Vision processing based baggage evaluation method, controller, medium and product

Abstract

A kind of based on visual processing's checked baggage evaluation method, controller, medium and product, it is related to intelligent control field.The method includes: generating the three-dimensional digital model of the checked baggage based on the three-dimensional point cloud and image data obtained;According to the extracted geometric and texture features, identify and mark the local area of the model as rigid mutation region or flexible transition region;The intersection volume is calculated by inputting the model into the virtual channel model;If the intersection volume is over-limit and the corresponding local area attribute is rigid mutation region, it is determined that there is a risk of jamming and the luggage is rejected;If the attribute is flexible transition region and the intersection volume ratio is less than the preset deformation threshold, the luggage is allowed to be checked.This application distinguishes rigid interference and flexible deformation through virtual simulation, solves the problem of high false positive rate of irregular luggage caused by size extreme static comparison in related technology, and improves the identification accuracy and pass rate of self-service check-in system.

Description

Technical Field

[0001] This application relates to the field of intelligent control, and in particular to a method, controller, medium, and product for assessing checked baggage based on vision processing. Background Technology

[0002] Self-service baggage check-in systems use automated conveyor equipment to transport passengers' baggage to the subsequent processing area. However, since checked baggage includes both regular boxes and irregularly shaped bags such as soft bags and backpacks, these bags may get stuck or become unstable when passing through the turns of the conveyor line or entering the security checkpoint, causing the equipment to stop and requiring manual intervention.

[0003] To address the aforementioned issues, relevant technologies typically incorporate a volume measurement subsystem at the front end of the conveyor equipment. This subsystem uses 3D scanning equipment to acquire point cloud data of the baggage and calculates the length, width, and height of the baggage's minimum bounding rectangle. These values ​​are then compared with the conveyor line's clearance threshold. If the limits are exceeded, the baggage is refused, thus intercepting oversized baggage.

[0004] However, the calculation of the circumscribed rectangle in related technologies essentially treats irregular areas such as recesses and hollows of luggage as solid fillers and equivalent to regular cuboids. In actual transportation, whether luggage can pass through does not entirely depend on the extreme values ​​of the calculated length, width, and height of the circumscribed rectangle, but more on whether its specific surface contour will spatially interfere with the boundary of the equipment in the movement trajectory. Therefore, it is difficult to distinguish whether the size overrun is caused by a compatible tilt or by a rigid structure with a risk of jamming by static comparison of the extreme values ​​of the dimensions alone. This increases the risk of misjudging irregular luggage by the automated check-in system and reduces the passage efficiency. Summary of the Invention

[0005] This application provides a visual processing-based baggage assessment method, controller, medium, and product to improve the recognition accuracy and throughput of irregular baggage in self-service baggage check-in systems.

[0006] In a first aspect, this application provides a visual processing-based baggage assessment method applied to the controller of a self-service baggage check-in system, comprising: generating a three-dimensional digital model of the baggage to be checked in based on acquired three-dimensional point cloud data and image data of the baggage; identifying and marking local region attributes of the three-dimensional digital model based on geometric feature data and texture feature data extracted from the three-dimensional digital model, wherein the local region attributes include rigid abrupt change regions or flexible transition regions; inputting the marked three-dimensional digital model into a preset virtual channel model and calculating the intersection volume between the three-dimensional digital model and the virtual channel model; when the intersection volume is greater than a preset intersection threshold, acquiring the local region attributes of the local region corresponding to the intersection part in the three-dimensional digital model; if the local region attribute is a rigid abrupt change region, determining that the baggage to be checked in has a risk of obstruction and generating a rejection instruction; if the local region attribute is a flexible transition region, calculating the ratio of the intersection volume to the total volume of the three-dimensional digital model to obtain the intersection volume ratio; when the intersection volume ratio is less than a preset deformation threshold, generating a permission instruction for the baggage to be checked in.

[0007] By adopting the above technical solution, the controller first constructs a high-precision 3D digital model using 3D point cloud data and image data. Then, instead of directly judging based on external dimensional extremes, it uses a virtual channel model for dynamic simulation to locate the intersection of baggage and the boundary of the conveyor equipment. Next, the controller performs attribute analysis on this intersection: for areas marked as rigid abrupt changes, it directly intercepts to protect the equipment; for flexible transition areas, it quantifies the deformation requirement by calculating the intersection volume ratio, allowing passage only when the deformation ratio is within a safe range. In summary, this solution, while ensuring that the conveyor equipment does not experience rigid jamming, utilizes the compressible characteristics of flexible baggage to improve the throughput of irregular checked baggage and the service efficiency of the self-service baggage check-in system.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the geometric feature data and texture feature data extracted from the three-dimensional digital model specifically include: extracting texture feature data of the three-dimensional digital model based on image data; and extracting geometric feature data of multiple cross-sectional contours of the three-dimensional digital model along a preset travel direction of the baggage to be checked.

[0009] By adopting the above technical solution, this solution constructs a multimodal data foundation that integrates visual texture and motion cross-sectional geometric attributes, enabling the three-dimensional digital model to not only possess static spatial coordinate information, but also dynamic features that can reflect its material properties and passability, providing data support for subsequent identification of local area attributes.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, based on geometric feature data and texture feature data extracted from the three-dimensional digital model, the local region attributes of the three-dimensional digital model are identified and marked. Specifically, this includes: identifying the material type of the baggage to be checked based on the texture feature data, where the material type includes hard materials and soft materials; when the material type is hard material, marking the local region of the three-dimensional digital model as a rigid abrupt change region; when the material type is soft material, calculating the rate of change of geometric moment features between adjacent cross-sectional contours based on the geometric moment features of each cross-sectional contour; if the rate of change of geometric moment features is greater than a preset change threshold, and the spatial dispersion of the point cloud of the local region of the three-dimensional digital model corresponding to the cross-sectional contour is greater than a preset dispersion threshold, then the local region is marked as a rigid abrupt change region; if the rate of change of geometric moment features is less than or equal to the preset change threshold, then the local region of the three-dimensional digital model corresponding to the cross-sectional contour is marked as a flexible transition region.

[0011] By employing the above technical solution, the controller first utilizes texture features to quickly distinguish materials with a distinctly hard appearance, directly identifying hard material regions as rigid abrupt change regions. Secondly, for regions where the texture appears as soft materials, the controller further analyzes their internal geometry, capturing dramatic fluctuations in cross-sectional shape by calculating the rate of change of geometric moments. Finally, combining point cloud spatial dispersion, regions that appear soft but exhibit abrupt geometric changes and high surface dispersion (such as rigid skeletons or fillers within soft packaging) are corrected and marked as rigid abrupt change regions, while regions with gradual shape changes are identified as flexible transition regions. In summary, this solution eliminates the blind spots of judging solely by appearance and material, accurately identifies potential rigid jamming points in soft-pack equipment, and improves the accuracy of local area attribute marking.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, if the rate of change of the geometric moment features is greater than a preset change threshold, and the spatial dispersion of the point cloud of a local region of the three-dimensional digital model corresponding to the cross-sectional profile is greater than a preset dispersion threshold, then the local region is marked as a rigid abrupt change region. Specifically, this includes: if the rate of change of the geometric moment features is greater than the preset change threshold, calculating the spatial dispersion of the point cloud of the local region of the three-dimensional digital model corresponding to the cross-sectional profile; when the spatial dispersion of the point cloud is less than or equal to the preset dispersion threshold, marking the local region as a flexible transition region; and when the spatial dispersion of the point cloud is greater than the preset dispersion threshold, marking the local region as a rigid abrupt change region.

[0013] By adopting the above technical solution, when the controller detects a large change rate in the geometric moment characteristics, i.e. a sudden change in shape, it does not make a hasty judgment, but further calculates the spatial dispersion of the point cloud in the local area, distinguishing between flexible wrinkles and rigid edges. This avoids misjudging irregularly shaped but passable soft bags as hard luggage with the risk of obstruction, and further reduces the false alarm rate of the automated baggage check-in system for irregular baggage.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, the marked three-dimensional digital model is input into a preset virtual channel model, and the intersection volume between the three-dimensional digital model and the virtual channel model is calculated. Specifically, this includes: along a preset transport path, calculating the area of ​​the overlapping region between the cross-sectional profile of the three-dimensional digital model at different positions and the corresponding boundary constraint cross-section in the virtual channel model, where the boundary constraint cross-section is the spatial boundary that allows baggage to pass through, and the overlapping area is the area of ​​the cross-sectional profile that extends beyond the spatial boundary; and integrating the area of ​​each overlapping region along the preset transport path to obtain the intersection volume between the three-dimensional digital model and the virtual channel model.

[0015] By employing the above technical solution, the controller reduces the dimension of collision detection in three-dimensional space, calculating the area of ​​the two-dimensional overlapping region between the cross-sectional profile and the boundary constraint cross-section along the preset transport path. Then, the controller integrates the discrete area values ​​to reconstruct the total intersection volume between the three-dimensional digital model and the virtual channel model. In summary, this solution calculates the physical quantity of luggage intruding into the transport clearance, providing quantitative data support for subsequent determination of whether flexible materials can be safely compressed and passed based on volume ratio.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, before the step of calculating the area of ​​the overlapping region between the cross-sectional profile of the three-dimensional digital model at different positions and the corresponding boundary constraint section in the virtual channel model along a preset transport path, the method further includes: obtaining the extreme values ​​of the edge coordinates of the cross-sectional profile and the preset boundary coordinate limit range of the boundary constraint section; if the extreme values ​​of the edge coordinates do not exceed the boundary coordinate limit range, then determining that the area of ​​the overlapping region between the cross-sectional profile and the boundary constraint section at the current position is zero; if the extreme values ​​of the edge coordinates exceed the boundary coordinate limit range, then performing the step of calculating the area of ​​the overlapping region between the cross-sectional profile and the corresponding boundary constraint section in the virtual channel model.

[0017] By adopting the above technical solution, before performing complex area overlap calculations, the controller first performs a rapid comparison between the extreme values ​​of the cross-sectional contour edge coordinates and the boundary coordinate limits. Cross-sections whose coordinate extreme values ​​are completely within the limits are directly determined to have zero overlap area, skipping time-consuming Boolean operations and geometric intersection analysis. Finally, refined calculations are only initiated for potentially risky cross-sections with out-of-limit coordinates. In summary, this solution, while ensuring calculation accuracy, reduces the computational load on the controller and improves the processing speed of the virtual simulation and evaluation process by quickly eliminating a large number of non-interference cross-sections.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, after calculating the ratio of the intersecting volume to the total volume of the three-dimensional digital model to obtain the intersecting volume ratio, the method further includes: generating a baggage rejection instruction when the intersecting volume ratio is greater than or equal to a preset deformation threshold.

[0019] By adopting the above technical solution, the controller identifies that although the baggage to be checked in is made of soft material, the required compression deformation exceeds the safety limit of its material or internal contents, and therefore refuses to check in the baggage. This solution not only considers the safety of the equipment that can pass through, but also takes into account the safety of the baggage that may be damaged under large-scale deformation, avoiding the risk of baggage being damaged due to excessive compression or being stuck in the passageway due to excessive rebound force.

[0020] In a second aspect, this application provides a controller for a self-service baggage check-in system. The controller includes: one or more processors and a memory; the memory is coupled to one or more processors and is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the controller to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, this application provides a computer-readable storage medium storing computer instructions that, when executed on a controller, cause the controller to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, this application provides a computer program product, including a computer program or instructions that, when executed on a controller, cause the controller to perform the method described in the first aspect and any possible implementation thereof.

[0023] It is understood that the controller provided in the second aspect, the computer-readable storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. Due to the adoption of a differentiated evaluation mechanism based on local area attribute recognition and virtual channel model, it can distinguish between rigid collisions and flexible contacts, effectively solving the problem of high misjudgment rate caused by static one-size-fits-all judgment based solely on the size of the circumscribed rectangle in related technologies. This allows for maximizing the throughput of irregular luggage while ensuring equipment safety.

[0026] 2. By adopting a material recognition logic that combines texture features with the rate of change of geometric moment features, it is possible to identify potential hard skeletons or abrupt structures inside through soft exteriors. This effectively solves the problem of difficulty in identifying rigid structures inside soft bags, which leads to the risk of jamming. As a result, it achieves accurate attribute marking of the physical properties of complex materials and irregularly shaped luggage.

[0027] 3. Because a shape mutation secondary verification method based on point cloud spatial discreteness is adopted, it can accurately distinguish between the natural wrinkles of flexible materials and the sharp edges of rigid materials, effectively solving the problem of misjudging irregularly shaped wrinkled soft bags as oversized baggage in related technologies, thereby reducing the false alarm and interception rate of self-service baggage check-in system for passable irregular baggage. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating a visual processing-based baggage assessment method according to an embodiment of this application.

[0029] Figure 2 This is another flowchart illustrating a visual processing-based baggage assessment method in this application embodiment;

[0030] Figure 3 This is a schematic diagram of the physical device structure of the controller in an embodiment of this application. Detailed Implementation

[0031] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0032] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0033] The following describes the process of the method provided in this implementation. Please refer to [link / reference]. Figure 1 This is a flowchart illustrating a visual processing-based baggage assessment method in an embodiment of this application.

[0034] S101. Based on the acquired 3D point cloud data and image data of the baggage to be checked, generate a 3D digital model of the baggage to be checked.

[0035] Among them, 3D point cloud data refers to the set of spatial coordinates of the surface of the baggage to be checked in, which is collected by a depth camera. Each point contains X, Y, and Z three-dimensional coordinate information and is used to represent the three-dimensional geometric outline of the baggage. Image data refers to the color image of the baggage to be checked in, which is acquired by an RGB camera and contains visual information such as texture, color, and surface pattern. 3D digital model refers to a digital baggage model with geometric structure and texture mapping generated by fusing 3D point cloud data and image data. It is usually stored in the form of a mesh. Each surface unit of the model contains both spatial coordinates and corresponding color or texture information.

[0036] Specifically, when baggage to be checked is placed within the scanning area of ​​the self-service baggage check-in device, the controller triggers the multimodal acquisition component to initiate the data acquisition process. The controller first receives 3D point cloud data from the depth camera and image data from the RGB camera. Using a pre-calibrated camera extrinsic matrix, the controller maps the pixel coordinates in the image data to the spatial coordinate system of the 3D point cloud data, achieving registration and fusion of the point cloud and texture, generating a point cloud with color attributes. Subsequently, the controller performs denoising, filtering, and downsampling preprocessing on the fused point cloud, and uses a triangulation reconstruction algorithm (such as Poisson reconstruction or greedy projection triangulation) to convert the discrete point cloud into a continuous 3D mesh model. This 3D digital model fully preserves the spatial morphological features and surface texture features of the baggage to be checked, providing a data foundation for subsequent attribute recognition and simulation evaluation.

[0037] S102. Based on the geometric feature data and texture feature data extracted from the three-dimensional digital model, identify and mark the local region attributes of the three-dimensional digital model. The local region attributes include rigid abrupt change regions or flexible transition regions.

[0038] Among them, geometric feature data refers to mathematical parameters describing the surface morphology extracted from the mesh structure of the 3D digital model, including the curvature of each vertex, normal direction, rate of change of normal, and edge sharpness; texture feature data refers to visual features extracted from image data mapped onto the surface of the 3D digital model, including texture roughness, edge intensity, color distribution, and material reflection characteristics; local region attributes are used to represent the physical attribute classification labels of different regions on the surface of the 3D digital model; rigid abrupt change regions refer to regions where the surface curvature changes abruptly, the normal changes drastically, and the texture exhibits the characteristics of hard materials, such as the metal edges of a suitcase, plastic wheels, and pull rods—incompressible hard structures; flexible transition regions refer to regions where the surface curvature changes continuously and has the texture characteristics of soft materials such as fabric or leather, such as the fabric surface of a soft bag and the side pockets of a backpack—structures that can undergo elastic deformation.

[0039] Specifically, after acquiring the 3D digital model, the controller initiates the feature extraction and semantic segmentation module. The controller first traverses all mesh faces of the 3D digital model, calculating the Gaussian curvature and average curvature of each vertex, calculating the normal angle between adjacent vertices, and identifying locations of curvature abrupt changes as candidate regions for rigid features. Simultaneously, the controller analyzes the texture images mapped to each facet, identifying rigid texture features such as metallic luster and regular geometric patterns through edge detection algorithms, or flexible texture features such as fabric weave textures through lightweight convolutional neural networks. The controller fuses the geometric feature analysis results with the texture feature analysis results, assigning attribute labels to each mesh facet: regions with abrupt curvature changes and hard textures are marked as rigid abrupt change regions, while regions with gradual curvature transitions and soft textures are marked as flexible transition regions, thus completing the local region attribute labeling for the entire 3D digital model.

[0040] S103. Input the marked 3D digital model into the preset virtual channel model, and calculate the intersection volume between the 3D digital model and the virtual channel model.

[0041] The virtual passageway model refers to a digital twin model established based on the physical structure of the actual transport line. This model contains key spatial boundary information of the transport line, such as the inner radius boundary at the turning point, the geometric constraints of the security inspection machine entrance frame, and the spatial position of the conveyor belt side baffles, which are used to represent the spatial bottleneck areas that luggage may encounter during transport. The intersecting volume refers to the volume value corresponding to the part in three-dimensional space where the marked three-dimensional digital model overlaps with the virtual passageway model, which is used to quantify the degree to which the luggage model exceeds the passageway boundary.

[0042] Specifically, after completing the local area attribute labeling of the 3D digital model, the controller initiates the virtual simulation evaluation process. The controller retrieves a pre-built virtual channel model from memory, which is constructed and stored based on the conveyor line CAD data of the current station. The controller places the labeled 3D digital model into the virtual channel model according to the actual placement posture of the luggage on the conveyor belt (determined based on the spatial position in the point cloud data), simulating the process of the luggage moving along the conveyor line trajectory through key sections. During the simulation, the controller checks frame-by-frame whether the boundary surfaces of the 3D digital model and the virtual channel model spatially overlap. Using Boolean operations for intersection or bounding box-based collision detection algorithms, the controller calculates the overlapping volume of the 3D digital model intruding into the bounded area of ​​the virtual channel model; this volume is the intersection volume.

[0043] Compare the intersection volume with the preset intersection threshold:

[0044] When the intersection volume is less than or equal to the preset intersection threshold, a baggage check-in permission instruction is generated.

[0045] S104. When the intersection volume is greater than the preset intersection threshold, obtain the local region attributes of the local region corresponding to the intersection part in the three-dimensional digital model.

[0046] The preset intersection threshold is a critical value for the intersection volume used to determine whether there is a potential obstacle to passage for luggage. The purpose of this threshold is to filter out intersections that have no actual impact due to measurement errors or extremely small contact. The setting scheme can be achieved by collecting the statistical distribution of the intersection volume of historical luggage that has passed normally and taking the upper limit of the 95% confidence interval as the threshold, or by setting it as the minimum allowable gap volume according to the mechanical tolerance of the equipment. The intersection part refers to the area in the three-dimensional digital model where the specific mesh surface or voxel unit that spatially overlaps with the virtual passage model is located.

[0047] Specifically, after the controller calculates the intersection volume, it immediately compares this value with a preset intersection threshold. If the intersection volume is less than or equal to the preset intersection threshold, it indicates that the interference between the baggage model and the channel clearance is negligible, and the controller directly generates a baggage allowance instruction. If the intersection volume is greater than the preset intersection threshold, it indicates that the baggage may come into substantial contact with the equipment boundary during transport, and the controller initiates a refined risk analysis process. Based on the spatial overlap position information recorded during the simulation, the controller locates which specific mesh patches in the 3D digital model intersect with the virtual channel model and identifies these patches as intersection locations. Subsequently, the controller queries the local region attribute labels assigned to these mesh patches in step S102 and extracts the local region attributes (rigid abrupt change region or flexible transition region) corresponding to the intersection locations.

[0048] S105. If the local area attribute is a rigid change area, it is determined that there is a risk of obstruction for the baggage to be checked, and a baggage refusal instruction is generated.

[0049] Among them, the risk of jamming refers to the potential malfunction that occurs when luggage gets stuck and cannot move forward or causes the equipment to stop due to incompressible hard contact between the rigid structure and the equipment boundary during the transportation process; the refusal to check in instruction refers to the control command issued by the controller to the execution mechanism of the self-service check-in equipment to prohibit the acceptance of the luggage into the transportation system. At the same time, it can be linked to the display screen or voice prompt device to give passengers a prompt message that they need to adjust the luggage posture or go to the manual counter for processing.

[0050] Specifically, after the controller acquires the local area attributes corresponding to the intersecting part, it first determines whether the attribute is a rigid abrupt change region. If the local area attribute of the intersecting part is marked as a rigid abrupt change region, it indicates that the rigid, non-deformable structural parts of the luggage (such as the hard edges of the suitcase, protruding metal handles, plastic wheels, etc.) will rigidly collide with the channel boundary during the conveyor process. Since these rigid abrupt change regions do not have the ability to compress or elastically deform, the controller determines that the collision will inevitably cause the luggage to get stuck at the turn or the security checkpoint entrance, posing a significant risk of obstruction. Based on this determination, the controller immediately generates a refuse-to-check-in instruction and sends the instruction to the conveyor belt control module through the internal communication bus to stop receiving the luggage. At the same time, it sends a prompt message to the human-machine interface, requesting the passenger to readjust the luggage's placement or proceed to the manual check-in counter, thereby intercepting potential obstruction faults in advance before the luggage enters the automated conveyor system.

[0051] Optionally, after generating a rejection command, the controller further initiates a pose adjustment analysis process. The controller performs multiple virtual rotations of the current 3D digital model within the virtual channel model (e.g., rotating 90 degrees or 180 degrees around the Z-axis, or simulating a switch between side-standing and flat placement), and recalculates the intersection volume for each transformed pose. If the controller detects a specific pose where the recalculated intersection volume is less than a preset intersection threshold (i.e., no rigid collision in this pose), the controller marks this feasible pose as a recommended solution and generates feedback information including specific content such as "Please place your luggage vertically" or "Please rotate your luggage 90 degrees." This information is displayed on the screen or announced via voice prompts to prompt passengers to make targeted adjustments, thereby avoiding baggage check-in failures due to placement issues and improving the success rate of baggage check-in.

[0052] S106. If the local region is a flexible transition region, calculate the ratio of the intersecting volume to the total volume of the 3D digital model to obtain the intersecting volume ratio.

[0053] The total volume refers to the overall spatial volume of the three-dimensional digital model, which is calculated by integrating the volume of the closed mesh of the model. The intersection volume ratio is the ratio of the intersection volume to the total volume, which is used to represent the proportion of the luggage model's encroachment on the passage clearance to the total volume of the luggage. This ratio reflects the amount of compression required to eliminate interference through deformation.

[0054] Specifically, when the controller determines that the local area of ​​the intersection is a flexible transition region, it indicates that the interference occurs at a deformable structure such as the fabric surface of a soft-pack or the side pocket of a backpack. Such collisions have the potential to eliminate interference through compressive deformation. The controller needs to further assess whether the required deformation is within the safe tolerance range of the flexible material. The controller first calculates the total volume of the 3D digital model, then divides the intersection volume by the total volume to obtain the intersection volume ratio. The larger this ratio, the greater the required compressive deformation.

[0055] The controller compares the calculated intersection volume ratio with a preset deformation threshold:

[0056] S107. When the ratio of intersecting volumes is less than the preset deformation threshold, a baggage check-in permission instruction is generated.

[0057] The preset deformation threshold refers to the critical value of the maximum deformation ratio that the flexible transition area can withstand. The function of this threshold is to determine whether the flexible material can safely pass through the channel without damage or blockage through its own elastic deformation. The setting scheme can be based on the elastic modulus and safe compression rate of common soft packaging materials (such as canvas and nylon fabric), combined with the dynamic pressure test data of the conveyor equipment, and set at 5%-8% as the upper limit of safe deformation. The allowance for baggage check-in instruction refers to the control command issued by the controller to the execution mechanism of the self-service check-in equipment, which allows the baggage to enter the automated conveyor system. This instruction triggers the normal execution of subsequent check-in processes such as conveyor belt start-up, label printing, and baggage weighing.

[0058] Specifically, if the cross-sectional volume ratio is less than the preset deformation threshold, it indicates that the required deformation is small, and the flexible material can safely pass through the channel bottleneck through its own elasticity. The controller then generates a clearance instruction. The controller sends this instruction to the conveyor belt drive module via the internal communication bus, starting the conveyor belt to send the luggage into the subsequent processing flow; simultaneously, it sends an instruction to the label printing module to complete the printing and affixing of luggage labels; and uploads the luggage's assessment record information to the data management module.

[0059] If the ratio of the intersecting volumes is greater than or equal to the preset deformation threshold, it indicates that although it is a flexible area, excessive compression may cause damage to luggage or blockage of the passage, and the controller generates a refusal to check in instruction.

[0060] Optionally, before calculating and comparing the intersection volume ratio, the controller initiates a material adaptability assessment logic. The controller calls the texture feature data extracted in S102 and uses an image classification algorithm to identify the material category of the main body of the baggage to be checked. If the identification result is a highly elastic material (such as nylon fabric for sports soft packs), the controller selects a higher dynamic deformation threshold (e.g., 8%); if the identification result is a low-elastic material (such as molded leather), the controller adjusts and selects a lower dynamic deformation threshold (e.g., 3%). Subsequently, the controller compares the calculated intersection volume ratio with the material-corrected dynamic deformation threshold. When the intersection volume ratio is less than the current dynamic deformation threshold, it indicates that the required compression amount is within the physical tolerance range of the specific material. Based on this, the controller determines that the baggage is allowed and generates a baggage clearance instruction, thereby achieving refined release control based on the physical characteristics of the material.

[0061] In this embodiment, by employing the identification of local region attributes of a 3D digital model based on geometric and texture feature data, and calculating the intersection volume using a virtual channel model, direct interception based on jamming risk or deformation potential assessment based on the intersection volume ratio is performed for intersections with different attributes. Therefore, it is possible to simulate the spatial interference of baggage during transportation and distinguish whether the interference source is an incompressible rigid structure or a flexible structure that allows a certain degree of deformation. This solves the problem in related technologies where static comparison based solely on the extreme values ​​of the circumscribed rectangle dimensions makes it difficult to distinguish whether the size exceeding the standard is caused by substantial jamming risk or by compatible posture tilting or flexible deformation. This, in turn, improves the recognition accuracy and pass rate of irregular baggage in the self-service baggage check-in system.

[0062] The above embodiments mainly illustrate the core logic of differentiated baggage release based on local area attribute recognition and virtual channel simulation. In practical applications, in order to more accurately identify the physical properties of baggage made of complex materials and improve the real-time performance of calculations, it is also necessary to combine specific geometric moment feature extraction and fast filtering mechanisms for further refinement.

[0063] Based on the above embodiments, the method provided in this embodiment will be described in further detail below. Please refer to... Figure 2 This is another flowchart illustrating a visual processing-based baggage assessment method in an embodiment of this application.

[0064] S201. Based on the acquired 3D point cloud data and image data of the baggage to be checked, generate a 3D digital model of the baggage to be checked.

[0065] This step and Figure 1 The description of step S101 in the embodiment is similar and will not be repeated here.

[0066] S202. Based on the image data, extract the texture feature data of the three-dimensional digital model, and along the preset travel direction of the baggage to be checked, extract the geometric feature data of multiple cross-sectional contours of the three-dimensional digital model.

[0067] Among them, the preset travel direction refers to the direction in which the checked baggage moves along the conveyor belt according to the planned direction of the conveyor system, usually with the direction of the conveyor belt movement as the positive direction of the Z-axis; the cross-sectional profile refers to the two-dimensional closed curve formed by the point cloud at a certain fixed position after orthogonally slicing the three-dimensional digital model along the preset travel direction, and this curve describes the cross-sectional shape of the baggage at that position; the geometric feature data refers to the mathematical parameters extracted from the cross-sectional profile that describe the shape features, including the area, perimeter, centroid coordinates, moment of inertia, eccentricity, circularity and other geometric moment features of the cross-sectional profile.

[0068] Specifically, after the controller completes the construction of the 3D digital model, it initiates a dual-stream feature extraction process. First, the controller performs texture analysis on the image data mapped onto the surface of the 3D digital model, extracting and storing texture feature data using a gray-level co-occurrence matrix algorithm or a pre-trained lightweight convolutional neural network. Then, the controller determines a preset travel direction based on the spatial orientation of the luggage in the point cloud data, establishing a coordinate system with this direction as the Z-axis. The controller slices the 3D digital model at equal intervals along the Z-axis with a preset step size (e.g., 5mm), extracting the cross-sectional contour formed by the point cloud projection at each slice location. For each cross-sectional contour, the controller calculates its zeroth moment (area), first moment (centroid coordinates), and second moment (moment of inertia), establishing an index association between these geometric moment features and the corresponding cross-sectional location, and storing them.

[0069] S203. Identify the material type of the baggage to be checked based on texture feature data. The material type includes hard materials and soft materials.

[0070] Material type refers to the classification of the physical properties of the surface material of the baggage to be checked in, based on texture feature data. Rigid materials refer to incompressible or rigid materials with minimal compressibility, including metals, hard plastics, hard-shell cases made of ABS / PC, and fiberglass, whose texture features are usually characterized by high light intensity reflection, smooth single-color blocks, regular geometric patterns, or metallic luster. Soft materials refer to flexible materials that are elastic and can deform, including fabrics, canvas, nylon, leather, and rubber, whose texture features are usually characterized by woven textures, irregular wrinkles, frosted surfaces, or light-absorbing materials.

[0071] Specifically, after the controller completes texture feature data extraction, it activates the material recognition module to classify the material of the checked baggage. The controller inputs the extracted texture feature data into a pre-trained material classifier (which can use a Support Vector Machine (SVM) or a lightweight CNN classification network). This classifier outputs the classification result and confidence score of the material type based on feature vectors such as texture roughness, edge intensity, and reflectivity. For different regions of the 3D digital model, the controller analyzes the corresponding texture features block by block: if high light intensity reflection is detected and the texture is uniform and smooth, it is determined to be a hard material; if woven texture or irregular wrinkles and shadows are detected, it is determined to be a soft material. The controller establishes a mapping relationship between the recognition results and the spatial regions of the 3D digital model, assigning a material type label to each surface region, completing global material type recognition, and providing a classification basis for subsequent material-based rigidity / flexibility attribute labeling.

[0072] S204. When the material type is hard material, mark local areas of the 3D digital model as rigid abrupt change regions.

[0073] Specifically, the controller traverses all surface areas of the 3D digital model, retrieving the corresponding material type label for each area. For local areas marked as hard materials, the controller directly determines that these local areas physically lack compressibility and will experience rigid contact if they collide with the equipment boundary during transport. Based on this physical property judgment, the controller batch-marks the mesh patches or voxel elements of the 3D digital model corresponding to these hard material areas as rigidity abrupt change regions.

[0074] S205. When the material type is soft material, calculate the rate of change of geometric moment characteristics between adjacent cross-sectional profiles based on the geometric moment characteristics of each cross-sectional profile.

[0075] Among them, geometric moment features refer to mathematically invariant moment parameters describing the shape of the cross-section profile, including the zeroth moment (representing the cross-sectional area), the first moment (representing the position of the center of mass), and the second moment (representing the moment of inertia and eccentricity of the shape). These features are invariant to translation, rotation, and scale changes, and can stably describe the cross-sectional shape. The rate of change of geometric moment features refers to the ratio of the difference in corresponding geometric moment features between adjacent cross-sectional profiles to the distance between the cross-sections. It is used to quantify the degree of change in the cross-sectional shape along a preset direction of travel. The calculation formula is as follows: Where M is the geometric moment eigenvalue of a certain cross-sectional profile, This represents the slice spacing.

[0076] Specifically, when the controller determines that the material of the baggage to be checked is soft, it cannot directly mark it as a flexible transition region. This is because when the baggage to be checked is made of purely soft material (such as a bag full of clothes), its cross-sectional shape changes gradually and continuously along the Z-axis (similar to a fluid or smooth surface). The calculated... The values ​​are small and fluctuate gently. When checked baggage contains rigid abrupt changes in structure (such as a square hard box placed inside a soft compartment, or hard wheels protruding from the side of a luggage bag), the shape distribution of the cross-section will undergo a step abrupt change when the slice scan reaches the edge of the hard object (e.g., from a circular cross-section to an angular cross-section instantaneously), resulting in a second-order central moment. The violent fluctuations necessitate further analysis of geometric features to identify potential rigid structures.

[0077] The controller reads the geometric moment characteristics of each cross-section profile sequentially according to their location, selecting either the area moment or the moment of inertia as the comparison object. The controller calculates the rate of change of the geometric moment characteristics between adjacent cross-section profiles. For example, if the first-order moment is used to calculate the rate of change of the geometric moment characteristics, then... ,in, Let be the first moment of the i-th cross section.

[0078] S206. If the rate of change of the geometric moment feature is greater than the preset change threshold, and the spatial dispersion of the point cloud of the local region of the three-dimensional digital model corresponding to the cross-sectional profile is greater than the preset dispersion threshold, then the local region is marked as a rigid abrupt change region.

[0079] This step specifically includes:

[0080] If the rate of change of the geometric moment features is greater than the preset change threshold, then the point cloud spatial dispersion of the local region of the three-dimensional digital model corresponding to the cross-sectional profile is calculated.

[0081] When the spatial dispersion of the point cloud is less than or equal to the preset dispersion threshold, the local region is marked as a flexible transition region.

[0082] When the spatial dispersion of the point cloud is greater than the preset dispersion threshold, the local region is marked as a rigid abrupt change region.

[0083] Among them, the spatial dispersion of point cloud refers to the degree of dispersion of point cloud from the fitting plane in a local area, which is obtained by calculating the standard deviation (RMSE) of the Euclidean distance from each neighboring point in the region to the least square fitting plane; the preset dispersion threshold is the critical value of point cloud dispersion used to distinguish the regularity of surface microstructure. The function of this threshold is to distinguish between smooth soft surfaces and hard surfaces with sharp edges or complex structures. Its setting scheme can be set to 3 times the measurement noise (about 6 mm) according to the measurement accuracy of the depth camera (e.g., ±2 mm), or determined by collecting point cloud samples of hard edges and soft fabrics and statistically analyzing the difference in dispersion.

[0084] Specifically, when the controller detects that the rate of change of geometric moment characteristics at a certain point exceeds a preset threshold, it indicates that a sudden change has occurred in the cross-sectional shape at that location, but further determination of the cause of the change is still required. The controller locates a local region of the 3D digital model corresponding to the cross-sectional contour and extracts the original point cloud data of that region. The controller selects all 3D point cloud data within this local region and uses Principal Component Analysis (PCA) or least squares method to fit a local tangent plane, calculates the perpendicular distance from each point to the local tangent plane, and calculates the root mean square error of these distances as the point cloud spatial dispersion. The controller compares the calculated point cloud spatial dispersion with a preset dispersion threshold:

[0085] If the dispersion is greater than the threshold, it indicates that there are micro-complex structures such as sharp edges, zipper teeth, and rigid fasteners on the surface of the local area, and it is determined to be a rigid abrupt change region.

[0086] If the dispersion is less than or equal to the threshold, it means that although the shape of the local area changes abruptly, the surface is smooth and continuous. It may be soft luggage with large deformation but small volume, such as shoulder straps of handbags, and is judged as a flexible transition area.

[0087] S207. If the rate of change of the geometric moment feature is less than or equal to the preset change threshold, then the local area of ​​the three-dimensional digital model corresponding to the cross-sectional profile is marked as a flexible transition area.

[0088] Specifically, when the controller iterates through the geometric moment characteristic change rate of all adjacent cross-sectional contours, for cross-sectional positions where the change rate is less than or equal to a preset change threshold, it determines that the cross-sectional shape at that location exhibits a smooth and continuous transition characteristic along a preset direction of travel, without any drastic geometric abrupt changes. Considering that the region has been identified as a soft material in step S203, the controller infers that this local area possesses both the texture characteristics of a soft material and the geometric characteristics of a smooth shape transition, conforming to the dual characteristics of typical flexible materials (such as fabric or backpack side pockets). Based on this comprehensive judgment, the controller marks the local area of ​​the 3D digital model corresponding to the cross-sectional contour as a flexible transition region.

[0089] This marking indicates that the area can undergo elastic deformation when subjected to external pressure, providing a flexible property criterion for the deformation potential assessment in subsequent virtual channel simulations, ensuring that the real soft-pack flexible part is not misjudged as a rigid jamming risk.

[0090] S208. Input the marked 3D digital model into the preset virtual passage model, obtain the extreme values ​​of the edge coordinates of the cross-sectional profile and the preset boundary coordinate limit range of the boundary constraint section in the virtual passage model. The boundary constraint section is the spatial boundary of the virtual passage model that allows baggage to pass through, and the overlapping area is the area of ​​the part of the cross-sectional profile that exceeds the spatial boundary.

[0091] Among them, the extreme values ​​of the edge coordinates refer to the maximum and minimum coordinate values ​​of the cross-sectional profile in the two-dimensional plane coordinate system, including the X-direction. and , Y direction and , is used to represent the spatial envelope range of the cross-sectional profile; the boundary coordinate limit range refers to the preset allowable coordinate range of the boundary constraint cross-section, which defines the spatial limit of the conveying channel at that location, and is usually represented by a two-dimensional polygon or rectangle, for example, the width of the security inspection machine entrance is (-400mm, +400mm) and the height is (0, 600mm).

[0092] Specifically, after the controller completes the local area attribute labeling of the 3D digital model, it initiates the virtual passageway simulation evaluation process. The controller loads a preset virtual passageway model from memory. This model consists of a series of boundary constraint sections distributed along a preset transport path, each defining the spatial boundary allowing luggage passage at its corresponding location. At this point, the 3D digital model and the virtual passageway model are in the same unified coordinate system. The controller places the labeled 3D digital model into the starting position of the virtual passageway model according to its actual placement posture, preparing for step-by-step simulation along the preset transport path. For each cross-sectional profile of the 3D digital model, the controller calculates its edge coordinate extreme values ​​on the 2D plane, and simultaneously obtains the boundary coordinate limit range of the boundary constraint section at the corresponding location in the virtual passageway model.

[0093] S209. If the extreme value of the edge coordinates does not exceed the limit range of the boundary coordinates, then the area of ​​the overlapping region between the cross-sectional profile at the current position and the boundary constraint cross-section is determined to be zero.

[0094] Specifically, the controller compares the extreme values ​​of the edge coordinates of the cross-sectional profile with the boundary coordinate limits of the boundary-constrained cross-section dimension by dimension. Controller checks... Is it greater than the lower limit of the x-direction boundary? Is it less than the upper limit of the boundary in the x-direction? Also check... and Is it within the boundary y-direction range? If all edge coordinate extreme values ​​do not exceed the boundary coordinate limit range, it indicates that the cross-sectional profile is completely within the allowable passage area of ​​the boundary-constrained cross-section, and there is no spatial interference.

[0095] Based on this determination, the controller directly sets the current position. The area of ​​the overlapping region between the cross-sectional profile and the boundary constraint cross-section is set to zero, eliminating the need for complex Boolean operations and improving computational efficiency. The controller stores this zero-value result in the overlapping area array and then continues processing the comparison of the next cross-sectional position.

[0096] S210. If the extreme values ​​of the edge coordinates exceed the limit range of the boundary coordinates, calculate the area of ​​the overlapping region between the cross-sectional profile of the three-dimensional digital model at different positions and the corresponding boundary constraint cross-section in the virtual channel model along the preset transport path.

[0097] Specifically, when the controller detects that the extreme values ​​of the edge coordinates of the cross-sectional profile exceed the boundary coordinate limits, it indicates that there is spatial interference at that cross-sectional location, requiring precise calculation of the overlapping area. The controller then initiates a refined simulation process, progressively advancing the 3D digital model from its starting position according to the time sequence of the preset transport path, at each discrete position. At that location, the controller extracts the cross-sectional profile of the 3D digital model. Simultaneously, obtain the corresponding boundary constraint sections in the virtual channel model. The controller controls the cross-sectional profile. With boundary constraint section Perform two-dimensional Boolean operations: First, calculate the complement of the boundary constraint sections, i.e., the forbidden region. Then find the intersection of the cross-sectional profile and the prohibited area. This intersection represents the portion of the cross-sectional profile that extends beyond the spatial boundary. The controller uses a polygon area calculation algorithm (such as Green's formula) to calculate the area of ​​this intersection region. The calculation results are stored in the overlapping area array, completing the area calculation of the overlapping area at that location. The controller performs area calculations for all key locations one by one along the preset conveyor path.

[0098] S211. Along the preset transport path, the area of ​​each overlapping region is integrated to obtain the intersection volume between the three-dimensional digital model and the virtual channel model.

[0099] Specifically, the controller retrieves the stored overlap area array. Each element corresponds to the area of ​​the overlapping region at a discrete location on the preset conveying path. The controller uses a discrete integration method (trapezoidal rule or rectangular rule) to numerically integrate these area values ​​to obtain the intersecting volume. The calculation formula is as follows: ,in This represents the spacing between adjacent cross-sectional positions (i.e., the slice step size). This integral operation essentially involves accumulating and multiplying all out-of-limit cross-sectional surfaces along the preset transport path direction by the micro-element thickness to obtain the total volume of the luggage model intruding into the passageway boundary in three-dimensional space. The controller will then calculate the intersecting volumes. The data is stored as a scalar value that quantifies the degree of spatial interference between the baggage to be checked and the virtual passageway model. The controller then transmits this intersection volume to the decision module, compares it with a preset intersection threshold, and provides a key quantitative indicator of volume interference for subsequent passage determination, thus completing the dimensional upgrade from two-dimensional cross-sectional analysis to three-dimensional volume assessment.

[0100] S212. When the intersection volume is greater than the preset intersection threshold, obtain the local region attributes of the local region corresponding to the intersection part in the three-dimensional digital model.

[0101] S213. If the local area attribute is a rigid change area, it is determined that there is a risk of obstruction for the baggage to be checked in, and a baggage refusal instruction is generated.

[0102] S214. If the local region is a flexible transition region, calculate the ratio of the intersecting volume to the total volume of the 3D digital model to obtain the intersecting volume ratio.

[0103] S215. When the ratio of intersecting volumes is less than the preset deformation threshold, a baggage check-in permission instruction is generated.

[0104] S216. When the ratio of intersecting volumes is greater than or equal to the preset deformation threshold, a rejection instruction for the baggage to be checked in is generated.

[0105] Steps S212 to S216 and Figure 1 The descriptions of steps S104 to S107 in the embodiment are similar and will not be repeated here.

[0106] In this embodiment, by employing secondary verification and marking of the local regional properties of soft materials, and by pre-comparing the extreme values ​​of the edge coordinates of the cross-sectional profile with the boundary coordinate limits before calculating the intersection volume, it is possible to quickly eliminate cross-sections without spatial interference to reduce computational load. At the same time, it can accurately identify potential rigid abrupt structures or distinguish safe surface wrinkles through the soft appearance, and quantify the deformation potential of the flexible transition region based on the intersection volume ratio. This solves the problems of high false alarm rate caused by the difficulty in distinguishing between soft folds and rigid edges in related technologies, as well as the large consumption of full-scale three-dimensional simulation computational resources and low efficiency. This improves the accuracy and real-time performance of risk assessment for irregular baggage.

[0107] The controller in the embodiments of this invention is described below from the perspective of hardware processing. Please refer to [link / reference]. Figure 3 This is a schematic diagram of the physical device structure of the controller in an embodiment of this application.

[0108] It should be noted that, Figure 3 The controller structure shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0109] like Figure 3As shown, the controller includes a CPU 301, which can perform various appropriate actions and processes based on a program stored in the read-only memory ROM 302 or a program loaded from the storage section 308 into the random access memory RAM 303, such as performing the methods described in the above embodiments. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An I / O interface 305 is also connected to the bus 304.

[0110] The following components are connected to I / O interface 305: input section 306 including audio input devices, push-button switches, etc.; output section 307 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 308 including a hard disk, etc.; and communication section 309 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 309 performs communication processing via a network such as the Internet. Drive 310 is also connected to I / O interface 305 as needed. Removable media 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 310 as needed so that computer programs read from them can be installed into storage section 308 as needed.

[0111] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program / instructions carried on a computer-readable medium, the computer program / instructions containing computer program / instructions for performing the methods shown in the flowcharts. In such embodiments, the computer program / instructions can be downloaded and installed from a network via communication section 309, and / or installed from removable medium 311. When the computer program is executed by CPU 301, it performs the various functions defined in the present invention.

[0112] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0113] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0114] Specifically, the controller in this embodiment includes a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements the visual processing-based baggage assessment method provided in the above embodiment.

[0115] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the controller described in the above embodiments; or it may exist independently and not assembled into the controller. The storage medium carries one or more computer programs that, when executed by a processor of the controller, cause the controller to implement a vision-based baggage assessment method provided in the above embodiments.

[0116] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0117] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0118] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A vision-based baggage assessment method, applied to the controller of a self-service baggage check-in system, characterized in that, include: Based on the acquired 3D point cloud data and image data of the baggage to be checked, a 3D digital model of the baggage to be checked is generated. Based on the geometric and texture feature data extracted from the three-dimensional digital model, the local region attributes of the three-dimensional digital model are identified and marked, including rigid abrupt regions or flexible transition regions. The marked three-dimensional digital model is input into a preset virtual channel model, and the intersection volume between the three-dimensional digital model and the virtual channel model is calculated. When the intersection volume is greater than a preset intersection threshold, the local region attributes of the local region corresponding to the intersection part in the three-dimensional digital model are obtained; If the local area attribute is the rigid abrupt change area, it is determined that the baggage to be checked in is at risk of being blocked, and a baggage refusal instruction is generated; If the local region attribute is the flexible transition region, calculate the ratio of the intersection volume to the total volume of the three-dimensional digital model to obtain the intersection volume ratio; When the intersection volume ratio is less than a preset deformation threshold, a check-in permission instruction is generated for the baggage to be checked.

2. The method according to claim 1, characterized in that, The geometric feature data and texture feature data extracted from the three-dimensional digital model specifically include: Based on the image data, extract the texture feature data of the three-dimensional digital model; Geometric feature data of multiple cross-sectional contours of the three-dimensional digital model are extracted along the preset travel direction of the baggage to be checked.

3. The method according to claim 2, characterized in that, The step of identifying and marking local region attributes of the 3D digital model based on the geometric and texture feature data extracted from the 3D digital model specifically includes: The material type of the baggage to be checked is identified based on the texture feature data, and the material type includes hard materials and soft materials; When the material type is a hard material, a local area of ​​the three-dimensional digital model is marked as a rigidity abrupt change region; When the material type is a soft material, the rate of change of geometric moment characteristics between adjacent cross-sectional profiles is calculated based on the geometric moment characteristics of each cross-sectional profile. If the rate of change of the geometric moment feature is greater than a preset change threshold, and the spatial dispersion of the point cloud of the local region of the three-dimensional digital model corresponding to the cross-sectional contour is greater than a preset dispersion threshold, then the local region is marked as a rigid abrupt change region. If the rate of change of the geometric moment feature is less than or equal to a preset change threshold, then the local area of ​​the three-dimensional digital model corresponding to the cross-sectional profile is marked as a flexible transition region.

4. The method according to claim 3, characterized in that, If the rate of change of the geometric moment feature is greater than a preset change threshold, and the spatial dispersion of the point cloud of a local region of the 3D digital model corresponding to the cross-sectional contour is greater than a preset dispersion threshold, then the local region is marked as a rigid abrupt change region, specifically including: If the rate of change of the geometric moment feature is greater than a preset change threshold, then the point cloud spatial dispersion of the local region of the three-dimensional digital model corresponding to the cross-sectional contour is calculated. When the spatial dispersion of the point cloud is less than or equal to a preset dispersion threshold, the local region is marked as a flexible transition region. When the spatial dispersion of the point cloud is greater than a preset dispersion threshold, the local region is marked as a rigid abrupt change region.

5. The method according to claim 1, characterized in that, The step of inputting the marked 3D digital model into a preset virtual channel model and calculating the intersection volume between the 3D digital model and the virtual channel model specifically includes: Along the preset transport path, calculate the overlapping area between the cross-sectional profile of the three-dimensional digital model at different positions and the corresponding boundary constraint section in the virtual channel model. The boundary constraint section is the spatial boundary that allows baggage to pass through, and the overlapping area is the area of ​​the portion of the cross-sectional profile that extends beyond the spatial boundary. Along the preset transport path, the area of ​​each overlapping region is integrated to obtain the intersection volume between the three-dimensional digital model and the virtual channel model.

6. The method according to claim 5, characterized in that, Before the step of calculating the area of ​​the overlapping region between the cross-sectional profile of the three-dimensional digital model at different positions along the preset transport path and the corresponding boundary constraint cross-section in the virtual channel model, the method further includes: Obtain the extreme values ​​of the edge coordinates of the cross-sectional profile and the preset boundary coordinate limit range of the boundary constraint cross section; If the extreme value of the edge coordinates does not exceed the limit range of the boundary coordinates, then the area of ​​the overlapping region between the cross-sectional profile at the current position and the boundary constraint cross-section is determined to be zero. If the extreme value of the edge coordinates exceeds the limit range of the boundary coordinates, then the step of measuring the overlapping area between the cross-sectional profile and the corresponding boundary constraint cross-section in the virtual channel model is executed.

7. The method according to claim 1, characterized in that, After the step of calculating the ratio of the intersecting volume to the total volume of the three-dimensional digital model to obtain the intersecting volume ratio, the method further includes: When the intersection volume ratio is greater than or equal to a preset deformation threshold, a rejection instruction for the baggage to be checked in is generated.

8. A controller for a self-service baggage check-in system, characterized in that, The controller includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the controller to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium storing computer instructions, characterized in that, When the computer instructions are executed on the controller, the controller causes the controller to perform the method as described in any one of claims 1-7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are run on the controller, the controller causes the controller to perform the method as described in any one of claims 1-7.

Images