A sandbox pouring positioning system based on three-dimensional point cloud

Abstract

The present application discloses a sandbox casting positioning system based on three-dimensional point cloud, which relates to the technical field of the casting industry. It includes a casting trolley, a laser radar, a laser rangefinder, an incremental encoder and a computing module. The casting trolley is arranged above the casting production line. The laser radar is used to scan the sandbox on the casting production line, the laser rangefinder is used to measure the distance of the sandbox on the casting production line, and the incremental encoder is used to detect and contour scan the sandbox on the casting production line. The present application uses a laser radar for non-contact positioning to obtain the sandbox pouring cup area, which improves the efficiency of sandbox casting, and uses a laser radar for contour detection instead of the traditional photoelectric sensor or mechanical limit to trigger the sandbox scanning, which adapts to the scanning trigger of sandboxes of different specifications and sizes, and uses a laser rangefinder in conjunction with a high-precision incremental encoder to detect the position of the casting trolley in real time, establishes a high-precision sandbox scanning model, and improves positioning accuracy.

Description

Technical Field

[0001] The present application relates to the technical field of the casting industry, and specifically to a sandbox casting positioning system based on three-dimensional point cloud. Background Art

[0002] The steel industry often uses sandboxes to make molds for castings. Sandboxes vary in size according to the size of the castings and are generally arranged on the casting production line. Figure 2 As shown, the pouring trolley is located above the pouring line and can move along the pouring line. The molten iron ladle is placed on the pouring trolley to hold the molten iron. When pouring begins, the pouring worker starts the pouring trolley and moves it along the pouring line. After it moves to the top of the sandbox to be poured, the worker manually aligns the outlet of the molten steel with the pouring cup on the upper surface of the sandbox, opens the valve of the molten iron ladle, and pours the molten iron into the sandbox. After the sandbox is poured, the pouring trolley continues to move, and the remaining sandboxes are poured in this way. This pouring method has the following problems:

[0003] 1. The efficiency of manually positioning the pouring cup is low. The height of the sandbox and the position of the pouring cup of the sandbox are different. Workers need to adjust them one by one according to the actual situation when pouring a single sandbox, which leads to low pouring efficiency.

[0004] 2. The temperature in the pouring environment is high, and there is a lot of dust during the pouring process. Once the molten steel is melted, it needs to be poured as quickly as possible. Therefore, the operator's proficiency and tolerance to harsh environments are high. Summary of the invention

[0005] The purpose of this application is to provide a sandbox casting positioning system based on three-dimensional point cloud to solve the technical problems raised in the above-mentioned background technology.

[0006] To achieve the above-mentioned purpose, the present application discloses the following technical solutions: a sandbox casting positioning system based on three-dimensional point cloud, comprising a casting trolley, on which a laser radar, a laser rangefinder and an incremental encoder are installed, and the laser radar, the laser rangefinder and the incremental encoder are respectively connected to a computing module for communication;

[0007] The pouring trolley is arranged above the pouring production line and moves along the pouring production line;

[0008] The scanning end of the laser radar is arranged toward the top side of the casting production line, and is used to scan the sandbox on the casting production line;

[0009] The laser rangefinder is arranged above the casting production line and is used to measure the distance of the sandbox on the casting production line;

[0010] The incremental encoder is arranged on both sides of the casting production line, and is used for detecting and scanning the contour of the sand box on the casting production line;

[0011] During pouring positioning, the pouring trolley moves along the pouring production line. When the laser radar scans the sandbox on the pouring production line, the two-dimensional contour and actual position of the sandbox are obtained in combination with the laser rangefinder and the incremental encoder. The calculation module integrates the two-dimensional contour to form a three-dimensional contour, and calculates and locates the pouring cup mouth data of the sandbox to complete the pouring positioning of the sandbox.

[0012] During pouring, the calculation module sends the calculation results to the PLC, and the PLC is used to control the movement of the pouring trolley and move the ladle iron nozzle on the pouring trolley to a position above the pouring cup mouth of the sand box, and control the ladle iron nozzle to descend to the pouring height to pour the sand box.

[0013] Preferably, the calculation module completes the calculation of the pouring cup data of one sandbox before the next sandbox is detected.

[0014] Preferably, after the ladle shroud completes pouring of a sand box, the PLC controls the ladle shroud to close the shroud, and controls the ladle shroud to reset its height, and after resetting, the pouring trolley moves along the pouring production line to perform inspection and pouring positioning for the next sand box.

[0015] Preferably, the step of obtaining the two-dimensional profile and actual position of the sandbox by combining the laser rangefinder and the incremental encoder specifically includes:

[0016] The laser radar detects the contour of the sandbox in real time, and based on the ranging result of the laser rangefinder, divides the contour detection result of the sandbox into high contour and low contour according to the number of high and low points in the contour points; if the current frame data is low contour data and the previous frame data is high contour data, it is determined that the rising edge trigger of the regional contour of the sandbox has arrived; if the previous frame data is high contour data and the current frame is low contour data, it is determined that the falling edge trigger of the sandbox has arrived.

[0017] Preferably, when the rising edge trigger of the area contour of the sandbox arrives, the initial position of the laser radar, the feedback position of the incremental encoder and the two-dimensional contour scanned by the laser radar are recorded; wherein, the two-dimensional contour scanned by the laser radar includes y-direction information and z-direction information, and the x-direction information in the two-dimensional contour is calculated by the initial position of the laser radar, the feedback position of the incremental encoder and the resolution, and then the single-frame contour is spliced ​​along the x-direction to obtain a real-time scanned three-dimensional model, wherein the y-direction is the width direction of the casting production line, the x-direction is the measurement direction of the laser rangefinder, and the z-direction is perpendicular to the ground and upward; and at the same time, it is detected in real time whether the falling edge trigger of the sandbox arrives. If the falling edge trigger of the sandbox arrives, it is determined that the sandbox acquisition is completed, and the calculation module performs the positioning calculation of the sandbox and calculates and locates the pouring cup mouth data of the sandbox.

[0018] Preferably, integrating the two-dimensional profile to form a three-dimensional profile specifically comprises:

[0019] When the rising edge of the area contour of the sandbox is triggered, the position of the casting trolley fed back by the laser rangefinder is recorded, and the single-frame contour and the value of the incremental encoder are collected. The single-frame contour is the yoz cross-sectional contour scatter point. The measured value of the laser rangefinder at this time is used as the x-direction information of the single-frame contour, combined with the y-direction information and the z-direction information in the single-frame contour, and converted into a three-dimensional coordinate system point;

[0020] In the contour after the rising edge of the area contour of the sandbox is triggered, the detection values ​​of the contour and the incremental encoder are recorded, and the increment in the x direction of the current contour is calculated by combining the detection value of the incremental encoder with the incremental encoder of the starting frame, and the measurement value of the laser rangefinder according to the starting frame is converted into the global x coordinate of the system, and then converted into a three-dimensional coordinate system point;

[0021] As the pouring trolley moves, multiple frames of contours are continuously stored in the sandbox three-dimensional model. When the falling edge trigger of the sandbox arrives, contour collection is stopped and the sandbox area contour collection is completed.

[0022] Preferably, the calculation and positioning of the pouring cup mouth data of the sandbox specifically includes:

[0023] S1: Rough cutting, cutting and removing invalid areas according to the height range of the sandbox area, wherein the invalid areas include the ground area;

[0024] S2: downsampling processing, downsampling in the height direction;

[0025] S3: Point cloud segmentation: the point cloud data is collected and cut into a single sub-area, and whether there is a sandbox area in the point cloud data segment is detected according to the size of the sandbox area. If there is no sandbox area, contour detection is continued. If there is a sandbox area, the average height of the sandbox area is calculated as the height of the sandbox.

[0026] S4: Segment the original dense point cloud based on the sandbox height obtained by detection to obtain the pouring cup mouth point cloud area and the upper surface point cloud area;

[0027] S5: segment the point cloud connected domain according to the point cloud area on the upper surface of the sandbox;

[0028] S6: Determine whether there is an area with the same size as the pouring cup mouth in the area segmented in S5. If so, use the area as the sandbox pouring cup mouth detection area to obtain the range of the circumscribed rectangle of the pouring cup mouth area;

[0029] S7: If the pouring cup mouth area is detected in S6, the upper point cloud of the pouring cup mouth is segmented based on the maximum value range of the x-direction and y-direction of the circumscribed rectangle of the pouring cup mouth area and the height in the z-direction, and the maximum height Zcup of the pouring cup mouth is calculated. At the same time, the center point of the pouring cup mouth area in the x-direction and y-direction is calculated as the Xcup and Ycup of the pouring area;

[0030] S8: If the pouring cup mouth area is detected in S6, the point cloud connected domain segmentation is performed on the area again, and the segmentation judgment is performed on the segmented sub-areas; if the number of areas in the segmented area that are consistent with the pouring cup mouth size is 0, it means that the sandbox pouring cup mouth area is not detected; if the number of areas in the segmented area that are consistent with the pouring cup mouth size is 1, S7 is executed to complete the pouring cup mouth positioning calculation.

[0031] Preferably, if the number of regions in the segmented area that is consistent with the size of the pouring cup mouth is greater than 1, the segmented areas are respectively defined as to-be-processed segmented areas, and the multiple to-be-processed segmented areas are projected onto the plane, wherein the x and y directions correspond to the col_index and row index of the image respectively. After the segmented areas are projected, component segmentation of the connected domain of the two-dimensional image is performed, and it is determined whether the roundness of the segmented components is within a set range. If the roundness of a component in the component segmentation result corresponding to a segmented area to be processed is within a set threshold range, the segmented area to be processed is determined to be a sandbox pouring cup mouth area, and S7 is executed to complete the sandbox pouring cup mouth positioning; if the roundness of the components in the component segmentation results corresponding to all the segmented areas to be processed is not within the set threshold range, it means that all the segmented areas to be processed are not pouring cup mouth areas, and contour detection is continued.

[0032] Beneficial effects: The sandbox pouring positioning system based on three-dimensional point cloud of the present application adopts laser radar for non-contact positioning to obtain the sandbox pouring cup area, which improves the efficiency of sandbox pouring, and uses laser radar for contour detection instead of traditional photoelectric sensors or mechanical limiters to trigger sandbox scanning, which is suitable for scanning triggering of sandbox sizes of different specifications, and uses laser rangefinder in conjunction with high-precision incremental encoder to detect the position of the pouring trolley in real time, establishes a high-precision sandbox scanning model, and improves positioning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 This is a structural block diagram of a sandbox casting positioning system based on three-dimensional point cloud in an embodiment of the present application;

[0035] Figure 2 It is a structural schematic diagram of a sandbox pouring line in the prior art;

[0036] Figure 3 A schematic diagram of the pouring positioning process in an embodiment of the present application;

[0037] Figure 4 A schematic diagram of the process of three-dimensional modeling in an embodiment of the present application;

[0038] Figure 5 This is a schematic diagram of the specific process of pouring positioning in an embodiment of the present application. DETAILED DESCRIPTION

[0039] The technical solutions in the embodiments of the present application are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0040] In this article, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprising..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0041] This embodiment discloses a Figure 1 The sandbox casting positioning system based on three-dimensional point cloud shown in the figure includes a casting trolley, on which a laser radar, a laser rangefinder and an incremental encoder are installed, and the laser radar, the laser rangefinder and the incremental encoder are respectively connected to the computing module for communication. The casting trolley is arranged above the casting production line and moves along the casting production line; the scanning end of the laser radar is arranged toward the top side of the casting production line, and is used to scan the sandbox on the casting production line; the laser rangefinder is arranged above the casting production line, and is used to measure the distance of the sandbox on the casting production line; the incremental encoder is arranged on both sides of the casting production line, and is used to detect and scan the contour of the sandbox on the casting production line.

[0042] like Figure 3 As shown, during pouring positioning, the pouring trolley moves along the pouring production line. When the laser radar scans the sand box on the pouring production line, the two-dimensional contour and actual position of the sand box are obtained in combination with the laser rangefinder and the incremental encoder. The calculation module integrates the two-dimensional contour to form a three-dimensional contour, and calculates and locates the pouring cup mouth data of the sand box to complete the pouring positioning of the sand box.

[0043] During pouring, the calculation module sends the calculation result to the PLC, and the PLC is used to control the movement of the pouring trolley and move the ladle iron nozzle on the pouring trolley to the position above the pouring cup mouth of the sandbox, and control the ladle iron nozzle to descend to the pouring height to pour the sandbox. In particular, the calculation module completes the calculation of the pouring cup mouth data of a sandbox before the next sandbox is detected. Moreover, after the ladle iron nozzle completes the pouring of a sandbox, the PLC controls the ladle iron nozzle to close the iron nozzle, and controls the ladle iron nozzle to reset the height. After resetting, the pouring trolley moves along the pouring production line to detect and position the next sandbox for pouring.

[0044] Combination Figure 4 and Figure 5 As shown, the combination of the laser rangefinder and the incremental encoder to obtain the two-dimensional profile and actual position of the sandbox specifically includes:

[0045] The laser radar detects the contour of the sandbox in real time, and based on the ranging result of the laser rangefinder, divides the contour detection result of the sandbox into high contour and low contour according to the number of high and low points in the contour points; if the current frame data is low contour data and the previous frame data is high contour data, it is determined that the rising edge trigger of the regional contour of the sandbox has arrived; if the previous frame data is high contour data and the current frame is low contour data, it is determined that the falling edge trigger of the sandbox has arrived.

[0046] When the rising edge trigger of the area contour of the sandbox arrives, the initial position of the laser radar, the feedback position of the incremental encoder and the two-dimensional contour scanned by the laser radar are recorded; wherein, the two-dimensional contour scanned by the laser radar includes y-direction information and z-direction information, and the x-direction information in the two-dimensional contour is calculated through the initial position of the laser radar, the feedback position of the incremental encoder and the resolution, and then the single-frame contour is spliced ​​along the x-direction to obtain a real-time scanned three-dimensional model, wherein the y-direction is the width direction of the casting production line, the x-direction is the measurement direction of the laser rangefinder, and the z-direction is perpendicular to the ground and upward; and at the same time, it is detected in real time whether the falling edge trigger of the sandbox arrives. If the falling edge trigger of the sandbox arrives, it is determined that the sandbox acquisition is completed, and the calculation module performs the positioning calculation of the sandbox and calculates and locates the pouring cup mouth data of the sandbox.

[0047] The step of integrating the two-dimensional profile to form a three-dimensional profile specifically includes:

[0048] When the rising edge of the area contour of the sandbox is triggered, the position of the casting trolley fed back by the laser rangefinder is recorded, and the single-frame contour and the value of the incremental encoder are collected. The single-frame contour is the yoz cross-sectional contour scatter point. The measured value of the laser rangefinder at this time is used as the x-direction information of the single-frame contour, combined with the y-direction information and the z-direction information in the single-frame contour, and converted into a three-dimensional coordinate system point;

[0049] In the contour after the rising edge of the area contour of the sandbox is triggered, the detection values ​​of the contour and the incremental encoder are recorded, and the increment in the x direction of the current contour is calculated by combining the detection value of the incremental encoder with the incremental encoder of the starting frame, and the measurement value of the laser rangefinder according to the starting frame is converted into the global x coordinate of the system, and then converted into a three-dimensional coordinate system point;

[0050] As the pouring trolley moves, multiple frames of contours are continuously stored in the sandbox three-dimensional model. When the falling edge trigger of the sandbox arrives, contour collection is stopped and the sandbox area contour collection is completed.

[0051] The calculation and positioning of the pouring cup mouth data of the sandbox specifically includes:

[0052] S1: Rough cutting, cutting and removing invalid areas according to the height range of the sandbox area, wherein the invalid areas include the ground area;

[0053] S2: downsampling processing, downsampling in the height direction;

[0054] S3: Point cloud segmentation: the point cloud data is collected and cut into a single sub-area, and whether there is a sandbox area in the point cloud data segment is detected according to the size of the sandbox area. If there is no sandbox area, contour detection is continued. If there is a sandbox area, the average height of the sandbox area is calculated as the height of the sandbox.

[0055] S4: Segment the original dense point cloud based on the sandbox height obtained by detection to obtain the pouring cup mouth point cloud area and the upper surface point cloud area;

[0056] S5: segment the point cloud connected domain according to the point cloud area on the upper surface of the sandbox;

[0057] S6: Determine whether there is an area with the same size as the pouring cup mouth in the area segmented in S5. If so, use the area as the sandbox pouring cup mouth detection area to obtain the range of the circumscribed rectangle of the pouring cup mouth area;

[0058] S7: If the pouring cup mouth area is detected in S6, the upper point cloud of the pouring cup mouth is segmented based on the maximum value range of the x-direction and y-direction of the circumscribed rectangle of the pouring cup mouth area and the height in the z-direction, and the maximum height Zcup of the pouring cup mouth is calculated. At the same time, the center point of the pouring cup mouth area in the x-direction and y-direction is calculated as the Xcup and Ycup of the pouring area;

[0059] S8: If the pouring cup mouth area is detected in S6, the point cloud connected domain segmentation is performed on the area again, and the segmentation judgment is performed on the segmented sub-areas; if the number of areas in the segmented area that are consistent with the pouring cup mouth size is 0, it means that the sandbox pouring cup mouth area is not detected; if the number of areas in the segmented area that are consistent with the pouring cup mouth size is 1, S7 is executed to complete the pouring cup mouth positioning calculation.

[0060] If the number of regions in the segmented area that are consistent with the size of the pouring cup mouth is greater than 1, the segmented areas are defined as to-be-processed segmented areas, and the multiple to-be-processed segmented areas are projected onto the plane, wherein the x and y directions correspond to the col_index and row index of the image respectively. After the segmented areas are projected, component segmentation of the connected domain of the two-dimensional image is performed to determine whether the roundness of the segmented components is within a set range. If the roundness of a component in the component segmentation result corresponding to a segmented area to be processed is within a set threshold range, the segmented area to be processed is determined to be a sandbox pouring cup mouth area, and S7 is executed to complete the sandbox pouring cup mouth positioning. If the roundness of the components in the component segmentation results corresponding to all the segmented areas to be processed is not within the set threshold range, it means that all the segmented areas to be processed are not pouring cup mouth areas, and contour detection continues.

[0061] The sandbox casting positioning system based on three-dimensional point cloud of this embodiment adopts laser radar for non-contact positioning to obtain the sandbox casting cup area, which improves the efficiency of sandbox casting, and uses laser radar for contour detection instead of traditional photoelectric sensors or mechanical limiters to trigger sandbox scanning, which adapts to the scanning trigger of sandboxes of different specifications and sizes, and uses a laser rangefinder with a high-precision incremental encoder to detect the position of the casting trolley in real time, establishes a high-precision sandbox scanning model, and improves positioning accuracy. That is, the positioning system of this embodiment combines high-precision laser radar, laser rangefinder and incremental encoder to achieve high-precision scanning of the sandbox area to be positioned, and ensures the rapid and accurate positioning of sandboxes of different heights and widths through multiple point cloud data segmentation and preprocessing, which improves the basis for sandbox casting automation.

[0062] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A sandbox pouring positioning system based on three-dimensional point cloud, including a pouring trolley, It is characterized in that The pouring trolley is equipped with a laser radar, a laser rangefinder and an incremental encoder, and the laser radar, the laser rangefinder and the incremental encoder are respectively connected to the computing module for communication; The pouring trolley is arranged above the pouring production line and moves along the pouring production line; The scanning end of the laser radar is arranged toward the top side of the casting production line, and is used to scan the sandbox on the casting production line; The laser rangefinder is arranged above the casting production line and is used to measure the distance of the sandbox on the casting production line; The incremental encoder is arranged on both sides of the casting production line, and is used in combination with a laser radar and a laser rangefinder to detect and scan the contour of the sandbox on the casting production line; During pouring positioning, the pouring trolley moves along the pouring production line. When the laser radar scans the sandbox on the pouring production line, the two-dimensional contour and actual position of the sandbox are obtained in combination with the laser rangefinder and the incremental encoder. The calculation module integrates the two-dimensional contour to form a three-dimensional contour, and calculates and locates the pouring cup mouth data of the sandbox to complete the pouring positioning of the sandbox. During pouring, the calculation module sends the calculation results to the PLC, and the PLC is used to control the movement of the pouring trolley and move the ladle iron nozzle on the pouring trolley to a position above the pouring cup mouth of the sand box, and control the ladle iron nozzle to descend to the pouring height to pour the sand box.

2. The sandbox pouring positioning system based on three-dimensional point cloud according to claim 1, It is characterized in that The calculation module completes the calculation of the pouring cup data of one sandbox before the next sandbox is detected.

3. The sandbox pouring positioning system based on three-dimensional point cloud according to claim 2, It is characterized in that After the ladle molten iron nozzle completes the pouring of a sand box, the PLC controls the ladle molten iron nozzle to close the molten iron nozzle and controls the ladle molten iron nozzle to reset its height. After resetting, the pouring trolley moves along the pouring production line to perform inspection and pouring positioning for the next sand box.

4. The sandbox pouring positioning system based on three-dimensional point cloud according to claim 1, It is characterized in that The method of combining the laser rangefinder and the incremental encoder to obtain the two-dimensional profile and actual position of the sandbox specifically includes: The laser radar detects the contour of the sandbox in real time, and based on the ranging result of the laser rangefinder, divides the contour detection result of the sandbox into high contour and low contour according to the number of high and low points in the contour points; if the current frame data is high contour data and the previous frame data is low contour data, it is determined that the rising edge trigger of the regional contour of the sandbox has arrived; if the previous frame data is high contour data and the current frame is low contour data, it is determined that the falling edge trigger of the sandbox has arrived.

5. The sandbox pouring positioning system based on three-dimensional point cloud according to claim 4, It is characterized in that When the rising edge trigger of the area contour of the sandbox arrives, the initial position of the laser radar, the feedback position of the incremental encoder and the two-dimensional contour scanned by the laser radar are recorded; wherein, the two-dimensional contour scanned by the laser radar includes y-direction information and z-direction information, and the x-direction information in the two-dimensional contour is calculated through the initial position of the laser radar, the feedback position of the incremental encoder and the resolution, and then the single-frame contour is spliced ​​along the x-direction to obtain a real-time scanned three-dimensional model, wherein the y-direction is the width direction of the casting production line, the x-direction is the measurement direction of the laser rangefinder, and the z-direction is perpendicular to the ground and upward; and at the same time, it is detected in real time whether the falling edge trigger of the sandbox arrives. If the falling edge trigger of the sandbox arrives, it is determined that the sandbox acquisition is completed, and the calculation module performs the positioning calculation of the sandbox and calculates and locates the pouring cup mouth data of the sandbox.

6. The sandbox pouring positioning system based on three-dimensional point cloud according to claim 5, It is characterized in that The step of integrating the two-dimensional profile to form a three-dimensional profile specifically includes: When the rising edge of the area contour of the sandbox is triggered, the position of the casting trolley fed back by the laser rangefinder is recorded, and the single-frame contour and the value of the incremental encoder are collected. The single-frame contour is the yoz cross-sectional contour scatter point. The measured value of the laser rangefinder at this time is used as the x-direction information of the single-frame contour, combined with the y-direction information and the z-direction information in the single-frame contour, and converted into a three-dimensional coordinate system point; In the contour after the rising edge of the area contour of the sandbox is triggered, the detection value of the contour and the incremental encoder is recorded, and the increment in the x direction of the current contour is calculated by combining the detection value of the incremental encoder with the value of the incremental encoder of the starting frame, and the measurement value of the laser rangefinder according to the starting frame is converted into the global x coordinate of the system, and then converted into a three-dimensional coordinate system point; As the pouring trolley moves, multiple frames of contours are continuously stored in the sandbox three-dimensional model. When the falling edge trigger of the sandbox arrives, contour collection is stopped and the sandbox area contour collection is completed.

7. The sandbox pouring positioning system based on three-dimensional point cloud according to claim 5, It is characterized in that The calculation and positioning of the pouring cup mouth data of the sandbox specifically includes: S1: Rough cutting, cutting and removing invalid areas according to the height range of the sandbox area, wherein the invalid areas include the ground area; S2: downsampling processing, downsampling in the height direction; S3: Point cloud segmentation: the point cloud data is collected and cut into a single sub-area, and whether there is a sandbox area in the point cloud data segment is detected according to the size of the sandbox area. If there is no sandbox area, contour detection is continued. If there is a sandbox area, the average height of the sandbox area is calculated as the height of the sandbox. S4: Segment the original dense point cloud based on the sandbox height obtained by detection to obtain the pouring cup mouth point cloud area and the upper surface point cloud area; S5: segment the point cloud connected domain according to the point cloud area on the upper surface of the sandbox; S6: Determine whether there is an area with the same size as the pouring cup mouth in the area segmented in S5. If so, use the area as the sandbox pouring cup mouth detection area to obtain the range of the circumscribed rectangle of the pouring cup mouth area; S7: If the pouring cup mouth area is detected in S6, the upper point cloud of the pouring cup mouth is segmented based on the maximum value range of the x-direction and y-direction and the height in the z-direction of the circumscribed rectangle of the pouring cup mouth area, and the maximum height Z of the pouring cup mouth is calculated. cup , and calculate the center point of the pouring cup area in the x and y directions as the X cup and Y cup ; S8: If the pouring cup mouth area is not detected in S6, the point cloud connected domain segmentation is performed on the area again, and the segmentation judgment is performed on the segmented sub-areas; if the number of areas in the segmented area that are consistent with the pouring cup mouth size is 0, it means that the sandbox pouring cup mouth area is not detected; if the number of areas in the segmented area that are consistent with the pouring cup mouth size is 1, S7 is executed to complete the pouring cup mouth positioning calculation.

8. The sandbox pouring positioning system based on three-dimensional point cloud according to claim 7, It is characterized in that If the number of regions in the segmented area that are consistent with the size of the pouring cup mouth is greater than 1, the segmented areas are defined as to-be-processed segmented areas, and the multiple to-be-processed segmented areas are projected onto the plane, wherein the x and y directions correspond to the col_index and row index of the image respectively. After the segmented areas are projected, component segmentation of the connected domain of the two-dimensional image is performed to determine whether the roundness of the segmented components is within a set range. If the roundness of a component in the component segmentation result corresponding to a segmented area to be processed is within a set threshold range, the segmented area to be processed is determined to be a sandbox pouring cup mouth area, and S7 is executed to complete the sandbox pouring cup mouth positioning. If the roundness of the components in the component segmentation results corresponding to all the segmented areas to be processed is not within the set threshold range, it means that all the segmented areas to be processed are not pouring cup mouth areas, and contour detection continues.

Images