the point cloud data, constructing a reconstruction model of the internal space of the material warehouse, wherein the reconstruction model includes materials stacked in the material warehouse; based on the reconstruction model and a reference model of the internal space of the material warehouse, determining a material stacking model, wherein the reference model is a model when no materials exist in the internal space of the material warehouse; and based on the material stacking model, taking inventory of material remainders.

[0007] In combination with the first aspect, in a first possible implementation method of the first aspect, a reconstruction model of the internal space of the material warehouse is constructed based on the point cloud data, including: filtering noise points on the point cloud data; performing three-dimensional reconstruction based on the point cloud data after noise filtering to determine the reconstruction model.

[0008] In combination with the first aspect, in a second possible implementation of the first aspect, the internal space of the material warehouse has a warehouse floor, surrounding walls and a ceiling top surface, there is an aisle area on the warehouse bottom surface, materials are not stacked in the aisle area, the range of material stacking does not exceed the boundary constructed by the surrounding walls, and the material stacking depth is lower than the ceiling top surface, and a material stacking model is determined based on the reconstructed model and a reference model of the internal space of the material warehouse, including: determining the warehouse floor, surrounding walls and ceiling top surface from the reconstructed model; aligning the reconstructed model with the reference model of the internal space of the material warehouse based on the warehouse floor, surrounding walls and ceiling top surface; and determining the material stacking model based on the aligned reconstructed model and reference model.

[0009] In combination with the second possible implementation method of the first aspect, in a third possible implementation method of the first aspect, point cloud data is obtained by scanning with several three-dimensional laser scanners arranged at the top of the internal space of the material warehouse, and the bottom surface of the warehouse, the surrounding walls and the top surface of the ceiling are determined from the reconstructed model, including: determining the aisle area from the reconstructed model, and determining the bottom surface of the warehouse based on the aisle area; determining the surrounding walls from the reconstructed model; and determining a reference plane based on the position of the three-dimensional laser scanner as the top surface of the ceiling.

[0010] In combination with the third possible implementation method of the first aspect, in a fourth possible implementation method of the first aspect, an aisle area is determined from the reconstructed model, and the bottom surface of the warehouse is determined based on the aisle area, including: determining a pending area whose depth is within a set range from the reconstructed model; if the area of ​​the pending area is not less than the set area, determining that the pending area is the aisle area; calculating the average depth of the aisle area, and determining the plane where the average depth is located as the bottom surface of the warehouse.

[0011] In combination with the third possible implementation method of the first aspect, in a fifth possible implementation method of the first aspect, a reference plane is determined based on the position of the three-dimensional laser scanner, including: obtaining the position center of each three-dimensional laser scanner and the corresponding reference distance, wherein the reference distance represents the distance from the position center of the three-dimensional laser scanner to the top surface set in the reference model of the internal space of the material warehouse; obtaining the spatial mapping point corresponding to each three-dimensional laser scanner by combining the position center of each three-dimensional laser scanner with the corresponding reference distance; if the spatial mapping point meets the judgment condition: half or more of the spatial mapping points are located in the same depth plane in the reconstructed model, then this depth plane is determined to be the reference plane; otherwise, the depth mean of the spatial mapping points corresponding to all the three-dimensional laser scanners is calculated, and the horizontal plane where the depth mean is located in the reconstructed model is determined to be the reference plane.

[0012] In combination with the fifth possible implementation manner of the first aspect, in a sixth possible implementation manner of the first aspect, based on the warehouse bottom surface, surrounding walls and ceiling top surface, aligning the reconstructed model with the reference model of the internal space of the material warehouse includes: calculating the distance d between the ceiling top surface in the reconstructed model and the set top surface in the reference model; mtop,rtop , and calculate the distance d between the warehouse bottom surface in the reconstructed model and the set bottom surface in the reference model mbase,rbase , where d mtop,rtop and d mbase,rbase Both positive and negative; based on the distance d mtop,rtop , align the ceiling top surface in the rebuilt model with the set top surface in the reference model, and based on the distance d mtop,rtop and distance d mbase,rbase , align the warehouse bottom surface in the rebuilt model with the set bottom surface in the reference model; align the surrounding walls in the rebuilt model with the corresponding set walls in the reference model.

[0013] In combination with the sixth possible implementation manner of the first aspect, in a seventh possible implementation manner of the first aspect, based on the distance d mtop,rtop , align the ceiling top surface in the rebuilt model with the set top surface in the reference model, and based on the distance d mtop,rtop and distance d mbase,rbase , align the warehouse bottom surface in the reconstructed model with the set bottom surface in the reference model, including: determining the distance d mtop,rtop The compensation distance d com ; Based on the compensation distance d com Align the ceiling top surface in the reconstructed model with the set top surface in the reference model; based on the compensation distance d com and distance d mbase,rbase, calculate the final bottom surface distance difference Δd, and use Δd to stretch or compress the reconstructed model along the depth direction to align the warehouse bottom surface in the reconstructed model with the set bottom surface in the reference model.

[0014] In combination with the seventh possible implementation manner of the first aspect, in an eighth possible implementation manner of the first aspect, if the ceiling top surface in the reconstructed model is determined by satisfying the determination condition, counting the material remainder based on the material accumulation model includes: for each point in the material accumulation surface of the material accumulation model, calculating the distance between each point and the ceiling top surface, and the distance between the i-th point and the ceiling top surface is recorded as d i ; Use the following formula to calculate the material balance:

[0015]

[0016] Among them, M rest is the material surplus, V is the material warehouse capacity, and n is the number of midpoints on the material stacking surface.

[0017] In combination with the seventh possible implementation manner of the first aspect, in a ninth possible implementation manner of the first aspect, if the ceiling top surface in the reconstructed model is not determined by satisfying the determination condition, counting the material remainder based on the material accumulation model includes: for each point in the material accumulation surface of the material accumulation model, calculating the distance between each point and the ceiling top surface, and the distance between the i-th point and the ceiling top surface is recorded as d i ; Use the following formula to calculate the material balance:

[0018]

[0019] Among them, M rest is the material surplus, V is the material warehouse capacity, n is the number of midpoints on the material accumulation surface, p i is d i The corresponding weights satisfy:

[0020]

[0021] Among them, h b Sets the depth of the bottom surface for the reference model.

[0022] Beneficial effects:

[0023] 1. By acquiring the point cloud data of the material warehouse, a reconstruction model of the internal space of the material warehouse is further constructed; then, based on the reconstruction model and the reference model of the internal space of the material warehouse (the model when there is no material in the internal space of the material warehouse), the material accumulation model is determined to further complete the inventory of material balances. The warehouse floor, surrounding walls and ceiling top of the internal space of the material warehouse (there is an aisle area on the bottom of the warehouse, materials are not stacked in the aisle area, the range of material stacking does not exceed the boundary constructed by the surrounding walls, and the material stacking depth is lower than the ceiling top) are aligned with the reference model of the internal space of the material warehouse to determine the material accumulation model, which can reduce errors and improve the accuracy of material inventory.

[0024] 2. In the process of determining the reference plane based on the location of the 3D laser scanner, obtain the location center of each 3D laser scanner and the corresponding reference distance (indicating the distance from the location center of the 3D laser scanner to the top surface set in the reference model of the internal space of the material warehouse); after combining the location center of each 3D laser scanner with the corresponding reference distance, obtain the space mapping point corresponding to each 3D laser scanner; if the space mapping point meets the judgment condition: half or more of the space mapping points are located in the same depth plane in the reconstructed model, then determine this depth plane as the reference plane. Otherwise, calculate the depth mean of the space mapping points corresponding to all 3D laser scanners, and determine the horizontal plane where this depth mean is located in the reconstructed model as the reference plane. This judgment condition can reflect and distinguish whether there is an error caused by a change in the position of the equipment, so as to provide targeted guidance for the alignment and final material inventory process, reduce errors, and improve the accuracy of material inventory.

[0025] 3. Determine the distance d mtop,rtop (calculated based on the ceiling top surface in the reconstructed model and the set top surface in the reference model) is the compensation distance d com ; Based on the compensation distance d com Align the ceiling top surface in the reconstructed model with the set top surface in the reference model; based on the compensation distance d com and distance d mbase,rbase(calculated based on the warehouse bottom surface in the reconstructed model and the set bottom surface in the reference model), calculate the final bottom surface distance difference Δd, and use Δd to stretch or compress the reconstructed model along the depth direction to align the warehouse bottom surface in the reconstructed model with the set bottom surface in the reference model. When taking inventory of materials, different material remainder calculation formulas are designed according to whether the ceiling top surface in the reconstructed model is determined by satisfying the judgment conditions. When the judgment conditions are met (indicating that the error mainly comes from the difference caused by the collected data, rather than the displacement of the equipment, even if multiple devices produce synchronous displacement at the same time (such as the height change of the horseway where the laser scanner is installed), it is overcome by aligning the ceiling top surface with the set top surface by compensating the distance), the material remainder is calculated by the weighted calculation method; when the judgment conditions are not met, the material remainder is calculated by the weighted method (this situation indicates that the error is likely to come from both the difference caused by the collected data and the error caused by the displacement of the equipment, and the error of the equipment displacement cannot be completely eliminated by alignment. Therefore, on the one hand, the error is reduced by using the mean calculation method when calculating the height of the ceiling top surface, and on the other hand, when the material inventory is finally taken, the material depth is calculated by designing the point depth weighted method). The larger the material depth (the distance from the point in the material accumulation surface of the material accumulation model to the set top surface depth is subtracted from the set bottom surface depth), the less the error impact of the unit depth is, so a higher weight is given to reduce the error and improve the accuracy of material inventory.

[0026] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 Schematic diagram of the material inventory system.

[0029] Figure 2 A flow chart of a material inventory method based on a three-dimensional model provided in an embodiment of the present application.

[0030] Figure 3 Schematic diagram of the material accumulation model.

[0031] Figure 4 This is a schematic diagram displayed on the terminal after taking inventory of materials in the material warehouse. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0033] See also Figure 1 , Figure 1 Schematic diagram of the material inventory system.

[0034] In this embodiment, the material inventory system includes a terminal, a platform server, and several three-dimensional laser scanners. For example, for each material yard (including a material warehouse, each material warehouse has a warehouse bottom, surrounding walls, and a ceiling top in its internal space, and there is an aisle area on the warehouse bottom, materials are not stacked in the aisle area, the range of material stacking does not exceed the boundary constructed by the surrounding walls, and the material stacking depth is lower than the ceiling top), multiple three-dimensional laser scanners are configured (for example, a material warehouse is equipped with 3 or 5 three-dimensional laser scanners, which are placed at the top of the internal space of the material warehouse) to complete the point cloud data scanning of the material yard (i.e., the internal space of the material warehouse), and the server is used to run the material inventory method based on the three-dimensional model to implement material inventory, and the material inventory results can be sent to the terminal for reference and supervision by management personnel. The material inventory method based on the three-dimensional model construction applied to the server will be introduced in detail below.

[0035] See also Figure 2 , Figure 2 A flow chart of a material inventory method based on a three-dimensional model provided in an embodiment of the present application. In this embodiment, the material inventory method based on a three-dimensional model may include step S10, step S20, step S30 and step S40.

[0036] First, the server may execute step S10.

[0037] Step S10: Acquire point cloud data of the material warehouse.

[0038] The 3D laser scanner installed on the top of the internal space of the material warehouse is used to scan the interior of the material warehouse downward to obtain the point cloud data of the material warehouse.

[0039] After obtaining the point cloud data of the material warehouse, the server may run step S20.

[0040] Step S20: constructing a reconstruction model of the internal space of the material warehouse based on the point cloud data, wherein the reconstruction model includes the materials piled up in the material warehouse.

[0041] In this embodiment, point cloud data can be used to construct a reconstruction model of the internal space of the material warehouse (including the materials piled in the material warehouse). In this process, the point cloud data can be firstly filtered for noise points, and then three-dimensional reconstruction can be performed based on the point cloud data after noise point filtering to determine the reconstruction model. Since the modeling process of this solution adopts the mature three-dimensional reconstruction technology, it will not be described here.

[0042] After reconstructing the reconstruction model of the internal space of the material warehouse, the server may execute step S30.

[0043] Step S30: determining a material stacking model based on the reconstructed model and a reference model of the internal space of the material warehouse, wherein the reference model is a model when no materials exist in the internal space of the material warehouse.

[0044] In this embodiment, the server can obtain a reference model of the internal space of the material warehouse, which is a model when there is no material in the internal space of the material warehouse, that is, a model obtained by scanning and reconstruction when there is no material, including a set bottom surface, a set wall surface, and a set top surface. The reference model can be understood as an empty container, which is used to compare the difference with the reconstructed model to determine the difference as the material stacking model.

[0045] After obtaining the reference model, the server can determine the material stacking model based on the reconstructed model and the reference model of the internal space of the material warehouse.

[0046] First, the server can determine the warehouse bottom surface, surrounding walls and ceiling top surface from the reconstructed model. For example, since there is an aisle area on the warehouse bottom surface of the internal space of the material warehouse, and materials are not stacked in the aisle area, the server can determine the aisle area from the reconstructed model and determine the warehouse bottom surface based on the aisle area.

[0047] Specifically, the server can determine the undetermined area whose depth is within the set range from the reconstructed model, wherein the set range is determined based on the maximum depth in the reconstructed model. For example, a range is determined based on the maximum depth in the reconstructed model. For example, the maximum depth is used as the upper limit of the set range, and 0.01 times (or other values) of the maximum depth are subtracted from this, that is, [0.99x, x], where x represents the maximum depth in the reconstructed model.

[0048] Then, the server can determine whether the area of ​​the pending area is lower than the set area. If the area of ​​the pending area is not lower than the set area (for example, the area of ​​the aisle area multiplied by a coefficient of 0.8 is used as the set area), the pending area can be determined to be the aisle area. Then the average depth of the aisle area is calculated (it can be a weighted average depth, for example, weighting is performed according to the range proportion of each depth in the aisle area to reduce the error caused by the edge of the material pile that is divided into the pending area), and the plane where this average depth is located is determined to be the bottom surface of the warehouse.

[0049] At the same time, the server can determine the surrounding walls from the reconstructed model. Since the modeling error of the surrounding walls is very small (the impact on material inventory is very small), and it is easy to determine the surrounding walls (the fitted surrounding walls have very obvious features, and generally speaking, there are points with continuous changes in height when the horizontal plane coordinates are fixed), they will not be elaborated here.

[0050] Furthermore, the server may determine a reference plane as the top surface of the ceiling based on the location of the three-dimensional laser scanner.

[0051] For example, the server can obtain the location center of each 3D laser scanner and the corresponding reference distance, wherein the reference distance represents the distance from the location center of the 3D laser scanner to the top surface set in the reference model of the internal space of the material warehouse. Then, the server combines the location center of each 3D laser scanner with the corresponding reference distance (for example, vertically descending this distance, i.e., increasing the depth) to obtain the space mapping point corresponding to each 3D laser scanner.

[0052] Then, the server can determine whether the spatial mapping points meet the judgment condition (half or more of the spatial mapping points are located in the same depth plane in the reconstructed model. The same depth plane here may have a certain slight error. For example, the distance between the planes does not exceed 1 cm, which is considered to be the same depth plane):

[0053] If the determination condition is met: the server may determine that this depth plane (ie, the depth plane where half or more of the spatial mapping points are located in the reconstructed model) is a reference plane.

[0054] If the judgment condition is not met: the server can calculate the average depth of the spatial mapping points corresponding to all three-dimensional laser scanners in the material warehouse, and determine the horizontal plane where the depth average is located in the reconstructed model as the reference plane.

[0055] After determining the reference plane (as the top surface of the ceiling), the server can align the reconstructed model with the reference model of the internal space of the material warehouse based on the warehouse bottom surface, surrounding walls and ceiling top surface.

[0056] For example, the server may calculate the distance d between the ceiling top surface in the reconstructed model and the set top surface in the reference model. mtop,rtop For example, the depth of the ceiling top surface in the reconstructed model minus the depth of the set top surface in the reference model is used to obtain the distance d mtop,rtop ; and, calculating the distance d between the warehouse bottom surface in the reconstructed model and the set bottom surface in the reference model mbase,rbase For example, the distance d is obtained by subtracting the depth of the set bottom surface in the reference model from the depth of the bottom surface of the warehouse in the reconstructed model. mbase,rbase , of course, mtop,rtop and d mbase,rbase Both have positive and negative signs.

[0057] Afterwards, the server can mtop,rtop , align the ceiling top surface in the rebuilt model with the set top surface in the reference model, and based on the distance d mtop,rtop and distance d mbase,rbase , align the bottom surface of the warehouse in the rebuilt model with the set bottom surface in the reference model.

[0058] For example, the server may determine the distance d mtop,rtop The compensation distance d com , and then use the compensation distance d com Align the ceiling top surface in the rebuilt model with the set top surface in the reference model (by changing the depth d of the ceiling top surface in the rebuilt model to mtop,rtop , minus the compensation distance d com , that is, aligning the ceiling top surface in the reconstructed model with the set top surface in the reference model). And, the server can be based on the compensation distance d com and distance d mbase,rbase , calculate the final bottom distance difference Δd, and use Δd to stretch or compress the reconstructed model in the depth direction to align the warehouse bottom surface in the reconstructed model with the set bottom surface in the reference model. As for the alignment between the surrounding walls in the reconstructed model and the corresponding set walls in the reference model, since this is easy to align, and the adjustment operation used during alignment has little effect on material inventory (because materials are generally rarely stacked against the wall, and the material stacking area accounts for a large proportion of the total area of ​​the material warehouse), no special instructions are given here, and alignment can be performed according to the commonly used alignment method. In this way, the alignment between the reconstructed model and the reference model can be completed.

[0059] After the reconstructed model is aligned with the reference model, the material accumulation model can be obtained, and the material accumulation model is presented as a material accumulation surface with different depths, such as Figure 3 shown.

[0060] Based on this, the server may execute step S40.

[0061] Step S40: Count the remaining material quantity based on the material accumulation model.

[0062] In this embodiment, the server may perform calculations in two different cases according to whether the ceiling top surface in the reconstructed model meets the determination conditions, so as to reduce the error of material inventory and improve the inventory accuracy.

[0063] For the first case (the ceiling top surface in the reconstructed model is determined by satisfying the judgment conditions):

[0064] If the ceiling top surface in the reconstructed model is determined by satisfying the judgment conditions (this situation indicates that there is almost no error caused by equipment displacement, or that the error caused by this part of equipment displacement is overcome by alignment), then for each point in the material accumulation surface of the material accumulation model, the server can calculate the distance between each point and the ceiling top surface. The distance between the i-th point and the ceiling top surface is recorded as d i , and then use the following formula to calculate the material balance:

[0065]

[0066] Among them, M rest is the material balance, V is the material warehouse capacity, and n is the number of points in the material accumulation surface.

[0067] For the second case (the top surface of the ceiling in the reconstructed model is not determined by satisfying the judgment conditions):

[0068] If the ceiling top surface in the reconstructed model is not determined by satisfying the judgment conditions (this situation indicates that the error is likely to come from both the difference caused by the collected data and the error caused by the displacement of the equipment, and alignment cannot completely eliminate the error of the equipment displacement. Therefore, on the one hand, the error is reduced by using the mean calculation method when calculating the height of the ceiling top surface, and on the other hand, when the material inventory is finally taken, the material depth is calculated by designing a point depth weighted method), then, for each point in the material accumulation surface of the material accumulation model, the server can calculate the distance between each point and the ceiling top surface, and the distance between the i-th point and the ceiling top surface is recorded as d i The server then calculates the material balance using the following formula:

[0069]

[0070] Among them, M rest is the material surplus, V is the material warehouse capacity, n is the number of midpoints on the material accumulation surface, p i is d i The corresponding weights satisfy:

[0071]

[0072] Among them, h b Sets the depth of the bottom surface for the reference model.

[0073] Of course, if we do not consider the amount of data calculation and sacrifice some calculation efficiency, in order to further optimize the error, p i It can also be designed to meet:

[0074]

[0075] Among them, h b Set the depth of the bottom surface in the reference model, h r Sets the depth of the top surface in the reference model. This takes into account the distance d i The depth h of the bottom surface is set b The relationship between the distance d mtop,rtop The influence of (through 0.005h r quantify the errors), thereby reducing the impact of the errors and improving the accuracy of material inventory.

[0076] Utilize the material inventory system (i.e. Figure 4 The digital material yard system in the material warehouse takes inventory of the materials and displays the schematic diagram through the terminal as follows Figure 4 As shown, the specific data has been coded.

[0077] In summary, the embodiment of the present application provides a material inventory method based on three-dimensional model construction, by acquiring the point cloud data of the material warehouse, further constructing a reconstruction model of the internal space of the material warehouse; then based on the reconstruction model and the reference model of the internal space of the material warehouse (the model when there is no material in the internal space of the material warehouse), determine the material accumulation model, and further complete the inventory of the material balance. The warehouse bottom surface, surrounding walls and ceiling top surface of the internal space of the material warehouse (there is an aisle area on the bottom surface of the warehouse, materials are not stacked in the aisle area, the range of material stacking does not exceed the boundary constructed by the surrounding walls, and the material stacking depth is lower than the ceiling top surface) are aligned with the reference model of the internal space of the material warehouse to determine the material accumulation model, thereby reducing errors and improving the accuracy of material inventory.

[0078] In the process of determining the reference plane based on the location of the 3D laser scanner, the location center of each 3D laser scanner and the corresponding reference distance (indicating the distance from the location center of the 3D laser scanner to the top surface set in the reference model of the internal space of the material warehouse) are obtained; the location center of each 3D laser scanner is combined with the corresponding reference distance to obtain the space mapping point corresponding to each 3D laser scanner; if the space mapping point meets the judgment condition: half or more of the space mapping points are located in the same depth plane in the reconstructed model, then this depth plane is determined to be the reference plane. Otherwise, the depth mean of the space mapping points corresponding to all 3D laser scanners is calculated, and the horizontal plane where this depth mean is located in the reconstructed model is determined to be the reference plane. This judgment condition can reflect and distinguish whether there is an error caused by a change in the position of the equipment, thereby specifically guiding the alignment and final material inventory process, reducing errors, and improving the accuracy of material inventory.

[0079] Determine the distance d mtop,rtop (calculated based on the ceiling top surface in the reconstructed model and the set top surface in the reference model) is the compensation distance d com ; Based on the compensation distance d com Align the ceiling top surface in the reconstructed model with the set top surface in the reference model; based on the compensation distance d com and distance d mbase,rbase (calculated based on the warehouse bottom surface in the reconstructed model and the set bottom surface in the reference model), calculate the final bottom surface distance difference Δd, and use Δd to stretch or compress the reconstructed model along the depth direction to align the warehouse bottom surface in the reconstructed model with the set bottom surface in the reference model. When taking inventory of materials, different material remainder calculation formulas are designed according to whether the ceiling top surface in the reconstructed model is determined by satisfying the judgment conditions. When the judgment conditions are met (indicating that the error mainly comes from the difference caused by the collected data, rather than the displacement of the equipment, even if multiple devices produce synchronous displacement at the same time (such as the height change of the horseway where the laser scanner is installed), it is overcome by aligning the ceiling top surface with the set top surface by compensating the distance), the material remainder is calculated by the weighted calculation method; when the judgment conditions are not met, the material remainder is calculated by the weighted method (this situation indicates that the error is likely to come from both the difference caused by the collected data and the error caused by the displacement of the equipment, and the error of the equipment displacement cannot be completely eliminated by alignment. Therefore, on the one hand, the error is reduced by using the mean calculation method when calculating the height of the ceiling top surface, and on the other hand, when the material inventory is finally taken, the material depth is calculated by designing the point depth weighted method). The larger the material depth (the distance from the point in the material accumulation surface of the material accumulation model to the set top surface depth is subtracted from the set bottom surface depth), the less the error impact of the unit depth is, so a higher weight is given to reduce the error and improve the accuracy of material inventory.

[0080] In this document, relational terms such as first and second, etc. are used merely to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0081] The above description is only an embodiment of the present application and is not intended to limit the protection scope of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A material inventory method based on a three-dimensional model, characterized in that: include: Obtain point cloud data of material warehouse; Based on the point cloud data, a reconstruction model of the internal space of the material warehouse is constructed, wherein the reconstruction model includes the materials piled up in the material warehouse; Based on the reconstructed model and the reference model of the internal space of the material warehouse, a material accumulation model is determined, wherein the reference model is a model when there is no material in the internal space of the material warehouse; Based on the material accumulation model, count the material balance; The internal space of the material warehouse consists of the warehouse floor, surrounding walls and ceiling top. There is an aisle area on the warehouse floor. Materials are not stacked in the aisle area. The material stacking range does not exceed the boundary constructed by the surrounding walls. In addition, the material stacking depth is lower than the ceiling top. Based on the reconstruction model and the reference model of the internal space of the material warehouse, the material accumulation model is determined, including: Determine the warehouse bottom surface, surrounding walls and ceiling top surface from the reconstructed model; Based on the warehouse floor, surrounding walls and ceiling, align the reconstructed model with the reference model of the material warehouse's interior space; Based on the aligned reconstructed model and the reference model, a material accumulation model is determined; The point cloud data is obtained by scanning with several 3D laser scanners installed on the top of the internal space of the material warehouse. The warehouse bottom surface, surrounding walls and ceiling top surface are determined from the reconstructed model, including: Determine the aisle area from the reconstructed model, and determine the warehouse floor based on the aisle area; Determine the surrounding walls from the reconstructed model; Based on the location of the 3D laser scanner, a reference plane is determined as the top surface of the ceiling; Based on the location of the 3D laser scanner, a reference plane is determined, including: Obtaining the location center of each 3D laser scanner and the corresponding reference distance, wherein the reference distance represents the distance from the location center of the 3D laser scanner to the top surface set in the reference model of the internal space of the material warehouse; By combining the location center of each 3D laser scanner with the corresponding reference distance, the spatial mapping point corresponding to each 3D laser scanner is obtained; If the spatial mapping points meet the judgment condition: half or more of the spatial mapping points are located in the same depth plane in the reconstructed model, then this depth plane is determined to be the reference plane; Otherwise, the depth mean of all the spatial mapping points corresponding to the three-dimensional laser scanner is calculated, and the horizontal plane where the depth mean is located in the reconstructed model is determined as the reference plane; Based on the warehouse floor, surrounding walls and ceiling, align the reconstructed model with the reference model of the material warehouse interior space, including: Calculate the distance d between the ceiling top surface in the reconstructed model and the set top surface in the reference model mtop,rtop , and calculate the distance d between the warehouse bottom surface in the reconstructed model and the set bottom surface in the reference model mbase,rbase , where d mtop,rtop and d mbase,rbase Both have positive and negative sides; Based on the distance d mtop,rtop , align the ceiling top surface in the rebuilt model with the set top surface in the reference model, and based on the distance d mtop,rtop and distance d mbase,rbase , align the warehouse bottom surface in the rebuilt model with the set bottom surface in the reference model; Align the surrounding walls in the rebuilt model with the corresponding set walls in the reference model; Based on the distance d mtop,rtop , align the ceiling top surface in the rebuilt model with the set top surface in the reference model, and based on the distance d mtop,rtop and distance d mbase,rbase , align the warehouse bottom surface in the rebuilt model with the set bottom surface in the reference model, including: Determine the distance d mtop,rtop The compensation distance d com ; Based on the compensation distance d com Align the ceiling top surface in the rebuilt model with the set top surface in the reference model; Based on the compensation distance d com and distance d mbase,rbase , calculate the final bottom surface distance difference Δd, and use Δd to stretch or compress the reconstructed model along the depth direction to align the warehouse bottom surface in the reconstructed model with the set bottom surface in the reference model; If the ceiling top surface in the rebuilt model is not determined by satisfying the judgment conditions, the material balance is counted based on the material accumulation model, including: For each point in the material accumulation surface of the material accumulation model, the distance between each point and the top surface of the ceiling is calculated. The distance between the i-th point and the top surface of the ceiling is recorded as d i ; The material balance is calculated using the following formula: Among them, M rest is the material surplus, V is the material warehouse capacity, S is the area of ​​the ceiling, n is the number of midpoints on the material stacking surface, and p i is d i The corresponding weights satisfy: Among them, h b Sets the depth of the bottom surface for the reference model.

2. The material inventory method based on three-dimensional model construction according to claim 1 is characterized in that: Based on the point cloud data, a reconstruction model of the internal space of the material warehouse is constructed, including: Filter the noise of point cloud data; Three-dimensional reconstruction is performed based on the point cloud data after noise filtering to determine the reconstructed model.

3. The material inventory method based on three-dimensional model construction according to claim 1 is characterized in that: Determine the aisle area from the reconstructed model, and determine the warehouse floor based on the aisle area, including: Determine a pending area with a depth within a set range from the reconstructed model; If the area of ​​the pending area is not less than the set area, the pending area is determined to be the aisle area; Calculate the average depth of the aisle area and determine the plane where this average depth lies as the bottom of the warehouse.

4. The material inventory method based on three-dimensional model construction according to claim 1 is characterized in that: If the ceiling top surface in the rebuilt model is determined by satisfying the judgment conditions, the material balance is counted based on the material accumulation model, including: For each point in the material accumulation surface of the material accumulation model, the distance between each point and the top surface of the ceiling is calculated. The distance between the i-th point and the top surface of the ceiling is recorded as d i ; The material balance is calculated using the following formula: Among them, M rest is the material surplus, V is the material warehouse capacity, S is the area of ​​the ceiling top surface, and n is the number of midpoints on the material stacking surface.

Images