Method and apparatus for rendering three-dimensional stockpile in real time, device, and medium

Abstract

The present disclosure relates to the field of production safety monitoring, and provides a method and apparatus for rendering a three-dimensional stockpile in real time, a device, and a medium. The method comprises: acquiring three-dimensional point cloud data of a stockpile to be detected; on the basis of asymmetric data in the three-dimensional point cloud data, layering the three-dimensional point cloud data to obtain a plurality of layers of planar layered data and a hierarchical mapping relationship of each layer of planar layered data; determining a convex hull corresponding to each layer of planar layered data; on the basis of the convex hull, determining a local convex hull of a corresponding layer of planar layered data; and eliminating the data of the planar layered data located outside the corresponding local convex hull to obtain valid stockpile data, and rendering the stockpile in real time on the basis of the valid stockpile data.

Description

A method, device, equipment and medium for real-time rendering of three-dimensional material piles

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] The present disclosure claims priority to Chinese patent application CN202311628211.5 filed on November 30, 2023, entitled “A method, device, equipment and medium for real-time rendering of three-dimensional material piles,” the entire contents of which are incorporated by reference into the present disclosure. Technical Field

[0003] The present disclosure relates to the field of safe production monitoring, and in particular to a method, device, equipment and medium for real-time rendering of a three-dimensional material pile. Background Art

[0004] Stacking and reclaiming equipment is the main large-scale equipment used in bulk cargo terminals for transshipping coal. Currently, during stacking and reclaiming operations, the real-time changes of the pile are mostly fed back through planar data. There are problems such as the pile point cloud data being too flat and the monitoring of pile deformation being unintuitive. Therefore, there are still many safety hazards in monitoring the real-time situation of the pile.

[0005] In order to ensure the safe operation of material handling operations, how to achieve real-time three-dimensional status monitoring of the material pile and enable real-time feedback of the real-time changes of the material pile in the three-dimensional monitoring system is a technical problem that needs to be solved urgently to realize the smart port.

[0006] Summary of the Invention

[0007] In view of this, the embodiments of the present disclosure provide a method, device, equipment and medium for real-time rendering of a three-dimensional stockpile to solve the technical problem that related technologies are difficult to accurately measure changes in the liquid level height of a circulation tank.

[0008] According to the first aspect, an embodiment of the present disclosure provides a method for real-time rendering of a three-dimensional material pile, the method comprising: obtaining three-dimensional point cloud data of the material pile to be detected; layering the three-dimensional point cloud data based on asymmetric data in the three-dimensional point cloud data to obtain several layers of plane layered data and a layered mapping relationship of each layer of plane layered data; the layered mapping relationship is a mapping relationship between the three-dimensional point cloud data and the plane layered data; determining the convex hull corresponding to each layer of plane layered data; determining the local convex hull of the corresponding layer of plane layered data based on the convex hull; eliminating the data of the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and rendering the material pile in real time based on the effective material pile data.

[0009] In combination with the first aspect, in the first embodiment of the first aspect, determining the convex hull corresponding to each layer of plane layered data further includes: determining the layered projection data with the smallest vertical coordinate value within a preset area, and determining the determined layered projection data as a reference point; determining the amplitude angle between the layered projection data and the reference point, sorting the layered projection data in ascending order based on the amplitude angle, and determining the sorting value corresponding to the layered projection data; creating an empty stack, and pushing the reference point and the layered projection data corresponding to the minimum sorting value into the stack in sequence; determining the connecting straight line between the layered projection data corresponding to the stack vertex and the adjacent stack elements, and determining the positional relationship between the layered projection data and the connecting straight line in the order of sorting; determining that the layered projection data is located on the left side of the connecting straight line, pushing the layered projection data into the stack, and updating the stack vertex and the connecting straight line; determining that the layered projection data is located on the right side of the connecting straight line, popping the stack vertex, and updating the stack vertex and the connecting straight line; traversing all layered projection data with sorting values, determining the elements in the stack, connecting the layered projection data corresponding to the adjacent elements, and obtaining the convex hull corresponding to the first projection data.

[0010] In combination with the first embodiment of the first aspect, in the second embodiment of the first aspect, determining the local convex hull of the corresponding layer plane layered data based on the convex hull further includes: determining the plane layered data located inside the convex hull; determining the vertical distances corresponding to the plane layered data and each edge of the convex hull, and based on the vertical distances, determining the edge closest to the plane layered data, and taking the closest edge as a candidate folding edge; connecting the two vertices of the plane layered data and the candidate folding edge to obtain a candidate straight line; determining the angle between the candidate straight line and the candidate folding edge; determining that the angles corresponding to the plane layered data are all acute angles, taking the plane layered data as a candidate connection point, and classifying the candidate connection point into the corresponding edge of the convex hull; disconnecting the two vertices on the edge corresponding to the candidate connection point, and connecting the two vertices to the candidate connection point respectively to obtain a local convex hull.

[0011] In combination with the first aspect, in the third embodiment of the first aspect, the data of the plane layered data located outside the corresponding local convex hull is eliminated to obtain effective material pile data, and the material pile is rendered in real time based on the effective material pile data, further including: determining the external projection data located outside the local convex hull and the internal projection data located inside the local convex hull; based on the layered projection relationship, eliminating the three-dimensional point cloud data corresponding to the external projection data to obtain effective material pile data, and rendering the material pile in real time based on the effective material pile data.

[0012] In combination with the third embodiment of the first aspect, in the fourth embodiment of the first aspect, the internal projection data is obtained by the following steps: determining the coordinate information of the vertices of the local convex hull, and based on the coordinate information, determining the extreme vertex of the local convex hull; determining the horizontal scan line that has an intersection with the local convex hull, determining the intersection corresponding to the horizontal scan line and the local convex hull and determining whether the horizontal scan line passes through the extreme vertex of the local convex hull; determining that the horizontal scan line passes through the extreme vertex, and dividing the extreme vertex into two extreme points based on the edge of the local convex hull that generates the extreme vertex; pairing the intersection points from left to right, determining the pairing points and connecting the pairing points with straight lines, eliminating the external projection data located on the horizontal scan line and not on the line connecting the pairing points, and obtaining the internal projection data.

[0013] In combination with the first aspect, in the fifth embodiment of the first aspect, based on the asymmetric data in the three-dimensional point cloud data, the three-dimensional point cloud data is layered to obtain several layers of plane layered data and the layered mapping relationship of each layer of plane layered data, further including: determining the symmetric data and asymmetric data in the three-dimensional point cloud data; based on the asymmetric three-dimensional point cloud data and a preset height value, determining the minimum distance association point corresponding to the asymmetric three-dimensional point cloud data; based on the minimum distance association point, the three-dimensional point cloud data is layered to obtain several layers of plane layered data and the layered mapping relationship of each layer of plane layered data.

[0014] In combination with the fifth embodiment of the first aspect, in the sixth embodiment of the first aspect, based on the asymmetric three-dimensional point cloud data and the preset height value, determining the minimum distance associated point corresponding to the asymmetric three-dimensional point cloud data further includes: based on the asymmetric three-dimensional point cloud data and the preset height value, determining the upper neighborhood point set and the lower neighborhood point set; finding the minimum distance of each point in the upper neighborhood point set in the lower neighborhood point set, and finding the minimum distance of each point in the lower neighborhood point set in the upper neighborhood point set; determining that the distances between the two are equal, and taking the point with the smaller value as the minimum distance associated point.

[0015] According to the second aspect, an embodiment of the present disclosure provides a device for real-time rendering of a three-dimensional material pile, the device including: an acquisition module, configured to acquire three-dimensional point cloud data of the material pile to be detected; a layering module, configured to layer the three-dimensional point cloud data based on the asymmetric data in the three-dimensional point cloud data, to obtain several layers of plane layered data and a layered mapping relationship of each layer of plane layered data; the layered mapping relationship is a mapping relationship between the three-dimensional point cloud data and the plane layered data; a first determination module, configured to determine the convex hull corresponding to each layer of plane layered data; a second determination module, configured to determine the local convex hull of the corresponding layer of plane layered data based on the convex hull; a real-time rendering module, configured to eliminate the data of the plane layered data located outside the corresponding local convex hull, to obtain effective material pile data, and to render the material pile in real time based on the effective material pile data.

[0016] According to the third aspect, an embodiment of the present disclosure provides an electronic device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, computer instructions stored in the memory, and the processor executing the computer instructions to execute the method for real-time rendering of a three-dimensional material pile as described in the first aspect or any preferred embodiment of the first aspect.

[0017] According to a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for real-time rendering of a three-dimensional material pile as described in the first aspect or any preferred embodiment of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The features and advantages of the present disclosure will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the present disclosure in any way. In the accompanying drawings:

[0019] FIG1 is a schematic flow chart of a method for real-time rendering of a three-dimensional stockpile according to an embodiment of the present disclosure;

[0020] FIG2 is a schematic diagram of obtaining a local convex hull based on a convex hull in a method for real-time rendering of a three-dimensional stockpile according to an embodiment of the present disclosure;

[0021] FIG3 is a flow chart of a convex hull algorithm in a method for real-time rendering of a three-dimensional stockpile according to an embodiment of the present disclosure;

[0022] FIG4 is a schematic structural diagram of an apparatus for real-time rendering of a three-dimensional stockpile according to an embodiment of the present disclosure;

[0023] FIG5 is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] To make the purpose, technical solutions, and advantages of the embodiments of the present disclosure more clear, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present disclosure.

[0025] The coal industry can have a wide-ranging impact on multiple industries, such as metallurgy, chemical industry, and electricity. After continuous construction, the port's specialized stacking and reclaiming equipment and production monitoring capabilities have been greatly improved, ensuring the transportation capacity of related sea transportation and laying the foundation for the corresponding supply of coal.

[0026] Among them, coal is stored in the form of piles in the port yard. Through the collected point cloud data of the coal pile, the changes in the pile can be rendered in real time, and the real-time changes in the pile can be monitored, which is also an important part of achieving safe and intelligent production.

[0027] Stacking and reclaiming equipment is the main large-scale equipment used in bulk cargo terminals for transshipping coal. Currently, during stacking and reclaiming operations, the real-time changes of the pile are mostly fed back through planar data. There are problems such as the pile point cloud data being too flat and the monitoring of pile deformation being unintuitive. Therefore, there are still many safety hazards in monitoring the real-time situation of the pile.

[0028] That is, in most current port monitoring equipment, the changes in the stockpile are only reflected through plane data. Therefore, in order to ensure the safe operation of the material handling operation, how to realize real-time three-dimensional status monitoring of the stockpile and make real-time feedback of the real-time changes of the stockpile in the three-dimensional monitoring system is an important factor in realizing the smart port.

[0029] In order to solve the above problems, a method, device, equipment and medium for real-time rendering of a three-dimensional material pile are provided in the present embodiment. The main solution of the embodiment of the present disclosure is: obtaining three-dimensional point cloud data of the material pile to be detected; based on the asymmetric data in the three-dimensional point cloud data, the three-dimensional point cloud data is layered to obtain several layers of plane layered data and a layered mapping relationship of each layer of plane layered data; the layered mapping relationship is a mapping relationship between the three-dimensional point cloud data and the plane layered data; determining the convex hull corresponding to each layer of plane layered data; determining the local convex hull of the corresponding layer of plane layered data based on the convex hull; eliminating the data of the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and rendering the material pile in real time based on the effective material pile data.

[0030] In the embodiment of the present disclosure, asymmetric data in the three-dimensional point cloud data is obtained, the three-dimensional point cloud data is layered, and several layers of plane layered data and a layered mapping relationship of each layer of plane layered data are obtained. Then, the convex hull corresponding to each layer of plane layered data is determined, and the local convex hull of the corresponding layer of plane layered data is determined based on the convex hull. Finally, the data of the plane layered data located outside the corresponding local convex hull is eliminated to obtain effective material pile data, and the material pile is rendered in real time based on the effective material pile data. The collected three-dimensional point cloud data structure is layered, and the convex hull algorithm is called to refine each layer of point cloud data. The convex hull algorithm is used to process the three-dimensional point cloud data. The generated three-dimensional model is more accurate and faster, which is convenient for subsequent rendering, so that the material pile rendering effect is better, ensuring the safe operation of the material picking operation, and the real-time changes of the material pile are fed back in real time in the three-dimensional monitoring system.

[0031] The following will describe in detail the technical solutions of the embodiments of the present disclosure and how the technical solutions of the embodiments of the present disclosure solve the above-mentioned technical problems using specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The following will describe the embodiments of the present disclosure in detail with reference to specific drawings.

[0032] Example 1

[0033] A method for real-time rendering of a three-dimensional material pile is provided in this embodiment. It can be understood that the method for real-time rendering of a three-dimensional material pile in the embodiment of the present disclosure can be used in electronic devices, including but not limited to computers, mobile terminals, etc. Figure 1 is a flow chart of the method for real-time rendering of a three-dimensional material pile according to the embodiment of the present disclosure. As shown in Figure 1, the method includes the following steps S10 to S50.

[0034] S10. Acquire three-dimensional point cloud data of the material pile to be inspected.

[0035] The 3D point cloud data can be pre-stored in the electronic device or acquired externally. For example, the electronic device can acquire the data from an external acquisition device, or extract the data from externally acquired data. The specific acquisition method of the 3D point cloud data is not restricted; it only requires that the electronic device be able to acquire the 3D point cloud data.

[0036] S20. Based on the asymmetric data in the three-dimensional point cloud data, the three-dimensional point cloud data is layered to obtain several layers of plane layered data and a layered mapping relationship between each layer of plane layered data; the layered mapping relationship is a mapping relationship between the three-dimensional point cloud data and the plane layered data.

[0037] Layered processing can reduce the dimensionality of three-dimensional data to obtain corresponding two-dimensional data, and establish a mapping relationship between three-dimensional data points and two-dimensional projection data points. In this method, the three-dimensional point cloud data is composed of several three-dimensional data points, and the layered projection data is composed of several layered projection data points. There are at least two three-dimensional layer data points corresponding to the same layered projection data point, that is, each three-dimensional layer data point corresponds to a layered projection data point, and a layered projection data point may correspond to multiple three-dimensional data points.

[0038] S30. Determine the convex hull corresponding to each layer of planar layered data. For a two-dimensional plane, the convex hull is the minimum convex polygon enclosing all layered projection data points. In one exemplary embodiment, given a set S of layered projection data points, the intersection T of all convex sets that contain set S is called the convex hull of set S. In the present disclosure, the convex hull of the first projection data can be determined based on methods such as Graham scanning, exhaustive enumeration, divide-and-conquer, Jarvis stepping, and Melkman algorithms. Again, there is no limitation on the specific convex hull algorithm.

[0039] S40 , determining the local convex hull of the corresponding layer plane layered data based on the convex hull.

[0040] For a point set Q containing N points and a small convex polygon K, there are usually three cases where every point in Q is contained, either inside K or on the boundary of K. Convex hull algorithms have important applications in imaging, geospatial computing, computer graphics, and other fields.

[0041] The convex polygon obtained by the convex hull algorithm is a subset of the point set and can be considered an approximation of the contour line of the object represented by the point set. However, there are still discrete points on the object's surface that are not included in the polyline. Therefore, these points need to be classified onto the current polyline. To more accurately represent the contour line, we first need to determine which edge within the polyline each point belongs to.

[0042] For each undetermined point Q, the relationship between the projection and distance between the point and each line segment of the current contour line needs to be calculated. Each undetermined point can only be classified into two line segments with an angle less than 90 degrees and the closest distance.

[0043] After classifying the current contour line and pending points using the above method, recalculate the convex hull of each segment of the polyline and the pending points classified into that segment to obtain the convex hull of the local area. All local convex hulls are then disconnected at the original contour line segment to obtain open polyline segments. Connect the previous and next polyline segments in sequence to obtain the next level of polyline.

[0044] If you're studying a layer of point cloud data, the first application of the convex hull algorithm yields a convex polygon. Continuing with the convex hull algorithm, you can further obtain a partially convex polygon. If you then integrate the convex polygons, you'll obtain a concave polygon composed of contour lines, rather than a convex polygon. Next, classify the remaining points in the concave polygon and apply the convex hull algorithm again, still yielding a concave polygon. In other words, only the first application of the convex hull algorithm yields a convex polygon. After each subsequent iteration, the resulting merging results in a concave polygon, no longer a convex polygon. After the convex hull algorithm completes, each remaining point maintains the same topological relationship with the current polyline and can be either inside or outside the polyline. If the number of iterations is odd, the remaining points are outside the polyline; if the number of iterations is even, the remaining points are inside.

[0045] By further fitting the obtained polyline polygon and the remaining undetermined points, the polyline can be further fitted to obtain smooth contour lines. There are many polyline fitting methods, such as cubic polynomials and spline function fitting. The method used in this article is the spline function fitting method.

[0046] Contour lines generally exhibit a distribution of detail and can be used to represent surface features. Complex contour lines on a contour line indicate that certain areas have more features, while simple contour lines indicate that certain areas have fewer features. However, as discussed previously, the complexity of the contour line construction process increases with the number of iterations. Therefore, the number of convex hull iterations is a criterion for distinguishing contour lines. In other words, a lower number of iterations indicates a lower level of detail in the contour line. Therefore, the initial contour line obtained from the first iteration of the convex hull algorithm is considered the one with the lowest level of detail. This contour line also has the fewest features and lacks significant detail. These contour lines and the remaining undetermined points are then classified. After this classification, the convex hull algorithm is applied to these points for a second iteration, again yielding contour lines with the next lowest level of detail. This process continues in this manner, resulting in a series of contour lines with increasing levels of detail.

[0047] Assume that the number of iterations for constructing the highest level contour line obtained by the convex hull algorithm for a certain layer of points is k. Then, according to the actual level of detail required, the level is set to m, depending on the number of iterations corresponding to each level of detail. For example, when k = 7 and m = 4, it can be seen that the number of iterations for the first level contour line, that is, the lowest detail contour line, is 1, the number of iterations for the second level is 3, and so on. The number of iterations for the third level is 5, and the number of iterations for the fourth level, that is, the highest detail contour line, is 7.

[0048] During the process of constructing contour lines, when the number of iterations is from 1 to k, each time the calculation is completed, the contour line of that level has been calculated. Therefore, it only needs to be stored in the memory. After the contour line containing all points is calculated, the contour lines of other levels have also been completed. Therefore, it is only necessary to match the number of iterations with the required contour line detail level to obtain the contour lines with high to low details.

[0049] S50 , eliminating the plane layered data outside the corresponding local convex hull to obtain valid stockpile data, and performing real-time rendering on the stockpile based on the valid stockpile data.

[0050] Since point cloud data is often scattered and disordered, and contains a lot of redundant data, it is very difficult to remove redundant data and find the topological relationship between points during the 3D modeling process, especially when studying algorithms from a 3D perspective. However, algorithms within the 2D plane are relatively mature, and the data on a scan line is in the same plane. This leads to the idea of ​​converting scattered point cloud data into scan line data. After the elimination process, the noise data can be better eliminated. This noise data needs to be filtered out to obtain more realistic stockpile data. In this method, the convex hull algorithm and the scan line algorithm are used for elimination. Particles located outside the layered projection data in the simulated fabric are eliminated to obtain valid stockpile data. The 3D point cloud data is also directly layered, so the convex hull algorithm is called to further refine each layer of plane layered data.

[0051] The disclosed method for real-time rendering of a three-dimensional material pile obtains asymmetric data in three-dimensional point cloud data, layers the three-dimensional point cloud data, obtains several layers of plane layered data and a layered mapping relationship of each layer of plane layered data, then determines the convex hull corresponding to each layer of plane layered data, determines the local convex hull of the corresponding layer of plane layered data based on the convex hull, and finally eliminates the data of the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and renders the material pile in real time based on the effective material pile data, layers the collected three-dimensional point cloud data structure, calls the convex hull algorithm to refine each layer of point cloud data, and uses the convex hull algorithm to process the three-dimensional point cloud data. The generated three-dimensional model is more accurate and faster, which is convenient for subsequent rendering, so that the material pile rendering effect is better, and the safe operation of the material retrieving operation is guaranteed. The real-time changes of the material pile are fed back in real time in the three-dimensional monitoring system.

[0052] In this embodiment, step S20 further includes the following steps S21 to S23.

[0053] S21. Determine symmetric data and asymmetric data in the three-dimensional point cloud data. S22. Determine, based on the asymmetric three-dimensional point cloud data and a preset height value, a minimum distance associated point corresponding to the asymmetric three-dimensional point cloud data. S23. Layer the three-dimensional point cloud data based on the minimum distance associated point to obtain multiple layers of planar layered data and a layered mapping relationship for each layer of planar layered data.

[0054] Before stratification, point cloud data is scattered, but it can be divided into two types of data according to its axisymmetry about the coordinate axis. One type of data is asymmetric with respect to the coordinate axis, and the other type of data is exactly symmetric about the X, Y, and Z axes. The asymmetric data with respect to the coordinate axis can be directly selected on the screen in sequence through human-computer interaction with the mouse, and then these selected points are fitted or interpolated into a curve. This curve is the stratification axis.

[0055] In an exemplary embodiment, step S22 further includes the following steps S221 to S223 .

[0056] S221. Based on the asymmetric three-dimensional point cloud data and a preset height value, determine the upper and lower neighboring point sets. S222. Find the minimum distance between each point in the upper neighboring point set and the lower neighboring point set, and find the minimum distance between each point in the lower neighboring point set and the upper neighboring point set. S223. If the distances between the two are equal, the point with the smaller value is used as the minimum distance associated point.

[0057] Assume that the cross-sectional height of a certain point cloud data is H, and take the height of the relevant area as ±△H (the value is determined according to the density of the point cloud). If the cross-section is the middle cross-section, a narrow annular band will be formed. Define the upper neighborhood point set E1 as the points in the area [H-△H, H], and the lower neighborhood point set E2 as the points in the area [H, H+△H]. First, find the minimum distance N of each point M in the upper neighborhood point set E1 to the lower neighborhood point set E2, and let D1 = MN. Similarly, find the minimum distance T of point N in E1 in E2, and let D2 = NT. If T coincides with M, then the minimum distance associated points are M and N, and the point pair corresponding to the smaller value is recorded as the minimum distance point pair, and the point pair corresponding to the other value is a non-associated point. Repeat the above process, and after traversing the entire upper neighborhood point set, the minimum distance associated point pair set of this layer section is obtained. Suppose a certain associated point pair is M1=(x1,y1,z1), The equation of the line connecting the points is:

[0058] It is known that the height of the section is h, and let k = (h-z1) / (z2-z1), then we have The x and y values ​​of the intersection of the point pair line and the section are

[0059] In this embodiment, step S30 further includes the following steps S31 to S37.

[0060] S31 . Determine the plane layered data (point) with the smallest vertical coordinate value within the preset area, and determine the determined plane layered data as the reference point p0 .

[0061] In the present disclosure, the ground plane is divided into four areas of completely equal area, namely the upper left area, the lower left area, the upper right area and the lower right area, among which the lower left area is the preset area, and the plane layered data with the smallest vertical coordinate value in the preset area is determined, and the first stolen data point is determined as the reference point p0.

[0062] S32. Determine the argument between the plane layered data and the reference point p0, sort the plane layered data in ascending order based on the argument, and determine the sort value corresponding to the plane layered data.

[0063] In the present disclosure, the coordinates of the plane layered data can be translated, and the reference point p0 can be used as the origin. The plane layered data and the reference point p0 can be connected, and the angle α between each plane layered data relative to the reference point p0 can be calculated. The plane layered data can be sorted in the order of the angle α from small to large. When the angle α is the same, the plane layered data farthest from the reference point p0 is retained, and the other plane layered data with the same angle α are not sorted. The plane layered data sorted in order are p1, p2 to p n

[0064] S33. Create an empty stack, and push the reference point p0 and the plane layered data p1 corresponding to the minimum sorting value into the stack in sequence. When the number of elements in the stack is less than 2, the elements are pushed in directly, that is, the reference point p0 is pushed into the stack first, and then the plane layered data p1 is pushed into the stack.

[0065] S34. Determine the connecting line between the plane layered data corresponding to the stack vertex and the adjacent stack element, and determine the positional relationship between the plane layered data and the connecting line according to the sorting order. Specifically, first push p0 and p1 onto the stack. Since there are only two elements in the stack initially, the plane layered data corresponding to the stack vertex is p1, and the plane layered data corresponding to the stack element adjacent to the stack vertex is p0. Then, connect p1 and p0 to obtain the connecting line p1p0.

[0066] S35. Determine that the plane layered data is located to the left of the connecting line, push the plane layered data into the stack, and update the stack vertices and the connecting line.

[0067] Determine whether the plane layered data p2 adjacent to p1 is on the left or right of the connecting line p1p0. If it is on the left, p2 is pushed into the stack. After pushing, the stack vertex is p2, the stack element adjacent to the stack vertex is p1, and the connecting line is p2p1.

[0068] S36: Determine that the plane layered data is located to the right of the connecting line, pop the stack vertex, and update the stack vertex and the connecting line.

[0069] Assuming that the plane layered data p3 is on the right side of the connecting line p2p1, then the stack vertex p2 is popped out. After popping, the stack vertex is p1, the stack element adjacent to the stack vertex is p0, and the connecting line is p1p0. Next, the positional relationship between the plane layered data p3 and the connecting line p1p0 will be determined, and the stack push and pop operations will be performed.

[0070] S37. Traverse all the plane layered data with sorted values, determine the elements in the stack, and use the determined elements in the stack as vertices of the convex hull, connect the plane layered data corresponding to adjacent elements, and obtain the convex hull corresponding to the first projection data.

[0071] Traverse all flat hierarchical data p with sorted values i , get all the elements in the stack, these elements are also the vertices of the convex hull, and the convex hull can be determined by connecting two adjacent vertices counterclockwise.

[0072] Please refer to FIG. 2 and FIG. 3 . In this embodiment, step S40 further includes the following steps S41 to S46 .

[0073] S41. Determine the plane layered data located inside the convex hull; S42. Determine the vertical distances corresponding to the plane layered data and each edge of the convex hull, and based on the vertical distances, determine the edge closest to the plane layered data, and use the closest edge as a candidate folding edge; S43. Connect the two vertices of the plane layered data and the candidate folding edge to obtain a candidate straight line; S44. Determine the angle between the candidate straight line and the candidate folding edge; S45. Determine that the angles corresponding to the plane layered data are all acute angles, use the plane layered data as a candidate connection point, and classify the candidate connection point into the corresponding edge of the convex hull; S46. Disconnect the two vertices on the edge corresponding to the candidate connection point, and connect the two vertices to the candidate connection point respectively to obtain a local convex hull.

[0074] If you're studying a layer of point cloud data, the first application of the convex hull algorithm yields a convex polygon. Continuing with the convex hull algorithm, further local convex polygons can be obtained. Integrating the convex polygons yields a concave polygon composed of contour lines, rather than a convex polygon. Then, by classifying the remaining points in the concave polygon and applying the convex hull algorithm again, a concave polygon can be obtained. This means that only the first application of the convex hull algorithm yields a convex polygon. With each subsequent iteration, the resulting merging is no longer convex, but concave. After the convex hull algorithm completes, the topological relationship between each remaining point and the current polyline remains the same, allowing them to be either inside or outside the polyline. If the number of iterations is odd, the remaining points are outside the polyline; if the number of iterations is even, the remaining points are inside.

[0075] By further fitting the obtained polyline polygon and the remaining undetermined points, the polyline can be further fitted to obtain smooth contour lines. There are many polyline fitting methods, such as cubic polynomials and spline function fitting. The method used in this article is the spline function fitting method.

[0076] Contour lines generally exhibit a distribution of detail and can be used to represent surface features. Complex contour lines on a contour line indicate that certain areas have more features, while simple contour lines indicate that certain areas have fewer features. However, as discussed previously, the complexity of the contour line construction process increases with the number of iterations. Therefore, the number of convex hull iterations is a criterion for distinguishing contour lines. In other words, a lower number of iterations indicates a lower level of detail in the contour line. Therefore, the initial contour line obtained from the first iteration of the convex hull algorithm is considered the one with the lowest level of detail. This contour line also has the fewest features and lacks significant detail. These contour lines and the remaining undetermined points are then classified. After this classification, the convex hull algorithm is applied to these points for a second iteration, again yielding contour lines with the next lowest level of detail. This process continues in this manner, resulting in a series of contour lines with increasing levels of detail.

[0077] In this embodiment, step S50 further includes the following steps S51 to S52.

[0078] S51. Determine the external projection data outside the local convex hull and the internal projection data inside the local convex hull; S52. Based on the layered projection relationship, eliminate the three-dimensional point cloud data corresponding to the external projection data to obtain effective stockpile data, and render the stockpile in real time based on the effective stockpile data.

[0079] In this embodiment, a polygonal scan line algorithm is used to determine whether the projection data point is located inside, outside, or on the edge of the convex hull. That is, the position of the projection data point relative to the convex hull is determined based on the scan line algorithm, and the external projection data located outside the convex hull is eliminated, while the internal projection data located inside and on the edge of the convex hull is retained, thereby obtaining valid stockpile data.

[0080] In an exemplary embodiment, the internal projection data is obtained through the following steps.

[0081] As an optional implementation of the present disclosure, step S43 further includes A10 to A40.

[0082] A10. Determine coordinate information of vertices of the local convex hull, and determine the extreme vertices of the local convex hull based on the coordinate information.

[0083] For a planar polygon, the vertex with the minimum / maximum x-coordinate value and the vertex with the minimum / maximum y-coordinate value are called the extreme vertices of the planar polygon. For the local convex hull obtained previously, a plane coordinate system can be established with p0 as the coordinate origin, and the coordinates of each vertex can be obtained. A20, S432, determine the horizontal scan line that intersects the local convex hull, determine the intersection point of the horizontal scan line with the local convex hull, and determine whether the horizontal scan line passes through the extreme vertex of the local convex hull.

[0084] The horizontal scan line, also called the X-scan line, is parallel to the x-axis and starts at y = 0. The intersections with the local convex hull are determined, dividing the horizontal scan line into segments. Some segments lie outside the local convex hull, while others lie inside.

[0085] A30. Determine that the horizontal scan line passes through the extreme vertex, and divide the extreme vertex into two extreme points based on the edge of the local convex hull that generates the extreme vertex.

[0086] Since the local convex hull is a convex polygon, for horizontal scan lines that do not pass through vertices, there is always an even number of intersections with the local convex hull, and the line segments inside the local convex hull and the line segments outside the local convex hull exist alternately; for horizontal scan lines that pass through vertices, the intersections with the local convex hull may be either even or odd. Further division, if the horizontal scan line passes through an extreme vertex, there is always an odd number of intersections with the local convex hull. If the horizontal scan line passes through a non-extreme vertex, there is always an even number of intersections with the local convex hull. In this case, all extreme vertices are regarded as two points, which can ensure that the intersections of the horizontal scan line and the local convex hull are always even. For example, any vertex is obtained by the intersection of two edges of the local convex hull. One of the endpoints of the two intersecting edges is the vertex, and the other endpoint of the edge is another vertex. Therefore, the extreme vertex can be marked by the other endpoint values ​​of the two edges A and B that generate the extreme vertex. The extreme vertex is regarded as two points. The endpoint values ​​of the two ends of edge A are A1 and A2 respectively, and the endpoint values ​​of the two ends of edge B are B1 and B2 respectively. A1 and B1 coincide to form an extreme vertex. Then the extreme vertex can be considered as the following two points: as well as

[0087] A40. Pair the intersections from left to right. It can be understood that extreme points are also intersections. Determine the pairing points and connect them with straight lines. Eliminate the external projection data that is located on the horizontal scan line and not on the line connecting the pairing points to obtain the internal projection data. Each intersection has one and only one other intersection to pair with, and the line between the pairing points is a horizontal line. For example, intersection A is paired with intersection B, so intersection A and intersection B are a pair of pairing points. The second scan data between a pair of successfully paired pairs is valid canopy data, and these second scan data are located inside the local convex hull or on the edge of the local convex hull. For example, the sequence of intersections between a horizontal scan line C and the local convex hull is EFGH. Pair them in pairs from left to right, then connect the intersections E and F and the intersections G and H respectively to obtain line segments EF and GH. Finally, eliminate the external projection data that is located on the horizontal scan line C and not on the line connecting the pairing points EF and GH.

[0088] Example 2

[0089] In this embodiment, a device for real-time rendering of three-dimensional material piles is provided. It can be understood that the device for real-time rendering of three-dimensional material piles in the embodiment of the present disclosure can be used in electronic devices, including but not limited to computers, mobile terminals, etc. Figure 4 is a structural schematic diagram of the device for real-time rendering of three-dimensional material piles according to the embodiment of the present disclosure. As shown in Figure 4, the device includes: an acquisition module 10, a layering module 20, a first determination module 30, a second determination module 40, and a real-time rendering module 50.

[0090] The acquisition module 10 is configured to acquire three-dimensional point cloud data of the material pile to be inspected. The three-dimensional point cloud data may be pre-stored in the electronic device or acquired externally. For example, the electronic device may acquire the data from an external acquisition device or extract the data from externally acquired data.

[0091] There is no restriction on the specific acquisition method of the three-dimensional point cloud data. It only needs to ensure that the electronic device can acquire the three-dimensional point cloud data.

[0092] The layering module 20 is configured to layer the three-dimensional point cloud data based on the asymmetric data in the three-dimensional point cloud data, and obtain several layers of plane layered data and a layered mapping relationship of each layer of plane layered data; the layered mapping relationship is the mapping relationship between the three-dimensional point cloud data and the plane layered data.

[0093] Layered processing can reduce the dimensionality of three-dimensional data to obtain corresponding two-dimensional data, and establish a mapping relationship between three-dimensional data points and two-dimensional projection data points. In this method, the three-dimensional point cloud data is composed of several three-dimensional data points, and the layered projection data is composed of several layered projection data points. There are at least two three-dimensional layer data points corresponding to the same layered projection data point, that is, each three-dimensional layer data point corresponds to a layered projection data point, and a layered projection data point may correspond to multiple three-dimensional data points.

[0094] The first determination module 30 is configured to determine the convex hull corresponding to each layer of planar layered data. For a two-dimensional plane, the convex hull is the minimum convex polygon enclosing all layered projected data points. In one exemplary embodiment, given a set S of layered projected data points, the intersection T of all convex sets that contain set S is called the convex hull of set S.

[0095] In the present disclosure, the convex hull of the first projection data may be determined based on Graham scanning method, exhaustive method, divide-and-conquer method, Jarvis stepping method, Melkman algorithm, etc., and again, no limitation is imposed on the specific convex hull algorithm.

[0096] The second determining module 40 is configured to determine the local convex hull of the corresponding layer plane layered data based on the convex hull.

[0097] For a point set Q containing N points and a small convex polygon K, there are usually three cases where every point in Q is contained, either inside K or on the boundary of K. Convex hull algorithms have important applications in imaging, geospatial computing, computer graphics, and other fields.

[0098] The convex polygon obtained by the convex hull algorithm is a subset of the point set and can be considered an approximation of the contour line of the object represented by the point set. However, there are still discrete points on the object's surface that are not included in the polyline. Therefore, these points need to be classified onto the current polyline. To more accurately represent the contour line, we first need to determine which edge within the polyline each point belongs to.

[0099] For each undetermined point Q, the relationship between the projection and distance between the point and each line segment of the current contour line needs to be calculated. Each undetermined point can only be classified into two line segments with an angle less than 90 degrees and the closest distance.

[0100] After classifying the current contour line and pending points using the above method, recalculate the convex hull of each segment of the polyline and the pending points classified into that segment to obtain the convex hull of the local area. All local convex hulls are then disconnected at the original contour line segment to obtain open polyline segments. Connect the previous and next polyline segments in sequence to obtain the next level of polyline.

[0101] If you're studying a layer of point cloud data, the first application of the convex hull algorithm yields a convex polygon. Continuing with the convex hull algorithm, you can further obtain a partially convex polygon. If you then integrate the convex polygons, you'll obtain a concave polygon composed of contour lines, rather than a convex polygon. Next, classify the remaining points in the concave polygon and apply the convex hull algorithm again, still yielding a concave polygon. In other words, only the first application of the convex hull algorithm yields a convex polygon. After each subsequent iteration, the resulting merging results in a concave polygon, no longer a convex polygon. After the convex hull algorithm completes, each remaining point maintains the same topological relationship with the current polyline and can be either inside or outside the polyline. If the number of iterations is odd, the remaining points are outside the polyline; if the number of iterations is even, the remaining points are inside.

[0102] By further fitting the obtained polyline polygon and the remaining undetermined points, the polyline can be further fitted to obtain smooth contour lines. There are many polyline fitting methods, such as cubic polynomials and spline function fitting. The method used in this article is the spline function fitting method.

[0103] Contour lines generally exhibit a distribution of detail and can be used to represent surface features. Complex contour lines on a contour line indicate that certain areas have more features, while simple contour lines indicate that certain areas have fewer features. However, as discussed previously, the complexity of the contour line construction process increases with the number of iterations. Therefore, the number of convex hull iterations is a criterion for distinguishing contour lines. In other words, a lower number of iterations indicates a lower level of detail in the contour line. Therefore, the initial contour line obtained from the first iteration of the convex hull algorithm is considered the one with the lowest level of detail. This contour line also has the fewest features and lacks significant detail. These contour lines and the remaining undetermined points are then classified. After this classification, the convex hull algorithm is applied to these points for a second iteration, again yielding contour lines with the next lowest level of detail. This process continues in this manner, resulting in a series of contour lines with increasing levels of detail.

[0104] Assume that the number of iterations for constructing the highest level contour line obtained by the convex hull algorithm for a certain layer of points is k. Then, according to the actual level of detail required, the level is set to m, depending on the number of iterations corresponding to each level of detail. For example, when k = 7 and m = 4, it can be seen that the number of iterations for the first level contour line, that is, the lowest detail contour line, is 1, the number of iterations for the second level is 3, and so on. The number of iterations for the third level is 5, and the number of iterations for the fourth level, that is, the highest detail contour line, is 7.

[0105] During the process of constructing contour lines, when the number of iterations is from 1 to k, each time the calculation is completed, the contour line of that level has been calculated. Therefore, it only needs to be stored in the memory. After the contour line containing all points is calculated, the contour lines of other levels have also been completed. Therefore, it is only necessary to match the number of iterations with the required contour line detail level to obtain the contour lines with high to low details.

[0106] The real-time rendering module 50 is configured to eliminate the plane layered data outside the corresponding local convex hull to obtain effective stockpile data, and perform real-time rendering of the stockpile based on the effective stockpile data.

[0107] Since point cloud data is often scattered and disordered, and contains a lot of redundant data, it is very difficult to remove redundant data and find the topological relationship between points during the 3D modeling process, especially when studying algorithms from a 3D perspective. However, algorithms within the 2D plane are relatively mature, and the data on a scan line is in the same plane. This leads to the idea of ​​converting scattered point cloud data into scan line data. After the elimination process, the noise data can be better eliminated. This noise data needs to be filtered out to obtain more realistic stockpile data. In this method, the convex hull algorithm and the scan line algorithm are used for elimination. Particles located outside the layered projection data in the simulated fabric are eliminated to obtain valid stockpile data. The 3D point cloud data is also directly layered, so the convex hull algorithm is called to further refine each layer of plane layered data.

[0108] The device for real-time rendering of a three-dimensional material pile disclosed in the present invention obtains asymmetric data in three-dimensional point cloud data, layers the three-dimensional point cloud data, obtains several layers of plane layered data and a layered mapping relationship of each layer of plane layered data, then determines the convex hull corresponding to each layer of plane layered data, determines the local convex hull of the corresponding layer of plane layered data based on the convex hull, and finally eliminates the data of the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and renders the material pile in real time based on the effective material pile data, layers the collected three-dimensional point cloud data structure, calls the convex hull algorithm to refine each layer of point cloud data, and uses the convex hull algorithm to process the three-dimensional point cloud data. The generated three-dimensional model is more accurate and faster, which is convenient for subsequent rendering, resulting in better rendering effect of the material pile, ensuring the safe operation of the material retrieving operation, and real-time changes in the material pile are fed back in real time in the three-dimensional monitoring system.

[0109] Example 3

[0110] FIG5 illustrates a schematic diagram of the physical structure of an electronic device. As shown in FIG5 , the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communications bus 540. The processor 510, the communications interface 520, and the memory 530 communicate with each other via the communications bus 540. The processor 510 may invoke logic commands in the memory 530 to execute a method for real-time rendering of a three-dimensional material pile. The method includes: obtaining three-dimensional point cloud data of a material pile to be inspected; layering the three-dimensional point cloud data based on asymmetric data in the three-dimensional point cloud data to obtain several layers of plane layered data and a layered mapping relationship for each layer of plane layered data; the layered mapping relationship being a mapping relationship between the three-dimensional point cloud data and the plane layered data; determining the convex hull corresponding to each layer of plane layered data; determining the local convex hull of the corresponding layer of plane layered data based on the convex hull; eliminating the plane layered data outside the corresponding local convex hull to obtain valid material pile data; and rendering the material pile in real time based on the valid material pile data.

[0111] In addition, the logical commands in the above-mentioned memory 530 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software medium, and the computer software medium is stored in a storage medium, including a number of commands to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0112] Example 4

[0113] On the other hand, the present disclosure also provides a computer program medium, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for real-time rendering of a three-dimensional material pile provided by the above methods, the method including: obtaining three-dimensional point cloud data of the material pile to be detected; based on the asymmetric data in the three-dimensional point cloud data, layering the three-dimensional point cloud data to obtain several layers of plane layered data and a layered mapping relationship of each layer of plane layered data; the layered mapping relationship is a mapping relationship between the three-dimensional point cloud data and the plane layered data; determining the convex hull corresponding to each layer of plane layered data; determining the local convex hull of the corresponding layer of plane layered data based on the convex hull; eliminating the data of the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and rendering the material pile in real time based on the effective material pile data.

[0114] Example 5

[0115] On the other hand, the present disclosure also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for real-time rendering of a three-dimensional material pile provided by the above-mentioned methods, the method comprising: obtaining three-dimensional point cloud data of the material pile to be detected; layering the three-dimensional point cloud data based on asymmetric data in the three-dimensional point cloud data to obtain several layers of plane layered data and a layered mapping relationship of each layer of plane layered data; the layered mapping relationship is a mapping relationship between the three-dimensional point cloud data and the plane layered data; determining the convex hull corresponding to each layer of plane layered data; determining the local convex hull of the corresponding layer of plane layered data based on the convex hull; eliminating the data of the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and rendering the material pile in real time based on the effective material pile data.

[0116] The present invention provides a method, device, equipment and medium for real-time rendering of a three-dimensional material pile, which obtains asymmetric data in three-dimensional point cloud data, layers the three-dimensional point cloud data, obtains several layers of plane layered data and a layered mapping relationship of each layer of plane layered data, then determines the convex hull corresponding to each layer of plane layered data, determines the local convex hull of the corresponding layer of plane layered data based on the convex hull, and finally eliminates the data of the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and renders the material pile in real time based on the effective material pile data, layers the collected three-dimensional point cloud data structure, calls the convex hull algorithm to refine each layer of point cloud data, and uses the convex hull algorithm to process the three-dimensional point cloud data. The generated three-dimensional model is more accurate and faster, which is convenient for subsequent rendering, so that the material pile rendering effect is better, and the safe operation of the material retrieving operation is guaranteed. The real-time changes of the material pile are fed back in real time in the three-dimensional monitoring system.

[0117] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units. That is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0118] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the above technical solution, in essence, or the part that contributes to the prior art, can be embodied in the form of a software medium. The computer software medium can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of commands for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or certain parts of the embodiment.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present disclosure.

Claims

1. A method for real-time rendering of a three-dimensional stockpile, comprising: Obtain three-dimensional point cloud data of the material pile to be inspected; Based on the asymmetric data in the three-dimensional point cloud data, the three-dimensional point cloud data is layered to obtain a plurality of layers of plane layered data and a layered mapping relationship of each layer of the plane layered data; The hierarchical mapping relationship is the mapping relationship between the three-dimensional point cloud data and the plane hierarchical data; Determine the convex hull corresponding to each layer of the plane layered data; Determine the local convex hull of the plane hierarchical data of the corresponding layer based on the convex hull; The data of the plane layered data outside the corresponding local convex hull is eliminated to obtain the effective stockpile data, and the stockpile is rendered in real time based on the effective stockpile data.

2. The method for real-time rendering of a three-dimensional stockpile according to claim 1, wherein: The determining of the convex hull corresponding to each layer of the planar hierarchical data further comprises: Determine the layered projection data with the smallest ordinate value in the preset area, and determine the determined layered projection data as a reference point; Determine the argument between the layered projection data and the reference point, sort the layered projection data in ascending order based on the argument, and determine the sort value corresponding to the layered projection data; Create an empty stack, and push the reference point and the layered projection data corresponding to the minimum sorting value into the stack in sequence; Determine a connecting line between the layered projection data corresponding to the stack vertex and the adjacent stack elements, and determine the positional relationship between the layered projection data and the connecting line according to the sorting order; Determine that the layered projection data is located on the left side of the connecting line, push the layered projection data into the stack, and update the stack vertices and the connecting line; Determine that the layered projection data is located on the right side of the connecting line, pop the stack vertex, and update the stack vertex and the connecting line; All hierarchical projection data with sorted values ​​are traversed to determine the elements in the stack, and the hierarchical projection data corresponding to adjacent elements are connected to obtain the convex hull corresponding to the first projection data.

3. The method for real-time rendering of a three-dimensional stockpile according to claim 2, wherein: The determining of the local convex hull of the plane hierarchical data of the corresponding layer based on the convex hull further comprises: Determine the planar layered data located inside the convex hull; Determine the vertical distances corresponding to the plane layered data and each edge of the convex hull, and based on the vertical distances, determine the edge closest to the plane layered data, and use the edge closest to the plane layered data as a candidate folding edge; Connect the plane layered data and the two vertices of the candidate folded edge to obtain the candidate straight line; Determine the angle between the candidate straight line and the candidate folded edge; Determine that the angles corresponding to the plane layered data are all acute angles, use the plane layered data as candidate connection points, and classify the candidate connection points into corresponding edges of the convex hull; Disconnect the two vertices on the edge corresponding to the candidate connection point, and connect the two vertices to the candidate connection point respectively to obtain the local convex hull.

4. The method for real-time rendering of a three-dimensional stockpile according to claim 1, wherein: The data of the plane layered data located outside the corresponding local convex hull is eliminated to obtain effective stockpile data, and the stockpile is rendered in real time based on the effective stockpile data, further comprising: Determine external projection data outside the local convex hull and internal projection data inside the local convex hull; Based on the hierarchical projection relationship, the three-dimensional point cloud data corresponding to the external projection data is eliminated to obtain the effective material pile data, and the material pile is rendered in real time based on the effective material pile data.

5. The method for real-time rendering of a three-dimensional stockpile according to claim 4, wherein: The internal projection data is obtained by the following steps: Determine coordinate information of vertices of the local convex hull, and determine the extreme vertices of the local convex hull based on the coordinate information; Determine a horizontal scan line that has an intersection with the local convex hull, determine the intersection point corresponding to the horizontal scan line and the local convex hull, and determine whether the horizontal scan line passes through the extreme vertex of the local convex hull; Determine that the horizontal scan line passes through the extreme vertex, and divide the extreme vertex into two extreme points based on the edge of the local convex hull that generates the extreme vertex; Pair the intersection points from left to right, determine the pairing points, connect the pairing points with straight lines, remove the external projection data that is located on the horizontal scan line and not on the line connecting the pairing points, and obtain the internal projection data.

6. The method for real-time rendering of a three-dimensional stockpile according to claim 1, wherein: The step of layering the three-dimensional point cloud data based on the asymmetric data in the three-dimensional point cloud data to obtain a plurality of layers of plane layered data and a layered mapping relationship of each layer of the plane layered data further includes: Determining symmetric data and asymmetric data in the three-dimensional point cloud data; Based on the asymmetric three-dimensional point cloud data and a preset height value, determining a minimum distance associated point corresponding to the asymmetric three-dimensional point cloud data; The three-dimensional point cloud data is layered based on the minimum distance associated points to obtain a plurality of layers of plane layered data and a layered mapping relationship of each layer of the plane layered data.

7. The method for real-time rendering of a three-dimensional stockpile according to claim 6, wherein: The step of determining the minimum distance associated point corresponding to the asymmetric three-dimensional point cloud data based on the asymmetric three-dimensional point cloud data and the preset height value further includes: Based on the asymmetric three-dimensional point cloud data and a preset height value, determining an upper neighborhood point set and a lower neighborhood point set; Find the minimum distance of each point in the upper neighborhood point set in the lower neighborhood point set, and find the minimum distance of each point in the lower neighborhood point set in the upper neighborhood point set; It is determined that the distances between the two are equal, and the point with the smaller value is used as the minimum distance associated point.

8. A device for real-time rendering of a three-dimensional stockpile, comprising: An acquisition module configured to acquire three-dimensional point cloud data of a material pile to be detected; A layering module is configured to layer the three-dimensional point cloud data based on the asymmetric data in the three-dimensional point cloud data to obtain a plurality of layers of plane layered data and a layered mapping of each layer of the plane layered data. Relationship; the hierarchical mapping relationship is the mapping relationship between the three-dimensional point cloud data and the plane hierarchical data; A first determining module is configured to determine the convex hull corresponding to each layer of the plane layered data; A second determination module is configured to determine a local convex hull of the plane hierarchical data of a corresponding layer based on the convex hull; The real-time rendering module is configured to eliminate the plane layered data located outside the corresponding local convex hull to obtain effective material pile data, and perform real-time rendering of the material pile based on the effective material pile data.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for real-time rendering of a three-dimensional material pile as claimed in any one of claims 1 to 7 when executing the program.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the method for real-time rendering of a three-dimensional stockpile are implemented as claimed in any one of claims 1 to 7.

Images