Stockyard stocktaking method and system

Abstract

The invention discloses a stockyard inventory making method and system, and relates to the technical field of stockyard inventory making. The method comprises the steps of obtaining original point cloud data of a stockyard; filtering the original point cloud data to obtain filtered point cloud data; segmenting the filtering point cloud data to obtain target modeling point data and non-modeling point data; constructing a three-dimensional model of the material pile in the storage yard based on the target modeling point data; and the volume of the material pile is calculated according to the three-dimensional model, and stocktaking is completed. Based on the scanner and the server, the influence of environmental factors and flying dust interference on the point cloud data can be removed, and the optimal point cloud denoising effect is realized, so that a more accurate three-dimensional model can be established, and the stocktaking precision is improved. Besides, the scanner is installed on a riding track at the top of the greenhouse, the problem that the stocktaking function of an existing remote control system of the bucket-wheel stacker-reclaimer occupies the operation time of the bucket-wheel stacker-reclaimer can be solved, real-time modeling stocktaking is achieved, the refreshing rate is higher, and the stocktaking speed is increased.

Description

Technical Field

[0001] The present invention relates to the technical field of storage yard inventory, in particular to a storage yard inventory method. Background Art

[0002] In existing technology, yard inventory functions are primarily implemented through remote control systems for bucket-wheel stackers and reclaimers. Specifically, a 2D or 3D laser scanner is installed on the stacker / reclaimer to scan and model the yard. A 2D laser scanner uses its boom to horizontally cover the material surface while the crane travels, achieving scanning and modeling. A 3D laser scanner is mounted on a platform above the crane's boom. The crane must remain stationary during scanning, while the scanner scans autonomously. The data is then calculated by a server based on the crane's 3D coordinates. However, this method relies on the stacker / reclaimer, which consumes operating time. Furthermore, the laser scanner is susceptible to weather conditions, such as dust from rainy or windy days, which can affect modeling accuracy. Furthermore, point cloud data is typically acquired using high-frequency millimeter-wave radar, and a series of algorithmic optimization schemes are employed to correct only the Z-axis coordinates of the point cloud, failing to account for environmental factors and dust interference. Dry bulk stockpile volume measurement methods using yard cameras reconstruct the yard's 3D structure through image matching, but do not process the point cloud data, thus failing to eliminate environmental factors and dust interference. Therefore, a yard inventory method is needed that can account for environmental factors and avoid dust interference, while maintaining high-precision inventory capabilities and minimizing the need for stacker / reclaimer operating time. Summary of the Invention

[0003] Therefore, the technical problem to be solved by the present invention is that inventorying will occupy the operating time of the stacker-reclaimer and will be easily affected by environmental factors and dust interference.

[0004] The above technical problems are solved by the following technical solutions:

[0005] The present invention proposes a method for inventorying a stockpile, which includes obtaining original point cloud data of the stockpile; filtering the original point cloud data to obtain filtered point cloud data; segmenting the filtered point cloud data to obtain target modeling point data and non-modeling point data; constructing a three-dimensional model of the stockpile in the stockpile based on the target modeling point data; and calculating the volume of the stockpile according to the three-dimensional model to complete the inventory.

[0006] In a preferred embodiment of the storage yard inventory method of the present invention: obtaining original point cloud data of the storage yard;

[0007] Filtering the original point cloud data to obtain filtered point cloud data;

[0008] Segmenting and processing the filtered point cloud data to obtain target modeling point data and non-modeling point data;

[0009] Constructing a three-dimensional model of the stockpile in the stockpile yard based on the target modeling point data;

[0010] The volume of the stockpile is calculated based on the three-dimensional model to complete the stockpile inventory.

[0011] In a preferred embodiment of the storage yard inventory method of the present invention: the filtering process of the original point cloud data includes:

[0012] Calculating the average distance between adjacent laser points in the original point cloud data;

[0013] determining an adaptive screening radius based on the average distance;

[0014] Obtaining the total number of the laser points within the screening radius;

[0015] Laser points whose total number is less than the preset threshold are removed as noise points.

[0016] In a preferred embodiment of the storage yard inventory method of the present invention: calculating the average distance between adjacent laser points in the original point cloud data includes:

[0017] Calculate distance: Use the Euclidean distance formula to calculate the distance between the i-th laser point and the j-th laser point in three-dimensional space;

[0018] Calculate the average distance: Calculate the distance between the i-th laser point and its k nearest neighbor laser points, add them together, and then divide by k to get the average distance of the local area around the i-th laser point.

[0019] In a preferred embodiment of the storage yard inventory method of the present invention, determining an adaptive screening radius based on the average distance includes:

[0020] Calculating a sparse coefficient of the original point cloud data based on an average distance between adjacent laser points;

[0021] The screening radius is calculated using the sparse coefficient of the original point cloud data.

[0022] In a preferred embodiment of the storage yard inventory method of the present invention: the sparse coefficient of the original point cloud data is calculated based on the average distance between adjacent laser points, including:

[0023] Calculate the local density of each laser point: sum the distances between the i-th laser point and all other laser points. The smaller the sum, the higher the local density.

[0024] Calculate the normalized local density of each laser point: divide the local density of each laser point by the average distance between it and its k neighboring points;

[0025] Calculate the sparse coefficient of the point cloud: sum the normalized local densities of all laser points and then divide it by the total number of laser points i to obtain the average normalized local density of the entire original point cloud data, that is, the sparse coefficient.

[0026] In a preferred embodiment of the storage yard inventory method of the present invention: the sparse coefficient of the original point cloud data is used to calculate the screening radius, including:

[0027] Calculating the standard deviation: calculating the standard deviation of the average distance to quantify the degree of dispersion of the local distance distribution of the original point cloud data;

[0028] Calculate the screening radius: Add an adjustment based on the standard deviation to the average distance to obtain the screening radius of the laser point.

[0029] In a preferred embodiment of the storage yard inventory method of the present invention: segmenting and processing the filtered point cloud data includes:

[0030] Obtaining the reflectivity of each laser point in the filtered point cloud data;

[0031] dividing the filtered point cloud data into a first group and a second group according to a reflectivity threshold;

[0032] Calculate the proportion of laser points with different reflectivity in the first group and the second group;

[0033] The segmentation function is constructed based on the proportion of filtered point cloud data under different reflectivity.

[0034] In a preferred embodiment of the storage yard inventory method of the present invention: calculating the proportion of laser points with different reflectivity includes:

[0035] Calculate the proportion of laser points with different reflectivity in the first group, sum the number of all laser points with reflectivity ranging from 1 to k, and then divide it by the total number of laser points to obtain the proportion of laser points with different reflectivity in the first group;

[0036] The proportion of laser points with different reflectivities in the second group is calculated. The number of all laser points with reflectivities from k+1 to L is summed up and then divided by the total number of laser points to obtain the proportion of laser points with different reflectivities in the second group.

[0037] In a preferred embodiment of the storage yard inventory method of the present invention: constructing a segmentation function, including:

[0038] Measures the difference between two sets of data after segmentation by reflectivity threshold k;

[0039] Adjust the reflectivity threshold k so that the two groups of filtered point cloud data after segmentation have the greatest difference in reflectivity.

[0040] The present invention also provides a storage yard disk library system, which includes a scanner and a server.

[0041] In a preferred embodiment of the stockpile inventory system of the present invention: there are a plurality of scanners evenly arranged on the road, for acquiring original point cloud data of the stockpile; a server is connected to the scanners, for receiving the original point cloud data and obtaining modeling point data through filtering and segmentation processing, and constructing a three-dimensional model of the stockpile in the stockpile through the modeling point data, and then calculating the volume of the stockpile based on the three-dimensional model to complete the inventory.

[0042] The beneficial effects of the present invention are that it can remove the effects of environmental factors and dust interference on point cloud data, achieving optimal point cloud denoising, thereby enabling the creation of a more accurate three-dimensional model and improving the accuracy of inventory. Furthermore, the scanner of the present invention is installed on the horseway at the top of the greenhouse, which can solve the problem of the inventory function of the existing bucket wheel stacker reclaimer remote control system occupying the bucket wheel stacker reclaimer's operating time. It can achieve real-time modeling and inventory, with a faster refresh rate, improving the speed of inventory, maximizing the scanning accuracy and stability of the scanner, and achieving full coverage of the scanner yard. At the same time, there is a certain overlap between the scanners, further improving the fault tolerance of the modeling effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings of the embodiments of the present invention. Obviously, the drawings described below only relate to some embodiments of the present invention, and are not intended to limit the present invention.

[0044] Figure 1 The overall process diagram of the yard inventory method is shown;

[0045] Figure 2 A front view of a shed for a yard storage system is shown;

[0046] Figure 3 A top view of a shed for a yard storage system is shown;

[0047] Figure 4 A schematic diagram showing the overall structure of the storage yard disk storage system. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the present invention, the present invention is further described in detail below with reference to specific embodiments and the accompanying drawings.

[0049] The terms used in the present invention are those commonly used in the art in view of the functions of the present invention, but these terms may vary according to the intentions of those skilled in the art, precedents, or new technologies in the art. In addition, specific terms may be selected by the applicant, and in such cases, their detailed meanings will be described in the detailed description of the present invention. Therefore, the terms used in the specification should not be understood as simple names, but rather as the meanings of the terms and the overall description of the present invention.

[0050] Example 1, with reference to Figure 1 , as one embodiment of the present invention, provides a yard inventory method, comprising:

[0051] S100: Obtaining the original point cloud data of the storage yard;

[0052] S200: filtering the original point cloud data to obtain filtered point cloud data;

[0053] S300: Segmenting and processing the filtered point cloud data to obtain target modeling point data and non-modeling point data;

[0054] S400: Constructing a three-dimensional model of the stockpile in the yard based on the target modeling point data;

[0055] S500: Calculate the volume of the stockpile based on the three-dimensional model and complete the inventory.

[0056] It should be noted that in real-world stockpile environments, the shape and volume of piled materials, as well as environmental factors such as lighting, weather, and dust, are complex and highly variable. These factors place high demands on the efficiency and accuracy of stockpile inventory. Traditional stockpile inventory methods require the acquisition equipment for 3D stockpile modeling to be installed on the stacker / reclaimer, making it highly dependent on the stacker / reclaimer and taking up its operating time. This means that during inventory operations, the stacker / reclaimer cannot perform normal material handling tasks, reducing operational efficiency. Without a covered greenhouse, the acquisition equipment for 3D stockpile modeling is susceptible to weather conditions. Rainy or windy days with high dust levels can obstruct the laser scanner, affecting the accuracy of the 3D laser modeling. Interference from environmental factors can lead to inaccurate scan data, which in turn affects inventory accuracy, resulting in suboptimal modeling results and an inability to guarantee round-the-clock unmanned automated material handling and inventory operations. This significantly limits the practical application of traditional stockpile inventory methods.

[0057] Therefore, to address the above-mentioned problems of stockpile inventory efficiency and accuracy, the original point cloud data of the stockpile is collected through steps S100-S500, and more accurate target modeling point data is obtained through filtering and segmentation processing, thereby constructing an accurate three-dimensional model, obtaining the volume of the stockpile, and completing the inventory.

[0058] Example 2, reference Figure 1 , which is an embodiment of the present invention, provides a yard inventory method based on the first embodiment.

[0059] In the embodiment of the present application, the original point cloud data of the stockpile is obtained in step S100 by scanning the stockpile in the stockpile with a scanner installed on the top of the greenhouse, thereby obtaining the original point cloud data to facilitate subsequent data processing.

[0060] In an optional embodiment, step S100 can be performed by using a 3D laser scanner mounted on a mobile device, including but not limited to a mobile robot, to obtain raw point cloud data of the storage yard. The mobile device can be fixed on a horse path on the top of the greenhouse or can be moved on the horse path on the top of the greenhouse.

[0061] In another alternative embodiment, step S100 can be performed using a 3D laser scanner mounted on a mobile device, including but not limited to a drone, to acquire the raw point cloud data of the storage yard. The mobile device can fly over the top of the storage yard, providing a more flexible scanning method when a greenhouse is not available.

[0062] In the embodiment of the present application, step S200 processes the original point cloud data by filtering to obtain filtered point cloud data, including the following steps A1-A4:

[0063] A1: Calculate the average distance between adjacent laser points in the original point cloud data;

[0064] A1:2 determines the adaptive screening radius based on the average distance;

[0065] A1:3 obtains the total number of laser points within the screening radius;

[0066] A1:4 removes laser points whose total number is less than the preset threshold as noise points.

[0067] Specifically, the average distance between adjacent laser points is calculated as follows:

[0068]

[0069] Where x i Indicates the coordinate of the i-th laser point on the x-axis, x j Indicates the coordinate of the jth laser point on the x-axis, y i Indicates the coordinate of the i-th laser point on the y-axis, y j represents the coordinate of the jth laser point on the y-axis, z i Indicates the coordinate of the i-th laser point on the z axis, z j represents the coordinate of the jth laser point on the z-axis, represents the average distance between the i-th laser point and its k adjacent laser points.

[0070] The i-th point has k adjacent laser points, which are filtered based on the calculation results. If the number of adjacent laser points within the screening radius of the i-th point is ≥ 0.6k, the point is retained; if it is < 0.6k, it is considered a noise point and removed.

[0071] Specifically, determining the screening radius based on the average distance between adjacent laser points includes calculating a sparse coefficient of the original point cloud data based on the average distance between adjacent laser points and calculating the screening radius using the sparse coefficient of the original point cloud data.

[0072] Specifically, the sparse coefficient of the original point cloud data is calculated based on the average distance between adjacent laser points. The sparse coefficient calculation formula of the point cloud is:

[0073]

[0074] in, represents the proportionality coefficient, d ij represents the distance between the i-th laser point and the j-th adjacent laser point, represents the distance between the i-th laser point and the k adjacent laser points, n represents the number of all laser points, Represents the sparse coefficient of the point cloud.

[0075] That is, first calculate the local density of each laser point: sum the distances between the i-th laser point and all other laser points. The smaller the sum, the higher the local density.

[0076] Then calculate the normalized local density of each laser point: divide the local density of each laser point by the average distance between it and its k neighboring points;

[0077] Finally, the sparse coefficient of the calculated point cloud is obtained: the normalized local density of all laser points is summed up, and then divided by the total number of laser points i to obtain the average normalized local density of the entire original point cloud data, that is, the sparse coefficient

[0078] The screening radius is calculated using the sparse coefficient of the original point cloud data, where the calculation formula for the screening radius is:

[0079]

[0080] Where σ is the standard deviation, r is the screening radius, represents the average distance between the i-th laser point and its k adjacent laser points.

[0081] Get the total number of laser points within the screening radius;

[0082] Laser points whose total number is less than the preset threshold are removed as noise points.

[0083] Traditional filtering algorithms estimate the denoising radius through multiple manual experiments. However, this method is influenced by subjective consciousness and therefore has poor denoising effects. The present invention can adaptively determine the filtering radius by calculating the sparsity coefficient of each laser point. In this way, the processing parameters can be dynamically adjusted according to the local density of the point cloud to achieve the best point cloud denoising effect.

[0084] In an optional implementation, step S200 may be implemented by following steps B1-B5 to further optimize the filtering effect:

[0085] B1: Calculate the curvature of each laser point in the original point cloud data;

[0086] B2: Segment the original point cloud data into planar and non-planar regions based on curvature;

[0087] B3: Filter the point cloud data in the plane area using a distance-based filtering method;

[0088] B4: For point cloud data in non-planar areas, a statistical filtering method is used. For example, the standard deviation of the distance within the neighborhood of each point is calculated, and then points with a standard deviation greater than a threshold are removed.

[0089] B5: Merge the filtered point cloud data of the planar area and the non-planar area to obtain filtered point cloud data.

[0090] In the embodiment of the present application, the segmentation and processing of the filtered point cloud data in step S300 to obtain target modeling point data and non-modeling point data includes the following steps C1-C4:

[0091] C.1: Obtain the reflectivity of each laser point in the filtered point cloud data;

[0092] C.2: Divide the filtered point cloud data into the first and second groups based on the reflectivity threshold;

[0093] C.3: Calculate the proportion of laser spots with different reflectivity in the first and second groups;

[0094] C.4: Construct a segmentation function based on the proportion of filtered point cloud data under different reflectivity.

[0095] Specifically, in C3, the calculation formula for the proportion of laser points with different reflectivity is:

[0096]

[0097] Where k represents the reflectivity threshold, n irepresents the number of laser points with reflectivity i, N represents the total number of laser points, ω0 represents the proportion of laser point cloud data under different reflectivities in the first group, ω1 represents the proportion of laser point cloud data under different reflectivities in the second group, and L represents the level of reflectivity.

[0098] Specifically, in C4, the segmentation function is constructed based on the proportion of filtered point cloud data under different reflectivity:

[0099] A segmentation function is constructed based on the mean reflectivity of the laser point cloud data in the first and second groups and the proportion of the laser point cloud data at different reflectivities, including:

[0100] Using the formula:

[0101]

[0102] Construct a partitioning function; where f max represents the objective function, μ0 represents the mean reflectivity in the first set of laser point cloud data, and μ1 represents the mean reflectivity in the second set of laser point cloud data.

[0103] Then, the target modeling points and non-modeling points are obtained by continuously adjusting the reflectivity threshold to maximize the segmentation function, because the reflectivity of the laser point cloud is a measure of the ratio of the laser signal intensity relative to the laser beam irradiating the surface of the object and returning to the sensor to the original laser signal intensity.

[0104] Reflectivity can be used to characterize how well different surfaces reflect laser light. In the storage yard field, this method uses the mean and proportion of reflectivity in laser point cloud data to segment the data. This eliminates points unrelated to the modeling target, such as chandeliers and transport mechanisms, and retains only laser points directly related to the cargo or stacking, helping to generate more accurate 3D models.

[0105] In this embodiment, the material is primarily dry mineral powder, stored in a closed greenhouse. Therefore, the partitioning function results are optimized to have a target modeling point ratio of ≥ 70% and non-target modeling points of < 30%. The reflectivity is relatively low. A reflectivity threshold k of 70 is used to separate the mineral powder from influencing factors such as metal impurities and pendant lights.

[0106] In an embodiment of the present application, in step S400, a three-dimensional model of the stockpile in the yard is constructed based on the target modeling point data. Software including but not limited to Blender, SketchUp, 3dsMax, Unity, and Maya can be used to generate a realistic three-dimensional model based on the point cloud data to accurately reflect the shape and spatial distribution of the stockpile.

[0107] In this embodiment of the present application, in step S500, the volume of the stockpile is calculated based on the 3D model, completing the inventory. Once the 3D model is constructed, the volume of the stockpile can be calculated using modeling software. The calculation results can be presented digitally or visually on a display terminal, thus completing the inventory process.

[0108] Example 3, reference Figure 2 , as an embodiment of the present invention, provides a yard disk library system, including a scanner 1 and a server 6.

[0109] It should be noted that existing remote control systems for bucket-wheel stackers and reclaimers have several issues with achieving yard inventory. Both 2D and 3D laser scanners typically require installation on the bucket-wheel stacker and reclaimer, which occupies operating time. Furthermore, without a covered greenhouse, laser scanners are susceptible to weather conditions. Rainy or windy days with high dust levels can obstruct the laser scanner, affecting the accuracy of 3D laser modeling and preventing round-the-clock unmanned automated material handling and inventory management.

[0110] Specifically, Scanner 1 is installed on the bridleway 2 atop the greenhouse. Scanner 1 is a 3D laser scanner used to acquire raw point cloud data of the stockpile. Bridleway 2 is a passageway or platform atop the greenhouse, serving as the mounting platform for Scanner 1. Its vertical projection on the ground flanks the material runway 4, the stacker-reclaimer track, which runs along the centerline of the greenhouse's interior. The specific number of scanners is determined based on the distribution of the stockpile and the scanner's coverage area to ensure comprehensive scanning of the entire stockpile. These scanners collect raw point cloud data of the stockpile within the yard, providing the data foundation for subsequent 3D modeling and inventory management.

[0111] The scanner 1 and the power distribution network box 3 are paired one by one, and the scanner 1 transmits the original point cloud data to the central control building scanner network aggregation switch 5 through the power supply cable and the communication optical cable. The three-dimensional modeling server 6 sets the automatic scanning frequency of the three-dimensional laser scanner through the background management software to perform scheduled automatic scanning.

[0112] After the scanning is completed, the computer modeling software in the 3D modeling server 6 performs modeling processing on the point cloud data obtained by the 3D laser scanner.

[0113] Generate a 3D image of the stockpile and calculate the volume of the stockpile to conduct a stockpile inventory.

[0114] Finally, it should be pointed out that the methods and devices described in detail above are merely embodiments, and those skilled in the art can modify these embodiments in different ways without departing from the scope of the present invention.

Claims

1. A method for inventorying a storage yard, characterized by: include, Obtain the original point cloud data of the yard; Filtering the original point cloud data to obtain filtered point cloud data; Segmenting and processing the filtered point cloud data to obtain target modeling point data and non-modeling point data; Constructing a three-dimensional model of the stockpile in the stockpile yard based on the target modeling point data; The volume of the stockpile is calculated based on the three-dimensional model to complete the stockpile inventory.

2. The method for inventorying in a storage yard according to claim 1, characterized in that: The filtering processes the original point cloud data, include, Calculating the average distance between adjacent laser points in the original point cloud data; determining an adaptive screening radius based on the average distance; Obtaining the total number of the laser points within the screening radius; Laser points whose total number is less than the preset threshold are removed as noise points.

3. The method for inventorying in a storage yard according to claim 2, characterized in that: Calculate the average distance between adjacent laser points in the original point cloud data, include, Calculate distance: Use the Euclidean distance formula to calculate the distance between the i-th laser point and the j-th laser point in three-dimensional space; Calculate the average distance: Calculate the distance between the i-th laser point and its k nearest neighbor laser points, add them together, and then divide by k to get the average distance of the local area around the i-th laser point.

4. The method for inventorying in a storage yard according to claim 3, characterized in that: Determining an adaptive screening radius based on the average distance includes, Calculating a sparse coefficient of the original point cloud data based on an average distance between adjacent laser points; The screening radius is calculated using the sparse coefficient of the original point cloud data.

5. The method for inventorying in a storage yard according to claim 4, characterized in that: Calculate the sparse coefficient of the original point cloud data based on the average distance between adjacent laser points, include, Calculate the local density of each laser point: sum the distances between the i-th laser point and all other laser points. The smaller the sum, the higher the local density. Calculate the normalized local density of each laser point: divide the local density of each laser point by the average distance between it and its k neighboring points; Calculate the sparse coefficient of the point cloud: sum the normalized local densities of all laser points and then divide it by the total number of laser points i to obtain the average normalized local density of the entire original point cloud data, that is, the sparse coefficient.

6. The method for inventorying in a storage yard according to claim 5, characterized in that: The screening radius is calculated using the sparse coefficient of the original point cloud data, including: Calculating the standard deviation: calculating the standard deviation of the average distance to quantify the degree of dispersion of the local distance distribution of the original point cloud data; Calculate the screening radius: Add an adjustment based on the standard deviation to the average distance to obtain the screening radius of the laser point.

7. The method for inventorying in a storage yard according to claim 6, characterized in that: Segmenting the filtered point cloud data, including: Obtaining the reflectivity of each laser point in the filtered point cloud data; dividing the filtered point cloud data into a first group and a second group according to a reflectivity threshold; Calculate the proportion of laser points with different reflectivity in the first group and the second group; The segmentation function is constructed based on the proportion of filtered point cloud data under different reflectivity.

8. The method for inventorying in a storage yard according to claim 7, characterized in that: Calculate the proportion of laser points with different reflectivity, including: Calculate the proportion of laser points with different reflectivity in the first group, sum the number of all laser points with reflectivity ranging from 1 to k, and then divide it by the total number of laser points to obtain the proportion of laser points with different reflectivity in the first group; The proportion of laser points with different reflectivities in the second group is calculated. The number of all laser points with reflectivities from k+1 to L is summed up and then divided by the total number of laser points to obtain the proportion of laser points with different reflectivities in the second group.

9. The method for inventorying in a storage yard according to claim 8, characterized in that: Construct a split function, including, Measures the difference between two sets of data after segmentation by reflectivity threshold k; Adjust the reflectivity threshold k so that the two groups of filtered point cloud data after segmentation have the greatest difference in reflectivity.

10. A yard inventory system, characterized by: include, A plurality of scanners (1) are evenly arranged on the horse path (2) for acquiring original point cloud data of the yard; The server (6) is connected to the scanner (1) and is used to receive the original point cloud data and obtain modeling point data through filtering and segmentation processing, and to construct a three-dimensional model of the material pile in the yard through the modeling point data, and then calculate the volume of the material pile according to the three-dimensional model to complete the inventory.

Images