Stacked piece detection method, system and equipment in automatic express sorting and medium

Abstract

The invention provides a stacked piece detection method, system and device in automatic express sorting and a medium, and belongs to the technical field of logistics sorting, and the method comprises the steps: obtaining a real-time depth image of an express piece on a conveying belt; converting the real-time depth image into point cloud data; setting a dynamic neighborhood radius in a DBSCAN clustering algorithm, and adding curvature similarity constraint in judgment of a core point in a neighborhood to obtain an improved clustering algorithm; and segmenting the point cloud data by using an improved clustering algorithm, calculating a gradient change track from the highest point to the lowest point, and carrying out stack detection according to the gradient change track of the stack. The overlapped express items can be quickly and accurately identified, and the subsequent express item processing efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of logistics sorting, and specifically relates to a method, system, equipment and medium for detecting overlapped items in automatic express sorting. Background Art

[0002] In the express logistics industry, sorting is a critical step in ensuring that packages reach their destinations accurately and efficiently. With the rapid development of e-commerce, express delivery volume has increased dramatically, and the requirements for sorting efficiency and accuracy are also increasing.

[0003] Traditional sorting methods rely primarily on manual labor, which is not only inefficient but also prone to errors. To improve sorting efficiency and accuracy, the express delivery industry has begun to introduce automated sorting systems. Automatic express delivery sorting technology is a key technology in modern logistics, significantly improving the efficiency of logistics distribution. In automated sorting systems, multiple packages may be stacked together, especially flat items such as documents and printed materials, which are prone to overlap and require accurate identification.

[0004] Existing deep learning algorithms for stacked parts detection require collecting and properly annotating large amounts of stacked and individual part images or point cloud data. This data is then trained and deployed into an artificial neural network model for stacked parts detection. The model's performance depends on data quality and model training, resulting in limited detection accuracy. Summary of the Invention

[0005] In order to solve the problem of accuracy in detecting overlapped items in automatic express sorting, the present invention provides a method, system, device and medium for detecting overlapped items in automatic express sorting.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] A method for detecting overlapping items in automatic express sorting, comprising the following steps:

[0008] Acquire a real-time depth image of the express item on the conveyor belt; convert the real-time depth image into point cloud data;

[0009] In the DBSCAN clustering algorithm, a dynamic neighborhood radius is set, and a curvature similarity constraint is added to the judgment of the core points in the neighborhood to obtain an improved clustering algorithm;

[0010] The point cloud data is segmented using the improved clustering algorithm, a gradient change trajectory from the highest point to the lowest point is calculated, and overlapping detection is performed based on the gradient change trajectory to obtain overlapping detection results.

[0011] Preferably, the processing steps of the improved clustering algorithm specifically include:

[0012] S1. Randomly select an unvisited point in the point cloud data and check the number of points within a neighborhood radius ε of the point; wherein the neighborhood radius is a dynamic neighborhood radius;

[0013] S2. If the number of neighborhood points is ≥ MinPts, mark it as a core point and generate a new cluster; otherwise, mark it as noise; where MinPts is the minimum number of points, indicating the minimum number of points that must be included in the ε neighborhood;

[0014] S3. Recursively traverse all neighboring points within the neighborhood radius of the current core point, and add the reachable core points and boundary points to the same cluster; by checking the curvature difference, determine whether the two core points in the neighborhood belong to the same cluster. If the curvature difference is less than the threshold, they belong to the same cluster, otherwise they belong to different clusters;

[0015] S4, repeat steps S1 to S3 until all points are visited;

[0016] S5. If the clustering result is 2 clusters or more, it is judged as a stacked part. If it is 1 cluster, further judgment is made by projecting the segmented point cloud into a 2.5D height map, obtaining the coordinates (x1, y1) of the maximum height point A and the coordinates (x2, y2) of the minimum height point B. Calculate the Sobel operator gradient of each point on any path from A to B. If there is a point with a gradient value greater than the gradient threshold, it is judged as a stacked part, otherwise it is a single part.

[0017] Preferably, the method further includes pre-processing the point cloud data before segmenting the point cloud data using the improved clustering algorithm, specifically including the following steps:

[0018] Filtering the point cloud data using a voxel grid to perform downsampling;

[0019] Set a height threshold to remove the background on the conveyor belt surface;

[0020] By calculating the neighborhood distance distribution of each point, points that deviate from the mean by more than a threshold are removed to eliminate noise points in the point cloud.

[0021] Preferably, the dynamic neighborhood radius is calculated based on the local height variance σ z Adjustments, specifically:

[0022]

[0023] ε=k·σ z ;

[0024] in, is the mean of the z coordinates in the neighborhood, i is the serial number of the point within the neighborhood radius, N is the number of neighborhood points, and k is the empirical coefficient.

[0025] Preferably, the curvature similarity constraint processing process is specifically as follows:

[0026] Use principal component analysis (PCA) to calculate the minimum eigenvalue λ of the point neighborhood covariance matrix min ;

[0027] The curvature C is calculated based on the minimum eigenvalue, specifically using the following formula:

[0028]

[0029] When judging whether two points belong to the same cluster, |C p -C q |<δ C ; Among them, δ C is the curvature difference threshold.

[0030] Preferably, the real-time depth image of the express items on the conveyor belt is obtained by installing a depth camera above the conveyor belt, adjusting the field of view to cover the entire width of the conveyor belt, and obtaining the depth image of the express items in real time.

[0031] The present invention also provides a system for detecting overlapping items in automatic express sorting, which specifically includes:

[0032] The data acquisition module is used to obtain real-time depth images of express items on the conveyor belt and convert the real-time depth images into point cloud data.

[0033] The algorithm improvement module is used to set the dynamic neighborhood radius in the DBSCAN clustering algorithm and add curvature similarity constraints to the judgment of core points in the neighborhood to obtain an improved clustering algorithm.

[0034] The overlapping part identification module is used to segment the point cloud data using the improved clustering algorithm, calculate the gradient change trajectory from the highest point to the lowest point, perform overlapping part detection based on the gradient change trajectory, and obtain overlapping part detection results.

[0035] The present invention also provides a computer device comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps described in the method for detecting stacked items in automatic express sorting.

[0036] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is loaded by a processor, it can execute the steps described in the method for detecting stacked items in automatic express sorting.

[0037] The present invention provides a method for detecting overlapping items in automatic express delivery sorting, which has the following beneficial effects:

[0038] This invention establishes a dynamic neighborhood radius within the DBSCAN clustering algorithm and incorporates curvature similarity constraints when determining core points within the neighborhood. This results in an improved clustering algorithm capable of discovering clusters of arbitrary size and shape, offering greater flexibility and accuracy when processing point cloud data of complex shapes. Real-time depth images of express items are converted into point cloud data; this point cloud data is processed using an improved clustering algorithm, and the gradient trajectory from the highest point to the lowest point is calculated. Overlap detection is performed based on the gradient trajectory of the overlapping items. Different regions within the image are identified based on the gradient changes, and the target object is accurately segmented, achieving efficient segmentation of overlapping items. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] To more clearly illustrate the embodiments of the present invention and its design, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be derived from these drawings without inventive effort.

[0040] Figure 1 The present invention is a flow chart of a method for detecting overlapping items in automatic express delivery sorting. DETAILED DESCRIPTION

[0041] In order to enable those skilled in the art to better understand the technical solution of the present invention and to be able to implement it, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention and are not intended to limit the scope of protection of the present invention.

[0042] Example

[0043] The present invention provides a method for detecting overlapping items in automatic express sorting, which specifically includes the following steps:

[0044] S1. Data acquisition: Use a depth camera to collect depth images of packages on the conveyor belt in real time.

[0045] S2. Read the depth image, substitute the camera internal parameters, and convert the depth image into point cloud data.

[0046] S3. Preprocess the point cloud data, specifically including the following steps:

[0047] Use voxel grid to filter point cloud data to reduce the amount of data and perform downsampling;

[0048] Set a height threshold to remove the background on the conveyor belt surface;

[0049] By calculating the neighborhood distance distribution of each point, points that deviate from the mean by more than a threshold are removed to eliminate noise points in the point cloud.

[0050] S4. Set a dynamic neighborhood radius in the DBSCAN clustering algorithm, and add a curvature similarity constraint in the judgment of core points within the neighborhood to obtain an improved clustering algorithm.

[0051] S5. Use the improved clustering algorithm to segment the point cloud data, calculate the gradient change trajectory from the highest point to the lowest point, and perform stack detection based on the gradient change trajectory of the stack.

[0052] According to the density characteristics, the data points in the point cloud data are divided into three categories: Core Point: The number of points within its own ε-neighborhood ≥ MinPts; Border Point: The number of points within its own neighborhood < MinPts, but located within the neighborhood of a certain core point; Noise Point: Neither a core point nor a border point. The specific steps are as follows:

[0053] S51. Randomly select an unvisited point and check the number of points within the neighborhood radius ε; where the neighborhood radius is the dynamic neighborhood radius. The dynamic neighborhood radius ε is adjusted according to the local height variance σ z by the following formula:

[0054]

[0055] ε = k·σ z ;

[0056] where, is the mean value of the z-coordinates within the neighborhood, i is the serial number of a certain point within the neighborhood radius, N is the number of neighborhood points, k is an empirical coefficient, usually taking 1.5 - 2.0. In flat areas (σ z is small), ε is reduced; in undulating areas (σ z is large), ε is enlarged.

[0057] S52. If the number of neighborhood points ≥ MinPts, mark it as a core point and generate a new cluster; otherwise, mark it as noise; where MinPts is the minimum number of points, indicating the minimum number of points that need to be included within the ε-neighborhood.

[0058] S53. Recursively traverse all neighborhood points of the core point, and add reachable core points and border points to the same cluster; when judging whether two core points within the neighborhood belong to the same cluster, check the curvature difference. If the curvature difference is less than the threshold, they belong to the same cluster, otherwise they belong to different clusters. The similarity constraint on the curvature specifically includes the following steps:

[0059] Use the principal component analysis PCA to calculate the minimum eigenvalue λ of the point neighborhood covariance matrix min ;

[0060] Calculate the curvature C based on the minimum eigenvalue, specifically through the following formula:

[0061]

[0062] When judging whether two points belong to the same cluster, |C p -C q |<δ C ; where δ C is the curvature difference threshold.

[0063] S54. Repeat steps S51 to S53 until all points are visited.

[0064] S55. If the clustering result is 2 clusters or more than 2 clusters, it is judged as a stacked part; if it is 1 cluster, further judgment is required. The specific method is: project the segmented point cloud into a 2.5D height map, calculate the coordinates of the maximum height point A (x1, y1) and the minimum height point B (x2, y2); calculate the Sobel operator gradient of each point on any path from A to B. If there is a point with a gradient value greater than the gradient threshold, it is judged as a stacked part, otherwise it is a single part.

[0065] S6. Output the overlapping detection result.

[0066] The present invention also provides a system for detecting overlapping items in automatic express sorting, which specifically includes:

[0067] The data acquisition module is used to obtain real-time depth images of express items on the conveyor belt and convert the real-time depth images into point cloud data.

[0068] The algorithm improvement module is used to set the dynamic neighborhood radius in the DBSCAN clustering algorithm and add curvature similarity constraints to the judgment of core points in the neighborhood to obtain an improved clustering algorithm.

[0069] The overlapping part identification module is used to segment the point cloud data using an improved clustering algorithm, calculate the gradient change trajectory from the highest point to the lowest point, perform overlapping part detection based on the gradient change trajectory, and obtain overlapping part detection results.

[0070] Each module in the aforementioned system for detecting overlapping items during automatic express delivery sorting can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a computer device memory in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0071] The present invention also provides a computer device comprising a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement the steps of an embodiment of a method for detecting overlapping items in automatic express delivery sorting. The specific implementation method can be found in the method embodiment and will not be further described here.

[0072] Furthermore, the present invention also provides a non-transitory computer-readable storage medium containing instructions, wherein a computer program is stored on the storage medium. For example, a memory containing instructions, the instructions can be executed by a processor of a computer device to complete the above method. For example, the non-transitory computer-readable storage medium can be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device. When the computer program is executed by the processor, it can implement the steps in an embodiment of a method for detecting overlapping parts in automatic express sorting. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0073] Those skilled in the art will appreciate that embodiments of the present invention may provide methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0074] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0075] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0076] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1A step that specifies a function in one or more boxes.

[0077] It should be pointed out that the specific implementation methods described above can enable those skilled in the art to understand the invention more comprehensively, but do not limit the invention in any way. Therefore, although the present specification and examples have described the invention in detail, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the scope of protection of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved. Any simple change or equivalent replacement of the technical solution that can be obviously obtained by any person familiar with the art within the technical scope disclosed in the present invention falls within the scope of protection of the present invention.

Claims

1. A method for detecting overlapping items in automatic express sorting, characterized in that: The following steps are involved: Acquire a real-time depth image of the express item on the conveyor belt; convert the real-time depth image into point cloud data; In the DBSCAN clustering algorithm, a dynamic neighborhood radius is set, and a curvature similarity constraint is added to the judgment of the core points in the neighborhood to obtain an improved clustering algorithm; The point cloud data is segmented using the improved clustering algorithm, a gradient change trajectory from the highest point to the lowest point is calculated, and overlapping detection is performed based on the gradient change trajectory to obtain overlapping detection results.

2. A method for detecting overlapping items in automatic express delivery sorting according to claim 1, characterized in that: The processing steps of the improved clustering algorithm specifically include: S1. Randomly select an unvisited point in the point cloud data and check the number of points within a neighborhood radius ε of the point; wherein the neighborhood radius is a dynamic neighborhood radius; S2. If the number of neighborhood points is ≥ MinPts, mark it as a core point and generate a new cluster; otherwise, mark it as noise; where MinPts is the minimum number of points, indicating the minimum number of points that must be included in the ε neighborhood; S3. Recursively traverse all neighboring points within the neighborhood radius of the current core point, and add all core points and boundary points within the radius to the same cluster; by checking the curvature difference, determine whether two core points in the neighborhood belong to the same cluster. If the curvature difference is less than the threshold, they belong to the same cluster, otherwise they belong to different clusters; S4, repeat steps S1 to S3 until all points are visited; S5. If the clustering result is 2 clusters or more, it is judged as a stacked part. If it is 1 cluster, further judgment is made by projecting the segmented point cloud into a 2.5D height map, obtaining the coordinates (x1, y1) of the maximum height point A and the coordinates (x2, y2) of the minimum height point B. Calculate the Sobel operator gradient of each point on any path from A to B. If there is a point with a gradient value greater than the gradient threshold, it is judged as a stacked part, otherwise it is a single part.

3. The method for detecting overlapping items in automatic express delivery sorting according to claim 1, characterized in that: The method further includes preprocessing the point cloud data before segmenting the point cloud data using the improved clustering algorithm, specifically including the following steps: Filtering the point cloud data using a voxel grid to perform downsampling; Set a height threshold to remove the background on the conveyor belt surface; By calculating the neighborhood distance distribution of each point, points that deviate from the mean by more than a threshold are removed to eliminate noise points in the point cloud.

4. The method for detecting overlapped items in automatic express delivery sorting according to claim 2, characterized in that: The dynamic neighborhood radius is based on the local height variance σ z Adjustments, specifically: e=k·s z ; in, is the mean of the z coordinates in the neighborhood, i is the serial number of the point within the neighborhood radius, N is the number of neighborhood points, and k is the empirical coefficient.

5. The method for detecting overlapped items in automatic express delivery sorting according to claim 1, characterized in that: The processing process of the curvature similarity constraint is specifically as follows: Use principal component analysis (PCA) to calculate the minimum eigenvalue λ of the point neighborhood covariance matrix min ; The curvature C is calculated based on the minimum eigenvalue, specifically using the following formula: When judging whether two points belong to the same cluster, |C p -C q |<δ C ; Among them, δ C is the curvature difference threshold.

6. The method for detecting overlapped items in automatic express delivery sorting according to claim 1, characterized in that: The real-time depth image of the express items on the conveyor belt is obtained specifically by installing a depth camera above the conveyor belt, adjusting the field of view to cover the entire width of the conveyor belt, and obtaining the depth image of the express items in real time.

7. A system for detecting overlapping items in automatic express delivery sorting, characterized in that: include: Data acquisition module, used to obtain real-time depth images of express items on the conveyor belt; Convert real-time depth images into point cloud data; The algorithm improvement module is used to set a dynamic neighborhood radius in the DBSCAN clustering algorithm and add a curvature similarity constraint to the judgment of core points in the neighborhood, thereby obtaining an improved clustering algorithm; The overlapping part identification module is used to segment the point cloud data using the improved clustering algorithm, calculate the gradient change trajectory from the highest point to the lowest point, perform overlapping part detection based on the gradient change trajectory, and obtain overlapping part detection results.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory, wherein: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is loaded into a processor, it can execute the steps of the method according to any one of claims 1 to 6.

Images