Method for determining the state of material in material unloading and control method for a material unloading drum

Abstract

The application provides a method for determining the state of material in material unloading and a control method for a material unloading cylinder. The determining method comprises the following steps: acquiring target carriage area scanning point cloud data collected by a point cloud data collection device at the current position of the material unloading cylinder; performing semantic segmentation on the scanning point cloud data to obtain material slope surface point cloud and carriage wall point cloud; performing angle normalization processing on the material slope surface point cloud, extracting the farthest effective points in each direction to form a material cross-section contour point set, and performing curve fitting to obtain a material cross-section contour; performing extreme value analysis based on the material cross-section contour to determine the height of the highest point of the material, and combining the material cross-section contour and the height of the highest point of the material to determine the material state determination result. The method can realize real-time and accurate identification of the material accumulation state in the carriage.

Description

Technical Field

[0001] This application relates to the field of material unloading technology, specifically to a method for determining the state of materials during material unloading and a control method for a material unloading cylinder. Background Technology

[0002] During grain harvesting, transshipment, and loading operations, unloading hoppers are typically used to transport materials such as grain into the cargo compartments of transport vehicles. As the unloading process continues, the material inside the compartment gradually accumulates, forming a pile with a certain slope. Due to factors such as vehicle parking posture, ground undulations, driving conditions, changes in unloading flow rate, and differences in the material's own fluidity, the shape and height of the pile will constantly change. If the unloading hopper is aimed at the same area for an extended period, the material height in some areas can easily rise rapidly and approach the upper edge of the cargo compartment wall, leading to spillage. This not only causes material loss but also affects operational efficiency and increases cleanup costs. In existing technologies, one method relies primarily on the driver or operator to visually observe and judge the loading situation inside the cargo compartment and manually adjust the position of the unloading hopper. This method is greatly affected by factors such as dust obstruction, lighting conditions, observation angle, and human reaction speed, making it difficult to timely and accurately grasp the true accumulation state of the material inside the cargo compartment. Another approach uses a preset trajectory, fixed time interval, or fixed angle switching strategy to control the movement of the unloading hopper. Although this reduces manual intervention to some extent, it still easily leads to problems such as switching too early, resulting in a decrease in loading efficiency, or switching too late, resulting in localized overflow, due to the lack of real-time perception of the actual material accumulation outline and height changes in the hopper. Summary of the Invention

[0003] The purpose of this application is to provide a method for determining the state of materials during material unloading and a control method for a material unloading cylinder.

[0004] To achieve the above objectives, the first aspect of this application provides a method for determining the state of materials during material unloading, the method comprising: Acquire scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder; Semantic segmentation is performed on the scanned point cloud data to obtain the point cloud of the material slope and the point cloud of the car wall inside the target car; Angle normalization processing is performed on the point cloud of the material slope. Extract the farthest effective points in each direction of the material slope point cloud to form a point set of the material cross-sectional profile; Curve fitting is performed based on the set of material cross-sectional profile points to obtain the material cross-sectional profile. Extreme value analysis is performed based on the cross-sectional profile of the material to obtain the highest point on the cross-sectional profile, and the height corresponding to the highest point is taken as the height of the highest point of the material. The material state determination result is based on the material cross-sectional profile and the height of the highest point of the material.

[0005] In this embodiment of the application, before performing semantic segmentation on the scanned point cloud data, the following steps are also included: Convert the polar coordinate point cloud data acquired by the point cloud data acquisition device into planar Cartesian coordinate point cloud data; Remove outliers that exceed the preset material unloading operation range; Dynamic occlusion filtering is applied to points of abrupt changes in reflection intensity and / or points of abrupt changes in distance. The initial ground height of the carriage is collected to establish a ground height reference benchmark. Based on the ground height reference benchmark, the scanned point cloud data is subjected to ground reference normalization processing.

[0006] In this embodiment of the application, the scanned point cloud data includes a series of consecutive scanned point clouds accumulated in time sequence. The series of consecutive scanned point clouds are converted into planar Cartesian coordinate point cloud data respectively, and compensated and projected onto the same coordinate system to form enhanced point cloud data for subsequent recognition.

[0007] In this embodiment of the application, semantic segmentation is performed on the scanned point cloud data to obtain the point cloud of the material slope inside the target carriage and the point cloud of the carriage wall, including: The scanning point cloud data is classified and extracted using either a point cloud density-based clustering method or a point cloud neural network-based point cloud semantic segmentation method to obtain the material slope point cloud and the point cloud of the carriage wall inside the target carriage.

[0008] A second aspect of this application proposes a control method for a material unloading cylinder, wherein the material unloading cylinder is equipped with a point cloud data acquisition device, and the control method includes: Acquire scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder; Semantic segmentation is performed on the scanned point cloud data to obtain the point cloud of the material slope and the point cloud of the car wall inside the target car; Angle normalization processing is performed on the point cloud of the material slope. Extract the farthest effective points in each direction of the material slope point cloud to form a point set of the material cross-sectional profile; Curve fitting is performed based on the set of material cross-sectional profile points to obtain the material cross-sectional profile. Extreme value analysis is performed based on the cross-sectional profile of the material to obtain the highest point on the cross-sectional profile, and the height corresponding to the highest point is determined as the height of the highest point of the material. Determine the height of the carriage wall based on the point cloud of the carriage wall; Determine the height difference between the highest point of the material and the height of the carriage wall, and compare the height difference with a preset safety threshold; If the height difference exceeds the preset safety threshold, control the material unloading cylinder to unload materials at the current position; If the height difference is less than or equal to the preset safety threshold, control the material unloading cylinder to move to the next position and return to execute the step of acquiring the scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder.

[0009] In this embodiment of the application, the method further includes: Real-time acquisition of the car tilt angle and material unloading flow rate at the current position of the material unloading cylinder; The preset safety threshold is determined based on the real-time acquired tilt angle of the carriage and the material unloading flow rate.

[0010] In this embodiment of the application, when the height difference is less than or equal to a preset safety threshold, controlling the material unloading cylinder to move to the next position includes: The height difference, preset safety threshold, real-time car tilt angle, real-time vehicle speed, real-time point cloud density change rate, current angle of the material unloading cylinder, and last search direction are input into the reinforcement learning decision model to output the target movement direction and target movement step size of the material unloading cylinder. Based on the motion delay and oscillation inertia during the movement of the material unloading cylinder, motion inertia compensation is performed on the material dropping position corresponding to the target movement direction and the target movement step length to obtain the next position after compensation. Control the material unloading cylinder to move to the next position after compensation.

[0011] In this embodiment, the preset security threshold is calculated according to the following formula: , Where p is the preset safety threshold, p0 is the basic threshold, θ is the real-time acquired car tilt angle, q is the real-time acquired material unloading flow rate, and k1 and k2 are weighting coefficients.

[0012] In this embodiment of the application, the method further includes: Real-time determination of material flowability based on point cloud density variation characteristics; If the material flowability is less than the preset flowability, the preset safety threshold will be adjusted to the second preset safety threshold. When the material flowability is greater than or equal to the preset flowability, the preset safety threshold remains the preset safety threshold determined based on the car tilt angle and the material unloading flow rate. The second preset security threshold is less than the preset security threshold.

[0013] A third aspect of this application discloses a material unloading system, comprising: Material unloading cylinder; A point cloud data acquisition device is located at the end of the material unloading cylinder and is used to acquire scanned point cloud data of the target carriage area when the material unloading cylinder is at the target position. The processing module performs semantic segmentation on the scanned point cloud data to obtain the material slope point cloud and the car wall point cloud inside the target car. It performs angle normalization on the material slope point cloud to extract the farthest effective points in each direction of the material slope point cloud, forming a material cross-sectional contour point set. Based on the material cross-sectional contour point set, it performs curve fitting to obtain the material cross-sectional contour. Based on the material cross-sectional contour, it performs extreme value analysis to obtain the highest point on the cross-sectional contour and determines the height corresponding to the highest point as the material's highest point height. It determines the car wall height based on the car wall point cloud, and then determines the height difference between the material's highest point height and the car wall height. Finally, it compares the height difference with a preset safety threshold. The control module is used to control the material unloading cylinder to unload materials at the current position when the height difference is greater than the preset safety threshold, and to control the material unloading cylinder to move to the next position when the height difference is less than or equal to the preset safety threshold, so as to continue to collect scanning point cloud data for the target carriage area.

[0014] A fourth aspect of this application provides a control device for a material unloading cylinder, comprising: The memory is configured to store instructions; The processor is configured to retrieve instructions from memory and, when executing the instructions, implement a control method for the material unloading cylinder.

[0015] A fifth aspect of this application provides a material unloading vehicle, the material unloading vehicle comprising: Material unloading cylinder; Point cloud data acquisition device, located at the end of the material unloading cylinder; and Control device for material unloading cylinder.

[0016] A sixth aspect of this application provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform a control method for a material unloading cylinder.

[0017] This application proposes a control method for a material unloading cylinder. The method involves acquiring scanned point cloud data of the target carriage area from a point cloud data acquisition device at the current position of the material unloading cylinder; performing semantic segmentation on the scanned point cloud data to obtain point clouds of the material slope and carriage walls; performing angle normalization on the material slope point cloud to extract the farthest effective points in each direction to form a material cross-sectional profile point set, and then performing curve fitting to obtain the material cross-sectional profile; performing extreme value analysis based on the material cross-sectional profile to determine the height of the highest point of the material; and combining the material cross-sectional profile and the height of the highest point to determine the material state. This method enables real-time and accurate identification of the material accumulation state within the carriage.

[0018] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 The illustration shows a flowchart of a method for determining the state of materials during material unloading according to an embodiment of this application; Figure 2 A schematic flowchart of a control method for a material unloading cylinder according to an embodiment of this application is shown. Figure 3 The schematic diagram illustrates the steps of a control method for a material unloading cylinder according to an embodiment of this application; Figure 4 The diagram illustrates the internal structure of a computer device according to an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0021] Figure 1 This illustration schematically depicts a flowchart of a method for determining the state of materials during material unloading according to an embodiment of this application. For example... Figure 1 As shown in one embodiment of this application, a method for determining the state of materials during material unloading is provided, comprising the following steps: Step 102: Obtain the scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder; Step 104: Perform semantic segmentation on the scanned point cloud data to obtain the point cloud of the material slope inside the target carriage and the point cloud of the carriage wall; Step 106: Perform angle normalization processing on the point cloud of the material slope; Step 108: Extract the farthest effective points in each direction of the material slope point cloud to form a material cross-sectional profile point set; Step 110: Perform curve fitting based on the material cross-sectional profile point set to obtain the material cross-sectional profile; Step 112: Perform extreme value analysis based on the cross-sectional profile of the material to obtain the highest point on the cross-sectional profile, and take the height corresponding to the highest point as the height of the highest point of the material. Step 114: Determine the material state determination result based on the material cross-sectional profile and the height of the highest point of the material.

[0022] In one embodiment, a point cloud data acquisition device is installed at the end of the material unloading cylinder. This device scans the target carriage area and outputs scanned point cloud data when the material unloading cylinder is in its current position. After acquiring the scanned point cloud data corresponding to the current position, the control system performs semantic segmentation on the scanned point cloud data, thereby distinguishing the point cloud within the target carriage into material slope point cloud and carriage wall point cloud. After obtaining the material slope point cloud, angle normalization processing is performed on the material slope point cloud to reduce the impact of changes in the material unloading cylinder's attitude, scanning angle, or vehicle attitude on the subsequent contour extraction results. After completing the angle normalization processing, the farthest effective points in each direction of the material slope point cloud are extracted, and a material cross-sectional contour point set is formed based on these farthest effective points. Subsequently, curve fitting is performed based on the material cross-sectional contour point set to obtain a continuous material cross-sectional contour. Further, extreme value analysis is performed based on the material cross-sectional contour to obtain the highest point on the cross-sectional contour, and the height corresponding to the highest point is determined as the material's highest point height. Finally, the material state determination result is determined based on the material cross-sectional contour and the material's highest point height. In this embodiment, the material state determination result can be used to characterize the material accumulation pattern, stack height level, or overflow risk status within the current target carriage area. Since the scanned point cloud data typically includes carriage wall boundaries, material accumulation surfaces, and some stray points in the environment, extracting the material slope point cloud through semantic segmentation helps separate the point cloud that truly reflects the material accumulation pattern from the complex scene, thereby improving the accuracy of subsequent contour construction. After angle normalization processing of the material slope point cloud, point cloud data obtained under different scanning postures can be analyzed under a unified angle reference, avoiding contour distortion caused by scanning direction deviation. Constructing a material cross-sectional contour point set based on the farthest effective points in each direction can accurately reflect the changes in the outer boundary of the material slope. Then, reconstructing the discrete point set into a continuous contour through curve fitting can reduce the impact of point cloud noise and local missing data on stack height judgment. By performing extreme value analysis on the fitted material cross-sectional contour, the height of the highest point of the material can be quickly obtained, and the material state determination result can be given in combination with the overall shape of the material cross-sectional contour.

[0023] In one embodiment, the control system continuously acquires scanned point cloud data output by the point cloud data acquisition device during material unloading, and repeatedly performs processing steps such as semantic segmentation, angle normalization, contour point extraction, curve fitting, and extreme value analysis according to the scanning area of ​​the carriage corresponding to the current position. As material continuously falls into the carriage, the cross-sectional contour of the material will dynamically change, and the height of the highest point will also change accordingly. Based on the cross-sectional contour of the material and the height of the highest point of the material at different time points, the control system can dynamically update the material status determination results, thereby providing a basis for subsequent unloading control, position switching, or overflow warning.

[0024] In one embodiment, the control system preprocesses the point cloud data before semantic segmentation. The raw point cloud data output by the point cloud data acquisition device can be polar coordinate point cloud data, which includes distance, angle, and reflection intensity information. To facilitate subsequent spatial analysis and geometric processing, the control system first converts the polar coordinate point cloud data into planar Cartesian coordinate point cloud data, enabling the point cloud to be represented in a unified planar coordinate system for positional and contour analysis. After coordinate transformation, the control system removes outliers that exceed the preset material unloading operation range to eliminate invalid point clouds from the background area outside the vehicle, the equipment structure area, or distant stray areas, thereby improving the effectiveness of the point cloud data. In this embodiment, since the material unloading operation is usually accompanied by dust, material falling and obstruction, and equipment vibration, there may be abrupt changes in reflection intensity and distance jumps in the scanned point cloud data. To reduce the interference of these outliers on subsequent semantic segmentation and contour extraction, the control system performs dynamic occlusion filtering on abrupt changes in reflection intensity and / or distance jumps. Dynamic occlusion filtering can filter out, reduce weight, or smooth out abnormal points based on changes in reflection intensity between adjacent points, changes in distance continuity, or neighborhood consistency characteristics in the scanning sequence. Through dynamic occlusion filtering, the impact of dust noise, occlusion edges, and instantaneous false detections can be reduced, making the retained point cloud closer to the actual boundaries of the carriage walls and material slopes. To ensure height analysis is based on a unified reference surface, the control system also collects the initial carriage ground height and constructs a ground height reference benchmark based on it. The initial carriage ground height can be obtained when the carriage is empty or through point cloud identification before material covers the ground. Based on the ground height reference benchmark, the control system performs ground reference normalization processing on the scanned point cloud data, ensuring that the point cloud height information obtained at different times and in different postures is uniformly expressed relative to the same ground reference surface.

[0025] In one embodiment, the scanned point cloud data includes consecutive multi-frame scanned point clouds accumulated sequentially. Since the point cloud data acquisition device uses planar scanning, single-frame scanned point clouds may exhibit sparseness, partial missing points, or discontinuous boundaries due to dust obstruction, material flow, equipment shaking, or changes in vehicle posture. To improve the stability of subsequent identification, the control system uses consecutive multi-frame scanned point clouds for fusion processing. Specifically, the control system converts consecutive multi-frame scanned point clouds into planar Cartesian coordinate point cloud data, ensuring that each frame is processed under the same coordinate representation. Subsequently, the converted consecutive multi-frame point clouds undergo compensating projection, aligning each frame to the same coordinate system, thereby forming enhanced point cloud data for subsequent identification. Compensating projection can eliminate coordinate deviations caused by material unloading cylinder movement, slight vehicle displacement, vehicle tilt changes, or changes in the installation posture of the point cloud data acquisition device during multi-frame sampling. In other words, although point clouds acquired at different times originate from the same target vehicle area, spatial offsets may exist between frames due to changes in acquisition conditions. By using compensated projection, historical frame point clouds can be mapped onto the current reference coordinate system, allowing point cloud data obtained at different times to be correctly superimposed without ghosting or misalignment at the boundaries of the carriage wall or material slope. The resulting enhanced point cloud data has higher point density and more complete boundary contours compared to single-frame point cloud data.

[0026] In one embodiment, semantic segmentation of the scanned point cloud data yields the point cloud of the material slope and the point cloud of the carriage wall within the target carriage. This can be achieved using either a point cloud density-based clustering method or a point cloud semantic segmentation method based on a point cloud neural network. The control system can select a suitable segmentation method based on the actual operating environment, processor computing power, and point cloud quality. When using a point cloud density-based clustering method, the control system clusters the point cloud based on the spatial distance relationships and local density distribution among the points in the scanned point cloud data, grouping spatially close and continuously distributed points into the same category. Subsequently, the clustered point clouds are further classified and extracted by combining the geometric distribution characteristics, boundary orientation characteristics, and height variation characteristics of each category of region, thus obtaining the point cloud of the material slope and the point cloud of the carriage wall within the target carriage. The point cloud of the carriage wall typically extends along the carriage boundary, characterized by relatively regular boundaries, strong continuity, and a relatively concentrated height distribution. The point cloud of the material slope, usually located inside the carriage, exhibits a sloping or bulging shape formed by changes in the stacking state, with its contour curvature and height variation trends significantly different from those of the carriage wall. Clustering methods based on point cloud density can first extract effective points from the complex background, and then use geometric features to distinguish between the material slope and the carriage wall, thus being suitable for application scenarios with relatively obvious point cloud structural features and limited computing resources. When using a point cloud semantic segmentation method based on point cloud neural networks, the control system inputs scanned point cloud data into the point cloud semantic segmentation model, which outputs a category label for each point to achieve point-level classification. The model input can include the point's location coordinates, distance information, reflection intensity information, and neighborhood geometric features; the model output includes at least the material slope category and the carriage wall category. Point cloud semantic segmentation methods based on point cloud neural networks can make full use of the local structural features and global distribution features of point clouds to perform more refined classification of point clouds under complex working conditions. They are especially suitable for scenarios with large amounts of dust, noise, or unclear boundary features.

[0027] like Figure 2 As shown, in one embodiment, this application also proposes a control method for a material unloading cylinder, comprising: Step 202: Obtain the scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder; Step 204: Perform semantic segmentation on the scanned point cloud data to obtain the point cloud of the material slope inside the target carriage and the point cloud of the carriage wall; Step 206: Perform angle normalization processing on the point cloud of the material slope; Step 208: Extract the farthest effective points in each direction of the material slope point cloud to form a material cross-sectional profile point set; Step 210: Perform curve fitting based on the material cross-sectional profile point set to obtain the material cross-sectional profile; Step 212: Perform extreme value analysis based on the cross-sectional profile of the material to obtain the highest point on the cross-sectional profile, and take the height corresponding to the highest point as the height of the highest point of the material. Step 214: Determine the height of the carriage wall based on the point cloud of the carriage wall; Step 216: Determine the height difference between the highest point of the material and the height of the carriage wall, and compare the height difference with a preset safety threshold. Step 218: If the height difference is greater than the preset safety threshold, control the material unloading cylinder to unload material at the current position; Step 220: If the height difference is less than or equal to the preset safety threshold, control the material unloading cylinder to move to the next position and return to the step of acquiring the scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder.

[0028] In one embodiment, the material unloading cylinder is equipped with a point cloud data acquisition device installed at the end of the material unloading cylinder. This device scans the target carriage area and outputs scanned point cloud data when the material unloading cylinder is in its current position. After acquiring the scanned point cloud data corresponding to the current position, the control system performs semantic segmentation on the scanned point cloud data, thereby dividing the point cloud within the target carriage into material slope point clouds and carriage wall point clouds. Subsequently, the material slope point cloud undergoes angle normalization processing to reduce the impact of material unloading cylinder posture changes, slight vehicle shaking, and scanning angle differences on the contour extraction results. After angle normalization, the farthest effective points in each direction of the material slope point cloud are extracted, and a material cross-sectional contour point set is formed based on these farthest effective points. Further, curve fitting is performed based on the material cross-sectional contour point set to obtain a continuous material cross-sectional contour. Then, extreme value analysis is performed based on the material cross-sectional contour to obtain the highest point on the cross-sectional contour, and the height corresponding to the highest point is determined as the highest point height of the material. Simultaneously, the height of the carriage wall is determined based on the point cloud data, and the height difference between the highest point of the material and the height of the carriage wall is determined accordingly. The control system compares the height difference with a preset safety threshold. If the height difference is greater than the preset safety threshold, the material unloading cylinder is controlled to unload the material at the current position; if the height difference is less than or equal to the preset safety threshold, the material unloading cylinder is controlled to move to the next position and return to execute the scanning point cloud data acquisition step at the current position, thereby re-perceiving and judging at the new position. In this embodiment, the material unloading cylinder does not mechanically swing according to a fixed time or fixed trajectory during operation, but decides whether to continue unloading at the current position in real time based on the material accumulation state in the target carriage area. That is to say, the control system continuously senses the material slope shape and carriage wall boundary position in the carriage through the point cloud data acquisition device, and judges whether there is still sufficient safety margin at the current position based on the height difference between the highest point of the material and the height of the carriage wall. When the height difference is large, it indicates that there is still considerable space between the current unloading position and the top edge of the truck bed, and continuing to unload at the current position can improve loading efficiency. When the height difference is small, it indicates that the material pile in the current unloading area is already close to the top edge of the truck bed, posing a risk of overflow. In this case, by controlling the material unloading cylinder to move to the next position, subsequent unloading can be transferred to areas with lower pile heights within the truck bed, avoiding excessive local accumulation. The material cross-sectional profile not only reflects the current highest accumulation position but also the overall undulation of the material surface. The control system makes control decisions based on both the material cross-sectional profile and the height of the highest point of the material, which can more accurately reflect the accumulation risk than a judgment based solely on a single height value. For example, when the material surface has local peaks but the overall distribution is relatively flat, the control strategy can combine the profile trend and height difference to comprehensively determine whether to move the position; when the material surface forms a steep slope and the highest point rapidly approaches the height of the truck bed wall, the unloading position can be switched more promptly.By constructing a material unloading control process based on point cloud perception, contour reconstruction, and height difference comparison, it is possible to achieve real-time monitoring of the material stacking height in the carriage and dynamic adjustment of the unloading position, thereby improving the control accuracy under complex working conditions.

[0029] In one embodiment, the control system also acquires the tilt angle of the unloading cylinder and the unloading flow rate of the carriage at the current position in real time, and determines a preset safety threshold based on the acquired tilt angle and unloading flow rate. That is, the preset safety threshold is not a fixed constant, but can be dynamically adjusted according to changes in vehicle posture and unloading conditions. By introducing the tilt angle and unloading flow rate as two parameters, the control system can dynamically determine a safety margin requirement more suitable for the current scenario based on the actual operating conditions. The tilt angle reflects the current posture change of the vehicle or carriage. For example, when the vehicle is parked on a slope, soft ground, or in an area with local elevation differences, the carriage will tilt forward, backward, or sideways relative to the horizontal plane. Carriage tilt changes the material accumulation pattern and the positional relationship of the highest point of the material relative to the carriage wall, resulting in different overflow risks for the same material height under different tilt angles. The unloading flow rate reflects the amount of material falling into the carriage per unit time. The larger the unloading flow rate, the faster the local accumulation rate increases, and the more timely the response of the control system to position changes is required. Therefore, when the car body tilts at a large angle or the material unloading flow rate is high, the control system can appropriately increase the safety margin requirement to trigger position adjustment earlier; when the car body is relatively stable and the unloading flow rate is low, the safety margin judgment conditions can be appropriately relaxed to improve unloading efficiency.

[0030] In one embodiment, when the height difference is less than or equal to a preset safety threshold, the control system moves the material unloading cylinder to the next position. When the safety threshold is not met and the material unloading cylinder needs to be moved to the next position, the control system introduces an intelligent decision-making and inertial compensation mechanism, enabling the material unloading cylinder to find a safe landing point more quickly under complex working conditions and reducing the risk of material spillage. In this embodiment, the control system uses the height difference between the highest point of the material and the height of the truck bed wall. Preset safety threshold Real-time acquisition of carriage tilt angle Real-time vehicle speed Real-time acquisition of point cloud density change rate Current angle of material unloading cylinder Last search direction indicator The input is a deep reinforcement learning decision model, which uses a DQN structure. The input-output relationship can be represented as follows:

[0031] in represents a deep neural network, represents an action in the action set, represents the state, such as maintaining the current angle, slightly adjusting by 2°, moderately adjusting by 5°, largely adjusting by 10°, or reversing the search direction. The decision-making cycle can be 500 ms. The decision-making device selects the optimal action in the preset action set every cycle. When the decision result is an angle adjustment action, the control system converts this action into the target angle corresponding to the next position of the material unloading cylinder. and drives the angle actuator to complete the movement of the material unloading cylinder; when the decision result is a holding action, the material unloading cylinder maintains the current position and enters the next round of point cloud acquisition and determination. In this embodiment, based on the vibration, inertia, and attitude changes of the material unloading vehicle during material unloading, the control system introduces motion inertia compensation before performing angle adjustment, and the compensation angle is jointly determined by the vehicle driving speed and the tilt angle change rate, and the calculation formula can be: , is the tilt angle change rate, ; >is the time interval between two adjacent tilt angle samplings, is the current carriage tilt angle, is the tilt angle of the previous control cycle, and the final execution angle satisfies , and at the same time, safety verification and amplitude limit constraints are added to make the single execution angle change satisfy , is the execution angle of the previous cycle. By superimposing the compensation angle on the policy angle and restricting the change amplitude of the execution angle, the risk of misjudgment caused by landing point deviation and material overflow can be reduced during vehicle acceleration and deceleration, vehicle body shaking, or slope operation, and the risk of large-amplitude and frequent jitter of the material unloading cylinder can be avoided.

[0032] In one embodiment, a hierarchical material unloading control strategy is executed according to the relationship between the height difference and the safety threshold. For example, when h≥p, material unloading is maintained; when 0.8p≤h<p, decelerated material unloading is triggered, and the material unloading flow rate is reduced to 50% of the normal flow rate; when h<0.8p, material unloading is stopped and the position adjustment of the material unloading cylinder is triggered. During the stage of stopping material unloading and triggering position adjustment, the decision-making model outputs an angle adjustment action, and the control system drives the material unloading cylinder to move to the next position after performing inertia compensation and amplitude limit, and then obtains the point cloud data again and updates the height difference judgment result.

[0033] In one embodiment, the detection frequency, angle adjustment range, and material unloading flow control method are dynamically adjusted according to the risk level, thereby improving the response speed when the height difference approaches the safety line. For example, in the free operation stage, which corresponds to a height difference exceeding the safety threshold by more than 20%, the detection frequency is once per second, allowing the material unloading cylinder to make relatively large adjustments within a 10° range; in the early warning stage, which corresponds to a height difference approaching the safety line, the detection frequency is increased to twice per second, while a 3-second trend prediction is enabled, and the maximum adjustment range is reduced to 8°; in the critical stage, which corresponds to a height difference touching the safety line, the detection frequency is increased to three times per second, combined with 5°-level fine adjustments, and the linear decay control of the material unloading flow is simultaneously initiated; in the emergency stage, which corresponds to a height difference exceeding the safety line, five monitoring cycles per second are triggered, and a rapid scanning mode is initiated while suspending material unloading, quickly locking in a new safe landing point through a priority compensation mechanism.

[0034] In one embodiment, to enable the safety threshold to dynamically adjust with changes in the carriage's posture and material unloading conditions, the control system uses a formula to calculate the preset safety threshold. When the material unloading cylinder is in its current position, the control system acquires the carriage's tilt angle and material unloading flow rate in real time, and calculates the preset safety threshold using the following formula: , Where p is the preset safety threshold, p0 is the basic threshold, θ is the real-time acquired car tilt angle, q is the real-time acquired material unloading flow rate, and k1 and k2 are weighting coefficients.

[0035] In this embodiment, the basic threshold p0 is used to characterize the basic safety margin when the carriage is in a horizontal state and the material unloading flow rate is at a normal level. The basic threshold can be calibrated according to the carriage structural dimensions, the installation height of the point cloud data acquisition device at the end of the material unloading cylinder, and the desired overflow safety margin. For example, when it is desired to maintain a certain distance as a safety margin between the upper edge of the carriage and the allowable stacking height of the material, this distance can be set as the basic threshold p0. The carriage tilt angle θ is used to reflect the degree of tilt of the vehicle in scenarios such as slopes, undulating fields, or uneven roads. The larger the tilt angle, the easier it is for the material to accumulate on the lower side, the local stacking height approaches the upper edge of the carriage faster, and the effective safety margin in the carriage is more unevenly distributed at different locations. By introducing the k1·θ term in the threshold calculation, the preset safety threshold p increases accordingly when the tilt angle increases, thereby triggering position adjustment or stopping material unloading in advance under tilted conditions, reducing the risk of material overflow. The weighting coefficient k1 is used to adjust the degree of influence of the tilt angle on the safety threshold. k1 can be obtained through experimental calibration or set according to the carriage width, material stacking angle, and the stacking height variation under tilted conditions. The material unloading flow rate q reflects the speed at which material enters the carriage per unit time. A higher unloading flow rate leads to faster material height growth, potentially causing the stacking height to rapidly approach the top edge of the carriage within a single sensing and decision-making cycle of the control system. By introducing the term k2·q into the threshold calculation, the preset safety threshold p increases accordingly with increasing material unloading flow rate, thus reserving a larger safety margin under high flow conditions and reducing the risk of lag due to system response cycles. The weighting coefficient k2 adjusts the influence of the material unloading flow rate on the safety threshold. k2 can be obtained through material unloading tests at different flow rates, or determined by establishing a mapping relationship between material unloading speed, control cycle, and stacking height growth rate. For example, in a specific example, the basic threshold p0 is 25 cm, k1 is 0.3, and k2 is 0.05. When the real-time carriage tilt angle θ is 10° and the real-time material unloading flow rate q is 50, the preset safety threshold p = 25 + 0.3 × 10 + 0.05 × 50 = 25 + 3 + 2.5 = 30.5 cm. At this point, the control system compares the height difference with 30.5 cm. Compared with a fixed threshold, it can identify the risk of material overflow caused by tilting and high flow rate earlier, and control the material unloading cylinder to switch positions or stop material unloading in a timely manner.

[0036] In one embodiment, to further adapt to the differences in the accumulation behavior of different materials during the unloading process, the control system, in addition to considering the influence of the carriage tilt angle and material unloading flow rate on the safety threshold, also determines the material's flowability in real time based on the point cloud density change characteristics, and adjusts the safety threshold accordingly. Different materials or materials with different moisture contents exhibit different flowability: when flowability is high, the material is more likely to spread out quickly and flow to lower areas, resulting in rapid changes in material shape and a localized pile height that may quickly approach the top edge of the carriage; when flowability is low, the material is more likely to form a more stable accumulation shape, with a relatively slow change in pile height. Point cloud density change characteristics are used to characterize the rate of change of material shape over a short period. The control system, in continuously collected point cloud data, statistically analyzes the number of points, point distribution density, or density change over time within a preset area on the material slope point cloud, and infers the material's flowability based on this change. For example, within the same control cycle, if the density distribution of the point cloud on the material slope changes rapidly, the density peak position shifts rapidly, or the point cloud density change rate remains above a preset range, the material can be determined to be in a high-flowability state. Conversely, if the point cloud density distribution changes slowly, the density change rate is low, and the outline shape is stable, the material can be determined to be in a low-flowability state. In this way, the material flowability can be estimated in real time using the dynamic characteristics of the point cloud itself without the need for additional sensors. After determining the material flowability, the control system compares this flowability with a preset flowability and selects different safety threshold strategies accordingly. When the material flowability is less than the preset flowability, the control system adjusts the preset safety threshold calculated based on the carriage tilt angle and material unloading flow rate to a second preset safety threshold, which is smaller than the first preset safety threshold. A smaller second preset safety threshold means that the control system allows the height difference to be closer to the upper edge of the carriage before triggering position adjustment, thereby reducing unnecessary frequent switching and improving the continuity and efficiency of material unloading when the material flowability is low and the stacking height changes slowly. Conversely, when the material flowability is greater than or equal to the preset flowability, the control system maintains the preset safety threshold calculated based on the real-time carriage tilt angle and real-time material unloading flow rate. This allows for a larger safety margin under high flowability conditions and timely triggering of anti-overflow control. The second preset safety threshold can be obtained by multiplying the preset safety threshold by a preset proportional coefficient or subtracting a preset compensation amount, or it can be directly selected based on the flowability level using a lookup table. For example, when the flowability is low, the second preset safety threshold can be set to 0.9 or 0.8 times the preset safety threshold to appropriately reduce the threshold; when the flowability is high, the original threshold remains unchanged to increase the safety margin. By determining the material flowability in real time based on the point cloud density change characteristics, and using a smaller second preset safety threshold when the flowability is low, and a preset safety threshold calculated based on the tilt angle and flow rate when the flowability is high, the material unloading cylinder control can balance anti-overflow safety and material unloading efficiency under different material conditions, reducing misjudgments and over-adjustments.

[0037] In one embodiment, a material unloading system is provided, including a material unloading cylinder, a point cloud data acquisition device, a processing module, and a control module. The point cloud data acquisition device is located at the end of the material unloading cylinder and is used to acquire scanned point cloud data of the target carriage area when the material unloading cylinder is at the target position. The processing module is connected to the point cloud data acquisition device and is used to receive and process the scanned point cloud data. Specifically, the processing module performs semantic segmentation on the scanned point cloud data to obtain the material slope point cloud and carriage wall point cloud within the target carriage; performs angle normalization on the material slope point cloud; extracts the farthest effective points in each direction of the material slope point cloud to form a material cross-sectional contour point set; performs curve fitting based on the material cross-sectional contour point set to obtain the material cross-sectional contour; performs extreme value analysis based on the material cross-sectional contour to obtain the highest point on the cross-sectional contour and determines the height corresponding to the highest point as the material's highest point height; determines the carriage wall height based on the carriage wall point cloud; further determines the height difference between the material's highest point height and the carriage wall height, and compares the height difference with a preset safety threshold. The control module is connected to the processing module and controls the material unloading cylinder based on the comparison results output by the processing module. When the height difference is greater than a preset safety threshold, the control module controls the material unloading cylinder to unload material at the current position; when the height difference is less than or equal to the preset safety threshold, the control module controls the material unloading cylinder to move to the next position to continue scanning point cloud data acquisition for the target carriage area. In other words, the material unloading system forms a closed-loop control process through the coordinated work of three parts: point cloud perception, data processing, and execution control. The point cloud data acquisition device is responsible for sensing the material accumulation state inside the carriage, the processing module is responsible for extracting material slope and carriage wall features from the point cloud data and calculating the height difference, and the control module is responsible for deciding whether to maintain the current unloading position or switch to the next position based on the comparison results.

[0038] For example, during grain unloading operations, the point cloud data acquisition device continuously collects scanned point cloud data of the current material unloading area inside the truck bed. The processing module identifies the boundaries of the truck bed walls and the material surface from this data and calculates the height difference between the highest point of the material and the height of the truck bed walls at the current position. If the comparison result indicates that there is still sufficient safety margin, the control module maintains the current position of the material unloading cylinder and continues unloading. If the comparison result indicates that the current safety margin is insufficient, the control module immediately drives the material unloading cylinder to switch to the next position, and the point cloud data acquisition device re-acquires point cloud data at the new position, entering the next round of processing and judgment. Through this modular system structure, automatic sensing, automatic judgment, and automatic position switching during the material unloading process can be achieved, solving the problems of high difficulty in manual monitoring and poor flexibility of traditional fixed strategies, thereby improving unloading uniformity, operational efficiency, and spill prevention safety.

[0039] This application also proposes a control device for a material unloading cylinder, comprising: The memory is configured to store instructions; The processor is configured to retrieve instructions from memory and, when executing the instructions, implement a control method for the material unloading cylinder.

[0040] This application also proposes a material unloading vehicle, which includes: a material unloading cylinder; A point cloud data acquisition device is located at the end of the material unloading cylinder; And a control device for the material unloading cylinder.

[0041] like Figure 3 As shown, in one embodiment, the automatic material unloading control process can be executed according to the steps corresponding to the flowchart. First, automatic material unloading is initiated, and the control system enters a cyclic detection state. In each control cycle, the control system acquires point cloud data of the target carriage area, calculates the height of the highest point of the material and the height of the carriage wall, and obtains the height difference h between the two. Then, the height difference h is compared with a preset safety threshold p: when the height difference h is greater than the preset safety threshold p, it indicates that there is still sufficient safety margin between the top of the material and the upper edge of the carriage. The material unloading cylinder maintains its current position and continues unloading. Simultaneously, the control system recalculates the height difference and repeats the comparison in the next cycle to achieve continuous monitoring. When the height difference h is less than or equal to the preset safety threshold p, it indicates insufficient safety margin and a risk of material overflow. The control system performs offset calculation and generates a material unloading cylinder movement command based on the current covered area of ​​the carriage, the target uncovered area, or a preset traversal order, driving the material unloading cylinder to move to the next position to change the landing point. After the material unloading cylinder completes its movement, the control system re-enters the cycle of height difference calculation and threshold comparison. In this embodiment, after generating the movement command, the control system also determines whether all areas of the carriage have been traversed: if not, it continues to execute the next round of height difference calculation, threshold comparison and necessary movement control until the material unloading coverage of the entire carriage is completed; if it has been traversed, it means that all places where materials can be unloaded have been unloaded, and then outputs the material unloading end command and ends the automatic material unloading process.

[0042] This application also proposes a storage medium on which a program is stored, which, when executed by a processor, implements the above-described control method for a material unloading cylinder.

[0043] Compared with existing technologies, this application proposes a method for determining the state of materials during unloading. This method involves acquiring scanned point cloud data of the target carriage area from a point cloud data acquisition device at the current position of the unloading cylinder, performing semantic segmentation on the scanned point cloud data to obtain point clouds of the material slope and carriage walls, further normalizing the angle of the material slope point cloud, extracting the farthest effective points in each direction to form a material cross-sectional contour point set, performing curve fitting based on the material cross-sectional contour point set to obtain the material cross-sectional contour, and then performing extreme value analysis based on the material cross-sectional contour to determine the height of the highest point of the material. The material state determination result is then determined by combining the material cross-sectional contour and the height of the highest point of the material. This allows for real-time and accurate identification of the material accumulation pattern and stacking height within the carriage, thereby solving the problem of untimely and inaccurate material state identification caused by reliance on manual observation or experience in existing technologies, and improving the automation and reliability of material state determination.

[0044] Figure 1 This is a flowchart illustrating a control method for a material unloading cylinder in one embodiment. It should be understood that, although... Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise explicitly stated herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0045] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 4As shown in the figure, the computer device includes a processor A01, a network interface A02, a display screen A04, an input device A05, and a memory (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01 and a computer program B02. The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A06. The network interface A02 is used for communication with external terminals via a network connection. When the computer program is executed by the processor A01, it implements a control method for a material unloading cylinder. The display screen A04 can be a liquid crystal display (LCD) or an e-ink display. The input device A05 can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0046] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0047] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0048] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0049] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0050] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0051] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0052] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0053] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0054] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0055] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for determining the state of materials during material unloading, characterized in that, The determination method includes: Acquire the scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder; Semantic segmentation is performed on the scanned point cloud data to obtain the point cloud of the material slope inside the target carriage and the point cloud of the carriage wall; The angle normalization process is performed on the point cloud of the material slope. Extract the farthest effective points in each direction from the point cloud of the material slope to form a point set of the material cross-section contour; Curve fitting is performed based on the set of material cross-sectional contour points to obtain the material cross-sectional contour. Based on the cross-sectional profile of the material, extreme value analysis is performed to obtain the highest point on the cross-sectional profile, and the height corresponding to the highest point is taken as the height of the highest point of the material. The material state determination result is based on the material cross-sectional profile and the height of the highest point of the material.

2. The method for determining the state of materials during material unloading according to claim 1, characterized in that, Before performing semantic segmentation on the scanned point cloud data, the following steps are also included: The polar coordinate point cloud data acquired by the point cloud data acquisition device is converted into planar Cartesian coordinate point cloud data; Remove outliers that exceed the preset material unloading operation range; Dynamic occlusion filtering is applied to points of abrupt changes in reflection intensity and / or points of abrupt changes in distance. The initial ground height of the carriage is collected to establish a ground height reference benchmark. Based on the ground height reference benchmark, the scanned point cloud data is subjected to ground reference normalization processing.

3. The method for determining the state of materials during material unloading according to claim 2, characterized in that, The scanned point cloud data includes multiple consecutive frames of scanned point clouds accumulated in time sequence. The multiple consecutive frames of scanned point clouds are converted into planar Cartesian coordinate point cloud data and compensated and projected onto the same coordinate system to form enhanced point cloud data for subsequent recognition.

4. The method for determining the state of materials during material unloading according to claim 1, characterized in that, Semantic segmentation was performed on the scanned point cloud data to obtain the point cloud of the material slope inside the target carriage and the point cloud of the carriage wall, including: The scanning point cloud data is classified and extracted using either a point cloud density-based clustering method or a point cloud neural network-based point cloud semantic segmentation method to obtain the material slope point cloud and the point cloud of the carriage wall inside the target carriage.

5. A control method for a material unloading cylinder, wherein the material unloading cylinder is equipped with a point cloud data acquisition device, characterized in that, The control method includes: Acquire scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder; Semantic segmentation is performed on the scanned point cloud data to obtain the point cloud of the material slope and the point cloud of the car wall inside the target car; Angle normalization processing is performed on the point cloud of the material slope. Extract the farthest effective points in each direction of the material slope point cloud to form a point set of the material cross-sectional profile; Curve fitting is performed based on the set of material cross-sectional profile points to obtain the material cross-sectional profile. Extreme value analysis is performed based on the cross-sectional profile of the material to obtain the highest point on the cross-sectional profile, and the height corresponding to the highest point is determined as the height of the highest point of the material. Determine the height of the carriage wall based on the point cloud of the carriage wall; Determine the height difference between the highest point of the material and the height of the carriage wall, and compare the height difference with a preset safety threshold; If the height difference exceeds the preset safety threshold, control the material unloading cylinder to unload materials at the current position; If the height difference is less than or equal to the preset safety threshold, control the material unloading cylinder to move to the next position and return to execute the step of acquiring the scanned point cloud data of the target carriage area collected by the point cloud data acquisition device at the current position of the material unloading cylinder.

6. The control method for a material unloading cylinder according to claim 5, characterized in that, The method further includes: Real-time acquisition of the car tilt angle and material unloading flow rate at the current position of the material unloading cylinder; The preset safety threshold is determined based on the real-time acquired tilt angle of the carriage and the material unloading flow rate.

7. The control method for a material unloading cylinder according to claim 5, characterized in that, When the height difference is less than or equal to a preset safety threshold, control the material unloading cylinder to move to the next position, including: The height difference, preset safety threshold, real-time car tilt angle, real-time vehicle speed, real-time point cloud density change rate, current angle of the material unloading cylinder, and last search direction are input into the reinforcement learning decision model to output the target movement direction and target movement step size of the material unloading cylinder. Based on the motion delay and oscillation inertia during the movement of the material unloading cylinder, motion inertia compensation is performed on the material dropping position corresponding to the target movement direction and the target movement step length to obtain the next position after compensation. Control the material unloading cylinder to move to the next position after compensation.

8. The control method for a material unloading cylinder according to claim 6, characterized in that, The preset security threshold is calculated using the following formula: ， Where p is the preset safety threshold, p0 is the basic threshold, θ is the real-time acquired car tilt angle, q is the real-time acquired material unloading flow rate, and k1 and k2 are weighting coefficients.

9. The control method for a material unloading cylinder according to claim 8, characterized in that, The method also includes: Real-time determination of material flowability based on point cloud density variation characteristics; If the material flowability is less than the preset flowability, the preset safety threshold will be adjusted to the second preset safety threshold. When the material flowability is greater than or equal to the preset flowability, the preset safety threshold remains the preset safety threshold determined based on the car tilt angle and the material unloading flow rate. Wherein, the second preset security threshold is less than the preset security threshold.

10. A material unloading system, characterized in that, include: Material unloading cylinder; A point cloud data acquisition device is located at the end of the material unloading cylinder and is used to acquire scanned point cloud data of the target carriage area when the material unloading cylinder is at the target position. The processing module performs semantic segmentation on the scanned point cloud data to obtain the material slope point cloud and the car wall point cloud inside the target car. It performs angle normalization on the material slope point cloud to extract the farthest effective points in each direction of the material slope point cloud, forming a material cross-sectional contour point set. Based on the material cross-sectional contour point set, it performs curve fitting to obtain the material cross-sectional contour. Based on the material cross-sectional contour, it performs extreme value analysis to obtain the highest point on the cross-sectional contour and determines the height corresponding to the highest point as the material's highest point height. It determines the car wall height based on the car wall point cloud, and then determines the height difference between the material's highest point height and the car wall height. Finally, it compares the height difference with a preset safety threshold. The control module is used to control the material unloading cylinder to unload materials at the current position when the height difference is greater than the preset safety threshold, and to control the material unloading cylinder to move to the next position when the height difference is less than or equal to the preset safety threshold, so as to continue to collect scanning point cloud data for the target carriage area.

11. A control device for a material unloading cylinder, characterized in that, include: The memory is configured to store instructions; The processor is configured to retrieve the instructions from the memory and, when executing the instructions, implement the control method for a material unloading cylinder according to any one of claims 5 to 9.

12. A material unloading vehicle, characterized in that, The material unloading vehicle includes: Material unloading cylinder; A point cloud data acquisition device is located at the end of the material unloading cylinder; and The control device for a material unloading cylinder according to claim 11.

13. A machine-readable storage medium storing instructions thereon, characterized in that, When executed by a processor, this instruction causes the processor to be configured to perform the control method for a material unloading cylinder according to any one of claims 5 to 9.

Images