Intelligent recovery cabinet and control method thereof

Abstract

The invention provides an intelligent recycling cabinet and a control method thereof. The control method comprises the steps that three-dimensional contour point clouds of thrown objects are collected, and volume data and weight data are generated; obtaining accumulation surface depth information of each branch bin, reconstructing a three-dimensional accumulation form, and calculating an accumulation volume fraction and an accumulation uniformity index; based on the local protrusion characteristics of the three-dimensional accumulation form, the contact structure parameters and the supporting stability, generating a branch warehouse bridging risk index; according to the residual accommodating space, the stacking balance improvement potential and the bridging risk suppression potential, target sub-bins of the thrown objects are determined; in the target sub-warehouse, generating a self-adaptive compression path including a compression starting point, a compression stroke, a compression direction and a compression speed according to the priority, the supporting structure and the risk state of the protruding area; and executing a compression path, monitoring the accumulation form change in real time in the compression process, and updating the accumulation volume fraction, the accumulation uniformity, the supporting stability and the bridging risk index according to a compression result.

Description

Technical Field

[0001] This invention relates to the field of intelligent recycling equipment technology, and in particular to an intelligent recycling cabinet and its control method. Background Technology

[0002] With the continuous development of urban household waste sorting and resource recycling systems, smart recycling bins are gradually becoming common recycling containers. Existing smart recycling bins typically possess basic functions such as waste identification, compartmentalized storage, capacity detection, and automatic compression, but they have the following shortcomings: Existing smart recycling bins typically perform simple sorting and compression control based on the type of items or the remaining capacity of the compartments. It is difficult to accurately analyze the three-dimensional shape data of the items, and it is also difficult to reconstruct the constantly changing three-dimensional morphology of the stack inside the compartments during the compression process in real time. As a result, complex spatial structures such as protruding areas, local void areas, and inclined support areas on the stacked surface are difficult to be effectively identified. Secondly, because the existing system lacks modeling of subtle parameters such as the contact surface, tilt angle, and support height of the stacked structure, it is also difficult to predict the risk of "bridging structures" forming between stacked materials, resulting in nonlinear faults such as jamming, suspended stacking, and ineffective compression that often occur under high loading rates.

[0003] Therefore, we propose an intelligent recycling cabinet and its control method. The information disclosed in the background section is only for enhancing the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by providing an intelligent recycling cabinet and its control method, thereby resolving the technical problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A control method for an intelligent recycling bin includes the following steps: S1. At the entry point of the recycling bin, the configured depth imaging module and weight sensing module are used to obtain the three-dimensional contour information and weight information of the current item being placed, and the volume data of the item is generated based on the three-dimensional contour information for subsequent accumulation analysis. S2. Based on the depth ranging array and weighing unit deployed inside each compartment, obtain the current stacking surface data and load-bearing weight data of each compartment; and reconstruct the three-dimensional stacking morphology of each compartment based on the stacking surface data to obtain the stacking volume fraction, stacking uniformity index and local stacking protrusion area of ​​each compartment. S3. Based on the stacking volume fraction, stacking uniformity index, and local stacking protrusion areas, calculate the bridging risk of each sub-compartment to obtain the corresponding bridging risk index, and use the bridging risk index as the input parameter for subsequent sub-compartment selection and compression path planning. S4. Based on the comprehensive data of the volume of the deployed material and the bridging risk index, determine the target compartment that is most suitable for the current deployment of the material; wherein, the target compartment is the compartment that can reduce local uneven stacking and reduce the risk of bridging while meeting the remaining capacity. S5. After determining the target compartment, an adaptive compression path is generated for the compression actuator based on the three-dimensional stacking morphology and bridging risk index. The adaptive compression path includes the compression start position, compression stroke, compression sequence and compression speed, and is used to guide the compression actuator to prioritize compression of the stacking protrusion area. S6. According to the target compartment, adjust the angle of the flow guiding mechanism at the delivery inlet to allow the delivery material to enter the target compartment; and drive the compression actuator to complete the compression action according to the adaptive compression path, so as to promote the stacking area to adjust in a uniform direction and expand the effective holding space.

[0006] S1 specifically involves: using a depth imaging module at the delivery entrance to perform a depth scan of the current delivery object, obtaining its 3D contour image data; based on the 3D contour image data, segmenting the delivery object from the background region and extracting the outer surface boundary point cloud of the delivery object; calculating the geometric dimensions of the delivery object, including length, width, and height, based on the outer surface boundary point cloud; generating estimated volume data of the delivery object based on the 3D geometric dimensions, which serves as input parameters for subsequent stacking assessment; and acquiring the weight data of the current delivery object through a weight sensing module, synchronizing the weight data with the volume data in time to enhance the reliability of subsequent stacking state calculations.

[0007] S2 specifically involves: acquiring the current stacking surface depth data of each compartment using a depth ranging array; generating a point cloud of the stacking region for each compartment based on the stacking surface depth data to represent the three-dimensional distribution of local stacking; reconstructing the three-dimensional stacking morphology within the compartment based on the stacking region point cloud, including the stacking shape, stacking edge surface, and the location of local stacking protrusions; calculating the stacking volume fraction of each compartment to assess the remaining capacity of the compartment; and generating a stacking uniformity index for each compartment to measure whether there are excessively high local stacking or void areas in the stacking region.

[0008] S3 specifically involves: identifying localized stacking protrusions in each compartment that could potentially lead to bridging based on the reconstructed three-dimensional stacking morphology; analyzing the contact surface structure between stacked materials based on these protrusions, and calculating the contact area, contact angle, and support height; calculating the support stability index between stacked materials based on the contact surface structure, serving as a key parameter for the likelihood of bridging formation; generating a bridging risk index for each compartment based on the stacking volume fraction, stacking uniformity index, and the aforementioned support stability index; and periodically updating the bridging risk index based on the reconstruction frequency and real-time monitoring data to maintain the real-time effectiveness of the prediction results.

[0009] S4 specifically involves: comparing the remaining capacity of each compartment based on the stacking volume fraction to determine if it meets the volume requirements of the material being placed; judging whether the material's placement can improve the local stacking balance of the target compartment based on the stacking uniformity index; analyzing whether the material's entry into each compartment has the potential to reduce bridging risk by using the bridging risk index as input; generating candidate compartment scores for all compartments while meeting the capacity requirements, taking into account both the local stacking balance requirement and the bridging risk suppression requirement; determining the target compartment corresponding to the current material based on the candidate compartment scores, and using the target compartment as the basis for subsequent compression path generation and flow control.

[0010] S5 specifically involves: calculating the compression priority of each protruding area based on the local accumulation protrusions; determining the compression starting position of the compression actuator based on the compression priority; planning the compression stroke of the compression actuator to effectively compress the protruding areas; generating the compression sequence of the compression actuator based on the positional relationship and priority between each protruding area; and setting the compression speed parameters of the compression actuator according to the bridging risk index to avoid unstable accumulation patterns under high-risk conditions.

[0011] S6 specifically involves: adjusting the angle of the delivery inlet guide mechanism according to the target compartment to ensure accurate delivery of the material into the target compartment; confirming that the material has entered the target compartment through the landing point monitoring sensor and feeding back the landing point data to the stacking status monitoring module; initiating the compression actuator to perform the compression action; during the compression process, monitoring the real-time changes in the stacking morphology through the depth ranging array to confirm whether the compression path is executed according to the plan; updating the stacking volume fraction, local stacking uniformity index, and bridging risk index based on the compressed stacking morphology to complete one control cycle.

[0012] The present invention also provides an intelligent recycling cabinet, which adopts the control method of the intelligent recycling cabinet described above.

[0013] The beneficial effects of this invention are as follows: This invention enables the system to obtain high-precision, structured shape feature data of the delivered items before they enter the sorting compartments by reconstructing 3D point clouds, extracting geometric parameters, and simultaneously collecting volume and weight data. This provides a stable and reliable foundation for subsequent stacking prediction and sorting decisions, improving the overall accuracy of sorting and stacking management. Through 3D stacking morphology, stacking volume fraction, and stacking uniformity indicators, the system can monitor the three-dimensional stacking structure inside the sorting compartments in real time, allowing compression actions to be adjusted based on the actual stacking state rather than relying on fixed presets.

[0014] This invention establishes a bridging risk index model by introducing comprehensive parameters such as local protrusion features, contact surface tilt angle, support stability, and stacking uniformity. This enables the system to identify potential risk points before the bridging structure is formed and to take mitigation measures in subsequent actions, effectively reducing failures such as material jamming, suspended stacking failures, and compression. By comprehensively scoring the remaining capacity space, the potential for improving stacking balance, and the potential for mitigating bridging risks, the selection of target compartments no longer relies on a single capacity threshold but is dynamically optimized based on multi-dimensional stacking structure indicators. This improves overall capacity efficiency and significantly reduces the probability of local overload in a particular compartment.

[0015] This invention utilizes parameter models such as compression starting point, compression stroke, compression direction, compression sequence, and compression speed to enable the compression mechanism to automatically generate the optimal compression path based on real-time protrusion height, stacking uniformity, support stability, and bridging risk index. This effectively reduces local protrusions, eliminates unstable stacking supports, and avoids stacking instability or compression failure due to incorrect paths. The deviation of the object's landing point, compression effect gain, changes in stacking morphology, and the attenuation of bridging risk are used as feedback signals to input into the next round of control, giving the system self-learning capabilities. During continuous operation, the system can continuously correct the guide angle, optimize the compression strategy, and dynamically reduce bridging risk, achieving long-term stable, efficient, and continuously high-capacity operation of the intelligent recycling bin. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a control method for an intelligent recycling cabinet according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1: As Figure 1 As shown, this embodiment provides a control method for an intelligent recycling cabinet, including the following steps: S1. Three-dimensional information acquisition steps of the deposited object: At the deposit entrance of the recycling bin, the configured depth imaging module and weight sensing module are used to acquire the three-dimensional contour information and weight information of the current deposited object, and the volume data of the deposited object is generated based on the three-dimensional contour information for subsequent accumulation analysis. S2, Reconstruction of Compartment Stacking Status: Based on the depth ranging array and weighing unit deployed inside each compartment, obtain the current stacking surface data and load-bearing weight data of each compartment; and reconstruct the three-dimensional stacking morphology of each compartment according to the stacking surface data to obtain the stacking volume fraction, stacking uniformity index and local stacking protrusion area of ​​each compartment. S3. Bridge Risk Prediction Step: Based on the stacking volume fraction, stacking uniformity index and local stacking protrusion areas obtained in step S2, the bridge risk of each sub-compartment is calculated to obtain the corresponding bridge risk index, and the bridge risk index is used as the input parameter for subsequent sub-compartment selection and compression path planning. S4. Target compartment determination step: Combining the volume data of the deployed material in step S1 and the bridging risk index in step S3, determine the target compartment most suitable for the current deployed material; wherein, the target compartment is a compartment that can reduce local uneven stacking and reduce bridging risk while meeting the remaining capacity requirements. S5. Adaptive Compression Path Generation Step: After determining the target compartment, based on the three-dimensional stacking morphology in step S2 and the bridging risk index in step S3, an adaptive compression path for the compression actuator is generated; the adaptive compression path includes the compression starting position, compression stroke, compression sequence, and compression speed, and is used to guide the compression actuator to preferentially compress the stacking protrusion area. S6. Target action execution steps: According to the target compartment obtained in step S4, adjust the angle of the flow guiding mechanism at the delivery inlet to allow the delivery material to enter the target compartment; and drive the compression execution mechanism to complete the compression action according to the adaptive compression path obtained in step S5, so as to promote the stacking area to adjust in a uniform direction and expand the effective holding space.

[0019] S1. Three-dimensional information acquisition steps for the disposed items: At the disposal entrance of the recycling bin, the configured depth imaging module and weight sensing module are used to acquire the three-dimensional contour information and weight information of the currently disposed items, and the volume data of the disposed items is generated based on the three-dimensional contour information for subsequent accumulation analysis; specifically including the following sub-steps: S110, 3D contour image acquisition sub-step: Use the depth imaging module at the delivery entrance to perform a depth scan of the current delivery object and obtain the corresponding point cloud set of the delivery object: in, A three-dimensional point cloud set of the deployed object; Point cloud sequence number; This represents the total number of points contained in the point cloud. , , They represent the first The X, Y, and Z coordinates of a point in a three-dimensional coordinate system.

[0020] In acquiring point cloud sets Before segmentation, statistical outlier removal or radius filtering is performed on the point cloud. Specifically, for each point, the average distance to its k nearest neighbors is calculated. If this distance is greater than a threshold equal to the sum of the global average distance and the standard deviation, the point is identified as noise and removed. This effectively prevents outliers caused by sensor noise from affecting subsequent processing. , , A huge error occurred in the calculation.

[0021] S120, Contour Segmentation and Boundary Extraction Sub-step: This involves processing the point cloud set obtained in step S110... Perform point cloud segmentation between the delivery area and the background area to obtain the delivery point cloud set: in, The point cloud set representing the outer surface of the object being deployed; The point cloud sequence number of the deployed object; The number of point clouds of the deployed objects, and satisfying the following conditions. ; These are the point cloud coordinates belonging to the outer surface of the object being placed.

[0022] in, Let be the point cloud coordinates belonging to the outer surface of the object, and These represent the components of the coordinate on the X, Y, and Z axes, respectively.

[0023] Segmentation can be performed using either background subtraction or pass-through filtering. The system pre-stores a background depth model of an empty cabinet. By comparing the current frame's depth data with the background model, point clouds with depth differences less than a preset threshold (e.g., 5mm) are marked as background and removed. The retained point clouds representing the differences are the object's point clouds. .

[0024] S130, 3D geometric parameter calculation sub-step: Based on the point cloud set from step S120 Calculate the length parameters of the objects to be placed. Width parameters and height parameters The calculation formula is as follows: in, Let X be the geometric length of the object along the X direction; The geometric width of the object along the Y direction; The geometric height of the object along the Z direction; and These are operations to retrieve the maximum and minimum values, respectively.

[0025] S140, Volume data generation sub-step: Based on the data obtained in step S130 , , Calculate the estimated volume data of the deployed material. In one embodiment, a bounding box volume estimation method can be used, the formula of which is: in, This is the estimated volume data for the material being placed.

[0026] In another alternative embodiment, a voxel integral estimation method can be used to estimate the point cloud. Discretize into several voxel units, and let the volume of each voxel be... The number of voxels falling into the drop zone is The volume can then be expressed as: in, The number of voxels within the area where the object is placed; The volume of a single voxel unit.

[0027] S150, Weight Data Synchronization Acquisition Sub-step: Acquire the weight data of the object being placed through the weight sensor module. and compare it with the volume data obtained in step S140. Time synchronization is performed. To ensure that both data belong to the same delivery process, timestamps are assigned to the volume data and weight data respectively. and And satisfy the synchronization conditions: in, This refers to the weight data of the items being delivered; The timestamp when the volume data was generated; This is the timestamp when the weight data was collected. This is the unified timestamp after synchronization.

[0028] like The system can then reject the current synchronization and re-collect data. This is the preset time synchronization allowable error threshold.

[0029] Considering the computational demands of 3D point cloud processing and real-time path planning, the main control unit of the control system employs an embedded edge computing module (such as an industrial computer with GPU acceleration or a high-performance ARM processor). The depth imaging module transmits high-frame-rate depth maps to the main control unit via USB 3.0 or a gigabit Ethernet port. The main control unit utilizes CUDA or OpenCL parallel computing frameworks to accelerate point cloud segmentation and reconstruction operations (S1-S2 steps), ensuring that the latency from deployment monitoring to path generation is controlled within 500ms.

[0030] Furthermore, in step S1, if the point cloud missing rate acquired by the depth imaging module exceeds a threshold (e.g., 20%), the system will automatically switch to a safety mode: suspend adaptive path planning, call the center point vertical compression path only based on weight information, and send a lens maintenance alarm to the background. This ensures that the device can still maintain basic retrieval functions even under extreme conditions such as damage to the visual sensor.

[0031] S2. Compartment Stacking State Reconstruction Steps: Based on the depth ranging array and weighing unit deployed inside each compartment, acquire the current stacking surface data and load-bearing weight data of each compartment; and reconstruct the three-dimensional stacking morphology of each compartment based on the stacking surface data to obtain the stacking volume fraction, stacking uniformity index, and local stacking protrusion areas of each compartment; specifically including the following sub-steps: S210, Sub-step for collecting depth data of the stacked surface: Inside each compartment, depth map data of the stacked surface of the compartment is acquired through a depth ranging array: in, For the first A collection of depth maps for each sub-warehouse; pixels in the depth map The corresponding depth value; , These are the horizontal and vertical pixel indices of the depth map, respectively; , These represent the horizontal and vertical pixel counts of the depth map, respectively. This is the warehouse number.

[0032] S220, Sub-step for generating point clouds in stacked regions: This involves generating depth map data... Convert to the corresponding 3D point cloud set: in, For the first A collection of point clouds representing the stacking areas of each sub-warehouse; Point cloud sequence number; This represents the total number of point clouds in this sub-warehouse; For the first The three-dimensional coordinates of each point in the compartment coordinate system.

[0033] Point cloud generation methods may include depth value back projection: in, , The coordinates of the principal point of the depth camera; , This refers to the focal length parameter of the depth camera.

[0034] In generating point cloud sets Next, the point cloud needs to be transformed from the camera coordinate system to the compartment physical coordinate system using extrinsic parameter matrices (rotation matrix R and translation vector T). In the compartment physical coordinate system, the XY plane is parallel to the bottom of the compartment, and the Z-axis is vertically upward. The transformed z-coordinate value represents the actual stacking height of that point relative to the bottom of the compartment, and is thus used to construct the height function. .

[0035] S230, Sub-step for 3D stacking morphology reconstruction: Based on point cloud set Construct a three-dimensional height function for the stacking surface inside the compartment: satisfy Among them, symbols This indicates an approximate matching or neighborhood attribution determination of coordinate values. For the first Each compartment is located on a plane coordinate system. The stacking height value at that location.

[0036] Based on this, construct the stacking volume region: in, For the first The domain of the volume region of the accumulated material in the compartment.

[0037] This step outputs a three-dimensional stacked morphology model, which serves as input for S240 and S250.

[0038] In constructing the height function At that time, for grid points with missing depth data An interpolation algorithm (such as bilinear interpolation or nearest neighbor interpolation) is used to fill in the height using the height values ​​of surrounding valid points; if there are no surrounding valid points (such as at the edge of a compartment), its height value is set to 0. This ensures the height function... In the domain The continuity within the volume ensures the effectiveness of subsequent volume integral and uniformity calculations.

[0039] S240, Sub-step for calculating bulk volume fraction: Based on the bulk region Calculate the stacked volume: in, For the first Current stacking volume of each sub-compartment This represents a double integral operation within the bottom planar region of the compartment.

[0040] Let the total usable volume of this sub-warehouse be... The bulk volume fraction is defined as: in, This represents the stacking volume fraction of the compartments, used to characterize the remaining storage space.

[0041] S250, Sub-step for generating packing uniformity index: To characterize the smoothness of the packing surface, based on the packing height function... Calculate the uniformity index.

[0042] Let the average height of the stack be in, For the first Average stacking height of each compartment; This refers to the bottom area of ​​the sub-warehouse.

[0043] The packing uniformity index can be defined as the standard deviation of height deviation: in, For the first The stacking uniformity index of the compartments indicates that the higher the value, the more uneven the stacking, and the possible presence of protrusions or local voids.

[0044] S3. Bridging Risk Prediction Step: Based on the stacking volume fraction, stacking uniformity index, and local stacking protrusion areas obtained in step S2, the bridging risk of each sub-compartment is calculated to obtain the corresponding bridging risk index. This bridging risk index is then used as input parameters for subsequent sub-compartment selection and compression path planning. Specifically, this includes the following sub-steps: S310, Local Stacking Protrusion Identification Sub-step: Based on the stacking height function of step S230 Calculate the local protrusion function: in, coordinates Local protrusion height deviation; For the deposited surface in Height; This represents the average stacking height of the compartment.

[0045] When satisfied Then the coordinates Marked as local protrusions, where, The threshold for identifying local protrusions.

[0046] Furthermore, all protrusion points are grouped into a set of protrusion regions: S320, Contact Surface Structure Analysis Sub-step: In the protruding area Inside, for point cloud collection Perform local plane fitting on the neighborhood of . Let the . The neighborhood fitting plane for each point is: in, , , , These are all parameters of the fitted plane.

[0047] The contact tilt angle is calculated based on this plane: in, For the first The local contact surface inclination angle at each protrusion point reflects the stability of the supporting surface; The larger the size, the steeper the surface. This is an operation on the inverse cosine function.

[0048] Simultaneously calculate the local support height in this area: in, For point The stacking height value at that location.

[0049] By combining the tilt angle and the support height, a local contact structure vector can be formed: .

[0050] The neighborhood refers to all point cloud data within a preset radius r (e.g., r = 5cm to 10cm) centered on the current point; or the k nearest neighboring points (e.g., k = 20 to 50) are selected as the fitting dataset. The points within this neighborhood are fitted using the least squares method or principal component analysis (PCA) to obtain the plane parameters a, b, c, and d.

[0051] S330, Support Stability Calculation Sub-step: For each protrusion point, based on its contact structure vector... Calculate the local support stability index: in, For the first Local support stability index of each protrusion point; The height influence coefficient (preferred value 0.05) reflects the trend that the greater the stacking height, the more unstable the structure becomes. The tilt angle influence coefficient (preferred value 0.02) reflects that the larger the angle, the more unstable the support.

[0052] By combining all the breakout points, the average support stability of the entire sub-compartment is obtained: in, This represents the number of points within the protruding area. This serves as an indicator of overall support stability for the sub-warehouses.

[0053] The smaller the value, the more unstable the whole system is, and the more likely bridging will occur.

[0054] S340, Bridge Construction Risk Index Generation Sub-step: Bridge construction risk depends on factors such as the degree of local protrusions, uniformity of stacking, and support stability. The following risk model is constructed: in, For the first The bridging risk index for each sub-account, the higher the value, the higher the bridging risk; The average height of the protrusion in the protruding area: This is an index of packing uniformity. To support stability indicators; , , As weighting parameters, preferred It is 0.4. It is 0.3. It is 0.3.

[0055] In particular, if the identified set of protruding regions It is an empty set (i.e.) If the surface of the accumulation is flat and without obvious protrusions, then directly set the average protrusion height. =0; Overall support stability of sub-accounts =1 (indicating the highest stability due to support from a flat bottom); at this point, the bridge construction risk index is... Take the preset minimum reference value (e.g.) =0).

[0056] S350, Bridge Construction Risk Update Sub-step: To ensure the real-time accuracy of bridge construction risk prediction, the bridge construction risk index is updated over time. Let the risk value of the previous period be... The current cycle risk value is The updated formula is: in, The updated bridge construction risk index; The weight parameter is updated over time (preferably 0.7) to satisfy the following conditions: This update mechanism can smooth out instantaneous fluctuations, making risk assessment more stable and reliable.

[0057] S4. Target Compartment Determination Step: Combining the volume data of the deployed material from step S1 and the bridging risk index from step S3, determine the target compartment most suitable for the current deployed material; wherein, the target compartment is a compartment that, while satisfying the remaining capacity, can reduce local uneven stacking and reduce bridging risk; specifically including the following sub-steps: S410, Remaining Capacity Comparison Sub-Step: Based on the calculation obtained in step S240, the... Individual compartment stacking volume fraction Its remaining capacity space is defined as: in, For the first The proportion of remaining storage space in each compartment; This represents the volume fraction of the stacked volume in this compartment.

[0058] If the volume data of the material to be disposed in step S140 is The item was placed on the The criteria for determining the capacity of a sub-warehouse are: in, For the first The maximum usable volume of the compartment.

[0059] Sub-warehouses that meet this condition constitute a set of sub-warehouses that can be accommodated: ; The vertical line represents the set construction operation. The left side of the vertical line represents the elements of the set, and the right side represents the conditions that the elements must satisfy.

[0060] Specifically, if the calculated set of accommodating compartments... If the system detects an empty set (meaning that the remaining space in all compartments is insufficient to accommodate the current item being placed), it will trigger a full-load exception handling mode: The system will pause the placement process, prompt the user to change the recycling bin via the human-machine interface, or automatically trigger a powerful compression and reorganization action on all compartments, and then re-execute step S2 to check the remaining space. If there is still no usable space after reorganization, the placement port will be locked.

[0061] S420, Local Stacking Equilibrium Demand Assessment Sub-step: Based on the stacking uniformity index obtained in step S250... Define the first item pair The potential indicators for improving the stacking balance of the sub-warehouses are: in, The larger the value, the more likely the material is to improve the uniformity of stacking after entering the compartment; A larger value indicates a more uneven distribution of material. It decreases accordingly.

[0062] S430, Sub-step for assessing bridge construction risk mitigation requirements: Based on the bridge construction risk index obtained in step S340... Define the potential index of the impact of the deployed material on the risk of bridge construction: in, For the first The potential indicator for mitigating bridging risk in sub-accounts; The current bridging risk index for this sub-warehouse; The larger the value, the less likely the risk of bridging will be after the delivered material enters the compartment.

[0063] S440, Candidate Sub-compartment Scoring Generation Sub-step: For the set of sub-compartments that can accommodate... For each sub-warehouse within the warehouse, a comprehensive score is generated by considering its remaining capacity, potential for improving stacking equilibrium, and potential for mitigating bridging risks. in, For the first Overall score of the sub-warehouse; This represents the proportion of remaining storage space. To improve the potential indicators for stacking equilibrium; As an indicator of potential for mitigating bridge construction risks; , , The weighting coefficients for the three indicators satisfy: .

[0064] In this embodiment, the weighting parameter is chosen to prioritize capacity while also considering stacking quality. Preferably, the weighting parameter is set to... (Focusing on remaining space) (Focus on improving the accumulation of debris) (Focus on risk mitigation). This parameter combination has been experimentally verified to effectively reduce the frequency of material jamming while maintaining a high loading rate.

[0065] It should be noted that, in order to eliminate the influence of different dimensions of the indicators, in the calculation... Previously, various indicators (especially) and All parameters are normalized or dimensionless. For example, packing uniformity. Normalization can be performed relative to the maximum height of the sub-compartment to ensure... They are all within the same order of magnitude (usually in the [0,1] range), thus ensuring the effectiveness of the weighted scoring.

[0066] S450, Final Determination of Target Warehouse Sub-step: Based on the comprehensive score in step S440, select from the set of warehouses that can accommodate the target warehouses. The highest-rated warehouse is selected as the target warehouse for the current delivery item. in, Number the optimal target warehouse; Score the candidate sub-warehouses. This represents the set of independent variables that maximize the objective function (i.e., the sub-compartment number corresponding to the maximum score). This selection method is explicit, computable, and fully implementable.

[0067] S5. Adaptive Compression Path Generation Step: After determining the target compartment, based on the three-dimensional stacking morphology from step S2 and the bridging risk index from step S3, an adaptive compression path is generated for the compression actuator. The adaptive compression path includes the compression starting position, compression stroke, compression sequence, and compression speed, and is used to guide the compression actuator to preferentially compress the stacking protrusion area. Specifically, it includes the following sub-steps: S510, Priority Evaluation of Protrusion Regions Sub-step: Based on the set of protrusion regions identified in step S310... For each protrusion point, calculate the protrusion priority index. : in, For the first Compression priority index for each protrusion point; This refers to the local deviation in the height of the protrusion at that point; The local contact surface inclination angle at that point; Risk index for bridging sub-warehouses; , , For priority weight parameters, satisfying: , 0.5 is preferred. 0.3 is preferred. 0.2 is preferred.

[0068] Define priority set: S520, Determining the Compression Starting Position Sub-step: Based on the priority index of step S510, select the highest priority point from the protruding region as the compression starting point: in This represents the parameter that maximizes the objective function (i.e., the burst point number corresponding to the highest priority). This is the optimal breakthrough point number.

[0069] The coordinates of the compression start position are: in, The optimal initiation point number; These are the starting spatial coordinates of the compressed path.

[0070] S530, Compressed Stroke Planning Sub-step: Define the compressed stroke length for: in, To compress the stroke length of the actuator; The deviation is the height of the protrusion at the starting point; The tilt angle at the starting point; This is an index for the uniformity of stacking in this compartment; , , Let the weight parameters satisfy: . The travel direction vector is set as follows: in , This represents the partial derivative of the stacking height function at the starting point; The direction pointing towards a locally steep descent is the most advantageous for "reducing protrusions and lowering bridge construction risks," among which... This indicates the operation of taking partial derivatives with respect to the variables x or y; parentheses Represents vector construction operations.

[0071] The end point of compression is: The calculated three-dimensional coordinates of the compression path endpoint are provided. To achieve the aforementioned three-dimensional adaptive compression path, the compression actuator employs a three-axis gantry structure or a hydraulic push rod structure with rotary joints. This mechanism not only possesses vertical compression capability along the Z-axis but is also equipped with displacement servo modules (or angle adjustment motors) for the X and Y axes, enabling it to compress according to the direction vector generated by the control system. By adjusting the spatial orientation of the compression plate or changing the direction of the resultant force of the downward pressure, directional oblique compression of the local protruding area can be achieved.

[0072] Furthermore, in generating the compression endpoint Afterwards, the system needs to perform kinematic constraint verification: if the calculated endpoint coordinates exceed the physical boundary of the compartment or are outside the stroke limit of the compression mechanism, the endpoint coordinates are truncated (Clamping) to limit them within the allowable safe range and prevent mechanical collisions.

[0073] Considering the stacking height function For discrete grid data, partial derivatives Numerical approximations are performed using the central difference method or the Sobel operator. For example, the partial derivative in the X-axis direction can be approximated as: in The grid spacing is [value]. This ensures that the descent direction of the gradient can be accurately obtained on discrete point cloud data.

[0074] S540, Compression Sequence Generation Sub-step: To ensure the compression path covers key protrusion regions, construct a distance metric between protrusion points: Define the compression order selection function: in, For the compressed path One access point; symbol This represents the set difference operation (i.e., removing visited points from the candidate set); The geometric distance between points; For the first Priority of each breakthrough point; , For distance weights and priority weights, satisfying , This represents the parameter that minimizes the objective function. The formula prioritizes compressing regions with high convexity while minimizing the travel distance, thus improving efficiency and stability.

[0075] The generation of the compression order is an iterative process, which stops generating subsequent paths when any of the following termination conditions are met: Candidate Set Empty (meaning all breakpoints have been planned); The current optimal breakthrough point priority Below the preset minimum motion threshold (indicating that the remaining protrusions are not obvious and no compression is needed); The planned number of compression points reaches the maximum allowed number of times in a single cycle (e.g., 5 times, to ensure system throughput efficiency).

[0076] S550, Compression Speed ​​Parameter Setting Sub-step: Based on the risk of compartment bridging With local stability Set the compression speed : in, To reduce the compression speed of the compression actuator; The maximum allowable compression speed; To support stability indicators; This is the bridge construction risk index.

[0077] When the risk is high Or low stability When the risk is low and the support is good, the compression speed will automatically decrease; when the risk is low and the support is good, the compression speed can be increased to improve efficiency.

[0078] S6. Target Action Execution Steps: Following the target compartmentation obtained in step S4, adjust the angle of the flow guiding mechanism at the delivery inlet to allow the material to enter the target compartment; and drive the compression actuator to complete the compression action according to the adaptive compression path obtained in step S5, thereby causing the accumulation area to adjust in a uniform direction and expand the effective carrying space. Specifically, this includes the following sub-steps: S610, Sub-step for adjusting the angle of the flow guiding mechanism: Based on the target compartment number obtained in step S450 Let the angle of the inlet guide mechanism be... The expected flow angle for different sub-compartments is The formula is adjusted as follows: in, The angle of the updated flow guiding mechanism; The current angle of the flow guiding mechanism; The guiding angle corresponding to the target warehouse; Adjust the step size coefficient to meet the following requirements for the guide angle: .

[0079] This formula enables smooth adjustment, avoiding deviation of the placed object caused by sudden large rotations.

[0080] S620, Sub-step for confirming the landing point of the released object: Assume the three-dimensional coordinates of the actual landing point are... The expected landing point coordinates (center of the distribution center) are: The landing point deviation is: like Then it is confirmed that the delivered goods have entered the target distribution box. This is the threshold for the allowable error of the landing point.

[0081] This deviation will serve as feedback for subsequent adjustments to the guide angle.

[0082] S630, Compression execution start sub-step: Using the compression path obtained in steps S520–S550, starting point Path endpoint Compression speed The compression displacement trajectory of the actuator is as follows: in, The unit vector in the direction of compression; This is a time variable for the compression process.

[0083] The trajectory is executed until the compression displacement reaches the stroke length: .

[0084] During the execution of the compression displacement trajectory, the drive current of the compression actuator (or the feedback value of the pressure sensor) is monitored in real time. If the drive current exceeds the preset safety threshold (indicating that an incompressible hard object has been encountered), the system immediately stops the current compression path, performs a reversal action, and marks the current position as a 'hard obstacle point'. In the next round of path planning, the system will automatically avoid the area or prompt manual clearing.

[0085] S640, Sub-step for monitoring morphological changes during compression: During compression, measure the height function after compression in real time. The instantaneous morphological change during compression is defined as: in, This is the height field before compression; For the compression process at any time The height field.

[0086] The compression effect evaluation function is defined as follows: in, This represents the compression effect gain; a higher value indicates more effective compression. This monitoring result will serve as real-time feedback for risk control and path updates.

[0087] Calculating the compression effect gain When this is the case, a noise threshold is introduced. Only when the change in height at a certain location Only when the value is 5mm (e.g., 5mm) is it included in the integration calculation; otherwise, it is considered as negligible sensor noise. This can improve the signal-to-noise ratio of the feedback gain.

[0088] S650, Compression Result Confirmation and Dynamic Update Sub-step: Define the update formula for key indicators after compression: (1) Update the stacked volume: in, This is the height function at the end of compression.

[0089] (2) Uniformity index update: in, This represents the average height after compression.

[0090] (3) Support stability update: The stability index of S330 Provide feedback and make corrections: in, For the updated support stability; The amount of stability improvement obtained based on the reduction in protrusion after compression; The stability coefficients are updated.

[0091] (4) Dynamic updates of bridge construction risk index: Using the risk smoothing model of S350, combined with the compression effect gain in this study : in, This is the immediate risk value after compression; This is the risk value from the previous period; This is the coefficient representing the impact of compression gain on risk attenuation. This formula reflects that the more effective the compression, the more significant the reduction in bridging risk.

[0092] It should be noted that after updating the above indicators, numerical boundary constraints (clamping) must be performed. For example, the updated support stability... Bridge construction risk index All must be limited to Within a closed interval: if the calculation result is greater than 1, then take 1; if it is less than 0, then take 0. This ensures that the state parameters are always kept within a physically meaningful effective range.

[0093] Example 2: This example provides a control system for an intelligent recycling cabinet, including: The three-dimensional information acquisition module for the object is used to perform depth imaging of the object at the delivery entrance, acquire the three-dimensional point cloud data of the object, and generate the geometric dimension parameters, volume data and weight data of the object based on the three-dimensional point cloud data. The compartment stacking state reconstruction module is used to acquire depth data of the stacking surface of each compartment, generate corresponding stacking area point clouds, reconstruct the three-dimensional stacking shape of each compartment, and calculate the stacking volume fraction and stacking uniformity index. The bridging risk prediction module is used to generate a bridging risk index for each compartment based on the local protrusion features, contact surface structural parameters, support stability index, and stacking uniformity index of the three-dimensional stacking morphology. The target compartment determination module is used to comprehensively score the available compartments based on the remaining capacity of each compartment, the potential index for improving stacking balance, and the potential index for mitigating bridging risks, and to determine the target compartments for the deployed materials based on the scoring results. An adaptive compression path generation module is used to generate an adaptive compression path for the compression actuator based on the three-dimensional stacking pattern, local protrusion priority, support structure parameters, and bridging risk index of the target compartment. The adaptive compression path includes compression start position, compression stroke, compression direction, compression sequence, and compression speed. The control execution module is used to adjust the angle of the delivery inlet guide mechanism according to the target compartment, and to control the compression execution mechanism to perform compression actions according to the adaptive compression path; The status update module is used to monitor changes in the stacking morphology during the compression process, obtain the compressed stacking volume fraction, stacking uniformity index, support stability index, and bridging risk index, and input the updated index into the compartment stacking status reconstruction module, the bridging risk prediction module, the target compartment determination module, and the adaptive compression path generation module to construct a dynamic closed loop of the control method.

[0094] This embodiment also provides an intelligent recycling cabinet, which adopts the control method of an intelligent recycling cabinet described in Embodiment 1 and the control system of an intelligent recycling cabinet described in Embodiment 2.

[0095] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0096] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0097] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0098] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0099] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0100] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0101] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0102] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0103] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0104] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A control method for an intelligent recycling cabinet, characterized in that, Includes the following steps: S1. At the entry point of the recycling bin, the configured depth imaging module and weight sensing module are used to obtain the three-dimensional contour information and weight information of the current item being placed, and the volume data of the item is generated based on the three-dimensional contour information for subsequent accumulation analysis. S2. Based on the depth ranging array and weighing unit deployed inside each compartment, obtain the current stacking surface data and load-bearing weight data of each compartment. The three-dimensional stacking morphology of each compartment is reconstructed based on the stacking surface data, and the stacking volume fraction, stacking uniformity index and local stacking protrusion area of ​​each compartment are obtained. S3. Based on the stacking volume fraction, stacking uniformity index, and local stacking protrusion areas, calculate the bridging risk of each sub-compartment to obtain the corresponding bridging risk index, and use the bridging risk index as the input parameter for subsequent sub-compartment selection and compression path planning. S4. Based on the comprehensive data of the volume of the deployed material and the bridging risk index, determine the target compartment that is most suitable for the current deployment of the material; wherein, the target compartment is the compartment that can reduce local uneven stacking and reduce the risk of bridging while meeting the remaining capacity. S5. After determining the target compartment, an adaptive compression path is generated based on the three-dimensional stacking morphology and bridging risk index for the compression actuator. The adaptive compression path includes the compression start position, compression stroke, compression sequence and compression speed, and is used to guide the compression actuator to prioritize compression of the stacking protrusion area.

2. The control method for an intelligent recycling cabinet according to claim 1, characterized in that, It also includes S6, which adjusts the angle of the flow guiding mechanism at the delivery inlet according to the target compartment so that the delivery material enters the target compartment; and drives the compression actuator to complete the compression action according to the adaptive compression path so as to make the accumulation area adjust to a uniform direction and expand the effective capacity space.

3. The control method for an intelligent recycling cabinet according to claim 1, characterized in that, S1 specifically refers to: The depth imaging module at the delivery entrance is used to perform a depth scan of the current delivery object to obtain the three-dimensional contour image data of the delivery object; Based on 3D contour image data, segmentation of the object and background region is performed, and the outer surface boundary point cloud of the object is extracted; Based on the point cloud of the outer surface boundary, calculate the geometric dimensions of the object, including length, width, and height. Based on the three-dimensional geometric dimensions, the estimated volume data of the deployed object is generated as input parameters for subsequent stacking assessment. The weight data of the currently deployed object is obtained by a weight sensor module, and the weight data is synchronized with the volume data in time to enhance the reliability of subsequent stacking state calculations.

4. The control method for an intelligent recycling cabinet according to claim 1, characterized in that, S2 specifically refers to: Within each compartment, the current stacking surface depth data of the compartment is obtained through a depth ranging array; Based on the depth data of the stacked surface, a point cloud of the stacked area corresponding to each compartment is generated to represent the three-dimensional distribution of the local stacking. Based on the point cloud of the stacking region, the three-dimensional stacking morphology inside the compartment is reconstructed, including the stacking shape, stacking edge surface and local stacking protrusions. Calculate the stacking volume fraction of each compartment to assess the remaining capacity of the compartment; Generate stacking uniformity indices for each compartment to measure whether there are areas of excessively high stacking or voids in the stacking area.

5. The control method for an intelligent recycling cabinet according to claim 1, characterized in that, S3 specifically refers to: Based on the reconstructed three-dimensional stacking morphology, identify local stacking protrusions in each compartment that may lead to bridging; Based on the localized accumulation protrusions, the contact surface structure between the accumulations is analyzed, and the contact area, contact angle, and support height are calculated. Based on the contact surface structure, the support stability index between the accumulations is calculated as a key parameter for the possibility of bridging formation. Based on the stacking volume fraction, stacking uniformity index, and the aforementioned support stability index, a bridging risk index is generated for each compartment. Based on the reconstruction frequency and real-time monitoring data, the bridge construction risk index is updated periodically to maintain the real-time effectiveness of the prediction results.

6. The control method for an intelligent recycling cabinet according to claim 1, characterized in that, S4 specifically refers to: Based on the stacking volume fraction, compare whether the remaining capacity of each compartment meets the volume data of the material to be placed. Based on the stacking uniformity index, it is determined whether the input of the material can improve the local stacking balance of the target compartment; Using the bridge-building risk index as input, we analyze whether the delivery material has the potential to reduce bridge-building risk after entering each compartment. Under the premise of meeting the accommodation space conditions, and taking into account the local stacking balance requirements and the bridging risk suppression requirements, candidate sub-compartments scores are generated for all sub-compartments. Based on the candidate distribution scores, the target distribution corresponding to the current delivery item is determined, and the target distribution is used as the basis for subsequent compression path generation and flow control.

7. The control method for an intelligent recycling cabinet according to claim 1, characterized in that, S5 specifically refers to: Calculate the compression priority of each protruding region based on the local accumulation protrusion regions; The compression start position of the compression actuator is determined based on the compression priority. The compression stroke of the compression actuator is planned to effectively compress the protruding area; Based on the positional relationship and priority between each protruding region, the compression sequence of the compression actuator is generated; Based on the risk index of bridge construction, the compression speed parameters of the compression actuator are set to avoid unstable accumulation patterns under high-risk conditions.

8. The control method for an intelligent recycling cabinet according to claim 2, characterized in that, S6 specifically refers to: Adjust the angle of the delivery inlet guide mechanism according to the target compartment to ensure that the delivery material is accurately directed to the target compartment; The landing point monitoring sensor confirms that the material has entered the target compartment and feeds back the landing point data to the stacking status monitoring module; then the compression actuator is activated to perform the compression action.

9. The control method for an intelligent recycling cabinet according to claim 8, characterized in that, S6 also includes: during the compression process, monitoring the real-time changes in the stacking morphology through a depth ranging array to confirm whether the compression path is executed according to the plan; updating the stacking volume fraction, local stacking uniformity index and bridging risk index based on the compressed stacking morphology to complete one control cycle.

10. An intelligent recycling bin, characterized in that, The control method for an intelligent recycling cabinet according to any one of claims 1-9 is adopted.

Images