Parcel sorting anomaly detection method, computer equipment and readable storage medium

Abstract

The invention relates to a parcel sorting anomaly detection method, computer equipment, a computer readable storage medium and a computer program product, and can be applied to the technical field of logistics. The method comprises the steps that target detection processing is conducted on video data of package sorting based on a preset target detection method, and position information of packages, position information of a balance wheel device and position information of a sliding groove are obtained; based on a preset target tracking method, performing trajectory tracking processing on the position information of the package to obtain motion trajectory information of the package; according to the movement track information of the parcel, the position information of the balance wheel device and the position information of the sliding groove, at least one falling parcel moving to the sliding groove from the balance wheel device is determined; and according to the number of the falling parcels and a balance wheel working signal, the abnormal sorting detection result of the falling parcels is determined, and the balance wheel working signal is a signal received by the sorting system and used for controlling a balance wheel to work. By adopting the method, the parcel sorting anomaly detection efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of logistics technology, and in particular to a method for detecting parcel sorting anomalies, a computer device, a computer-readable storage medium, and a computer program product. Background Technology

[0002] In modern logistics sorting systems, swing wheel sorters are widely used in express delivery, e-commerce warehousing, and other scenarios due to their high efficiency and flexibility. Their core working principle is to guide packages to the target slot (sorting destination) using a rotating swing wheel. However, due to mechanical failures, identification errors, or system scheduling problems, packages may be incorrectly sorted to non-target slots, leading to chaos in subsequent logistics processes and increased transportation costs. Therefore, how to efficiently detect anomalies during the sorting process has become an important research direction.

[0003] Traditional technology typically involves manual inspection for package sorting anomaly detection, using manual inspection stations at the end of the sorting process to identify missorted packages. However, manual inspection is inefficient, resulting in low efficiency in package sorting anomaly detection. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for detecting package sorting anomalies that can improve the efficiency of package sorting anomaly detection, in order to address the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for detecting parcel sorting anomalies. The method is applied to a sorting system, which includes a swing wheel device and a chute connected to the swing wheel device. The parcel moves to the chute under the control of the swing wheel device to complete the unloading process.

[0006] The method includes:

[0007] Based on a preset target detection method, the video data of package sorting is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel.

[0008] Based on a preset target tracking method, the location information of the package is processed to obtain the movement trajectory information of the package;

[0009] Based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the chute, at least one drop package that has moved from the balance wheel device to the chute is determined.

[0010] The sorting anomaly detection result of the dropped packages is determined based on the number of dropped packages and the working signal of the balance wheel, wherein the working signal of the balance wheel is the signal received by the sorting system to control the balance wheel to work.

[0011] In one embodiment, determining the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the balance wheel operating signal includes:

[0012] When the number of the unloaded packages is 1, the sorting anomaly detection result of the unloaded packages is determined based on the detection result of the preset balance wheel working signal.

[0013] When there are multiple unloaded packages, the sorting anomaly detection results of the multiple unloaded packages are determined based on the overlapping information of the positional areas between the multiple unloaded packages.

[0014] The location region overlap information includes the intersection-union ratio information among the multiple dropped packages, which is calculated based on the boundary box information of the multiple dropped packages in the balance wheel region.

[0015] In one embodiment, determining the sorting anomaly detection result of the dropped package based on the detection result of the preset balance wheel working signal includes:

[0016] If the balance wheel is detected to be working within a preset time window, indicating that the balance wheel is to drop the package, the sorting anomaly detection result of the dropped package is determined to be normal.

[0017] If no working signal of the balance wheel is detected within the preset time window to instruct the balance wheel to drop the package, the sorting anomaly detection result of the dropped package is determined to be a sorting anomaly.

[0018] In one embodiment, determining the sorting anomaly detection result of the multiple dropped packages based on the overlapping location areas among the multiple dropped packages includes:

[0019] If the crossover ratio is greater than a preset threshold, the sorting anomaly detection results of multiple unloaded packages are determined to be sorting anomalies.

[0020] If the crossover ratio information is less than or equal to the preset threshold, the sorting anomaly detection results of multiple dropped packages are determined to be normal.

[0021] In one embodiment, determining at least one package that has moved from the balance wheel device to the chute based on the package's trajectory information, the position information of the balance wheel device, and the position information of the chute includes:

[0022] Based on a preset time window, the movement trajectory information of the package is analyzed to obtain the trajectory analysis information of the package.

[0023] Based on the trajectory analysis information, the position information of the balance wheel device, and the position information of the chute, the at least one drop package that moves from the balance wheel device to the chute is determined.

[0024] In one embodiment, the method based on a preset target detection method performs target detection processing on the video data of package sorting to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute, including:

[0025] Based on a preset time interval, image extraction processing is performed on the video data to obtain image data of package sorting;

[0026] The image data is preprocessed to obtain the target image data;

[0027] Based on the target detection method, the target image data is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute.

[0028] In one embodiment, after determining the sorting anomaly detection result of the dropped packages based on the number of the dropped packages and the balance wheel operating signal, the method further includes:

[0029] If the sorting anomaly detection result is a sorting anomaly, a sorting anomaly warning message for the dropped package is generated based on the sorting anomaly detection result.

[0030] The sorting anomaly warning information is sent to a preset monitoring terminal.

[0031] Secondly, this application also provides a parcel sorting anomaly detection device. The device includes:

[0032] The target detection module is used to perform target detection processing on the video data of package sorting based on a preset target detection method to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel;

[0033] The trajectory tracking module is used to perform trajectory tracking processing on the position information of the package based on a preset target tracking method, so as to obtain the movement trajectory information of the package;

[0034] The package placement determination module is used to determine at least one package that has moved from the balance wheel device to the slide based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the slide.

[0035] The result determination module is used to determine the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the balance wheel working signal, wherein the balance wheel working signal is the signal received by the sorting system to control the balance wheel to work.

[0036] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0037] Based on a preset target detection method, the video data of package sorting is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel.

[0038] Based on a preset target tracking method, the location information of the package is processed to obtain the movement trajectory information of the package;

[0039] Based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the chute, at least one drop package that has moved from the balance wheel device to the chute is determined.

[0040] The sorting anomaly detection result of the dropped packages is determined based on the number of dropped packages and the working signal of the balance wheel, wherein the working signal of the balance wheel is the signal received by the sorting system to control the balance wheel to work.

[0041] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0042] Based on a preset target detection method, the video data of package sorting is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel.

[0043] Based on a preset target tracking method, the location information of the package is processed to obtain the movement trajectory information of the package;

[0044] Based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the chute, at least one drop package that has moved from the balance wheel device to the chute is determined.

[0045] The sorting anomaly detection result of the dropped packages is determined based on the number of dropped packages and the working signal of the balance wheel, wherein the working signal of the balance wheel is the signal received by the sorting system to control the balance wheel to work.

[0046] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0047] Based on a preset target detection method, the video data of package sorting is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel.

[0048] Based on a preset target tracking method, the location information of the package is processed to obtain the movement trajectory information of the package;

[0049] Based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the chute, at least one drop package that has moved from the balance wheel device to the chute is determined.

[0050] The sorting anomaly detection result of the dropped packages is determined based on the number of dropped packages and the working signal of the balance wheel, wherein the working signal of the balance wheel is the signal received by the sorting system to control the balance wheel to work.

[0051] The aforementioned parcel sorting anomaly detection method, apparatus, computer equipment, computer-readable storage medium, and computer program product, based on a preset target detection method, perform target detection processing on video data of parcel sorting to obtain the parcel's position information, the position information of the swing wheel device, and the position information of the chute; based on a preset target tracking method, perform trajectory tracking processing on the parcel's position information to obtain the parcel's motion trajectory information; based on the parcel's motion trajectory information, the position information of the swing wheel device, and the position information of the chute, determine at least one dropped parcel that has moved from the swing wheel device to the chute; and based on the number of dropped parcels and the swing wheel working signal, determine the sorting anomaly detection result of the dropped parcels, wherein the swing wheel working signal is the signal received by the sorting system to control the swing wheel to operate. This solution performs target detection processing on video data of package sorting based on a preset target detection method to obtain the location information of the packages, the position information of the swing wheel device, and the position information of the chute. This facilitates the accurate positioning of each target object in the package sorting area, avoiding omissions and misjudgments due to manual observation. It also performs trajectory tracking processing on the package location information based on a preset target tracking method to obtain the package's movement trajectory information, which is beneficial for continuously tracking the package's movement process. By determining at least one package that has moved from the swing wheel device to the chute based on the package's movement trajectory information, the position information of the swing wheel device, and the position information of the chute, it helps to accurately screen packages that have moved from the swing wheel device to the chute, reducing interference from irrelevant packages. Finally, by determining the sorting anomaly detection results of the sorted packages based on the number of sorted packages and the swing wheel's operating signal, it facilitates the automated judgment of whether there are anomalies in package sorting, thereby improving the efficiency of package sorting anomaly detection and reducing the workload and time cost of manual inspection. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart illustrating a parcel sorting anomaly detection method in one embodiment;

[0054] Figure 2 This is a flowchart illustrating the steps for determining the sorting anomaly detection result in one embodiment;

[0055] Figure 3 This is a flowchart illustrating a parcel sorting anomaly detection method in another embodiment;

[0056] Figure 4This is a structural block diagram of a package sorting anomaly detection device in one embodiment;

[0057] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0059] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0060] In one exemplary embodiment, such as Figure 1 As shown, a method for detecting parcel sorting anomalies is provided.

[0061] This method is applied to an automated logistics sorting system, which includes a swing wheel device and a chute connected to the swing wheel device. In normal operation, packages arrive at the swing wheel device via a conveyor belt. The sorting system sends a swing wheel operation signal to instruct the swing wheel to rotate to the target angle. The swing wheel device then guides the package to the corresponding chute, completing the sorting process.

[0062] Normal sorting scenario: When a single item is dropped, the balance wheel working signal is detected within the preset time window to instruct the balance wheel to drop the item; when multiple items are dropped, the overlap of multiple packages on the balance wheel device is low (the crossover ratio is less than or equal to the preset threshold), indicating that the packages pass through independently and sequentially.

[0063] Misclassification scenarios: Single item misclassification: The package moves to the chute but no balance wheel working signal is detected, indicating that the package entered the chute when the balance wheel was not working properly; Multiple item misclassification: The overlap of multiple packages on the balance wheel device is high (the crossover ratio is greater than the preset threshold), indicating that the packages are stacked or stuck together and are guided to the same chute together, resulting in misclassification.

[0064] This embodiment illustrates the application of this method to a sorting system. It is understood that this method can also be applied to terminals or servers, and can be applied to systems including both terminals and servers, implemented through interaction between the terminals and servers. The terminals can be, but are not limited to, various personal computers, laptops, smartphones, tablets, etc.; the servers can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing cloud computing services.

[0065] In this embodiment, the method is applied to a sorting system, which includes a swing wheel device and a chute connected to the swing wheel device. The package moves to the chute under the control of the swing wheel device to complete the unloading process. The method includes the following steps:

[0066] Step S101: Based on a preset target detection method, target detection processing is performed on the video data of package sorting to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel.

[0067] Step S102: Based on a preset target tracking method, the location information of the package is processed by trajectory tracking to obtain the movement trajectory information of the package;

[0068] Step S103: Based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the chute, determine at least one package that has moved from the balance wheel device to the chute.

[0069] Step S104: Determine the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the working signal of the balance wheel, wherein the working signal of the balance wheel is the signal received by the sorting system to control the balance wheel to work.

[0070] The sorting system can be an automated logistics sorting system, comprising a swing wheel device and a chute connected to the swing wheel device. The swing wheel device can be a rotatable swing wheel or pulley, used to guide packages to the target compartment; the chute can be the sorting destination channel into which the package enters after moving from the swing wheel device. The preset target detection method can be a detection method based on a target detection algorithm, or a method or target detection model based on a detection model, used to identify and locate target objects such as express packages, the swing wheel device, and the chute from video images. The video data for package sorting can be real-time video data collected by a camera device set above the swing wheel, positioned vertically downwards from the swing wheel. Target detection processing can be a process of inputting pre-processed monitoring video images into a trained target detection network for feature extraction. The location information of the package can be the coordinate information of the express package in the image obtained after feature extraction by the target detection network. The location information of the swing wheel device can be the coordinate information of the swing wheel device in the image obtained after feature extraction by the target detection network. The chute's position information can be the coordinates of the chute in the image obtained after feature extraction via a target detection network. The preset target tracking method can be a multi-target tracking algorithm-based method. Trajectory tracking processing can be the process of processing the acquired package position information using a multi-target tracking algorithm to obtain the package's movement path. The package's trajectory information can be the movement path information of the express package over a period of time, obtained through a multi-target tracking algorithm; this period can be the trajectory information of the express package within 3 seconds. At least one dropped package can be identified by analyzing the express package trajectory information within 3 seconds; if a package moves from the swing wheel device to the chute, it is considered a dropped package. The number of dropped packages can be the number of dropped packages that move from the swing wheel device to the chute within 3 seconds; this number can be one or more dropped packages. The swing wheel working signal can be the signal received by the sorting system that controls the swing wheel to operate; this signal indicates whether the swing wheel device swings according to normal instructions and guides the package to the target compartment; it can also be called the swing wheel dropping signal. The Intersection over Union (IOU) ratio can be an indicator used to calculate the degree of overlap between multiple packages on the swing wheel device. The sorting anomaly detection result can be the final sorting status of the dropped packages determined by the number of dropped packages and the working signal of the balance wheel, including normal dropping, single missorting, or multiple missorting. When single missorting or multiple missorting occurs, the relevant staff will be notified to handle the missorted express packages.

[0071] Optionally, the sorting system performs target detection processing on the video data of package sorting based on a preset target detection method to obtain the position information of the packages, the position information of the swing wheel device, and the position information of the chute. Specifically, firstly, real-time video data with a specification of 25 FPS (frames per second) and 1920×1080 is acquired through a camera device set above the swing wheel. This camera device is positioned vertically downwards with the swing wheel as the center. Then, the video data of package sorting is preprocessed by extracting images of size 640×640 at 100-millisecond intervals. The extracted images are then subjected to preprocessing operations such as standardization, image compensation, and image format conversion. To improve the effectiveness of subsequent processing, the preprocessed surveillance video images are input into a trained target detection network (such as YOLOv11) for feature extraction to obtain the position information of the express package, the balance wheel device, and the chute. Based on a preset target tracking method, the position information of the package is processed for trajectory tracking to obtain the motion trajectory information of the package. Specifically, the obtained position information of the package is used to obtain the trajectory information of the express package through a multi-target tracking algorithm. The multi-target tracking algorithm predicts the target position through Kalman filtering, and can resume tracking even if the express package is briefly obscured. Based on the motion trajectory information of the package, the position information of the balance wheel device, and the position information of the chute, the trajectory information of the express package is determined. The system identifies at least one dropped package that has moved from the balance wheel device to the chute. Specifically, it analyzes the package trajectory information within 3 seconds. If a package moves from the balance wheel device to the chute, it is considered a dropped package. The system determines the sorting anomaly detection result based on the number of dropped packages and the balance wheel's operating signal. The balance wheel's operating signal is the signal received by the sorting system that controls the balance wheel to operate. Specifically, when a dropped package exists, it checks if there are other dropped packages within 3 seconds. If there is only one dropped package, it checks if there is a balance wheel operating signal within 3 seconds. If there is a balance wheel operating signal within 3 seconds (i.e., the balance wheel operating signal instructs the balance wheel to drop the package), then the dropped package is identified. If the sorting anomaly detection result is normal, the package is considered to have been missorted. Otherwise, the package is considered to have been missorted as a single item. If there are multiple packages, the Intersection over Union (IOU) is calculated based on the movement trajectory information of the multiple packages and the boundary box information of the multiple packages on the balance wheel device. A preset threshold T is set. If the IOU is greater than the preset threshold T, the package is considered to have been missorted as multiple items. Otherwise, the package is considered to have been missorted as normal. When a single or multiple missorted items occur, the relevant staff are notified to handle the missorted packages to avoid increasing transportation costs.

[0072] In the above-mentioned parcel sorting anomaly detection method, based on a preset target detection method, target detection processing is performed on the video data of parcel sorting to obtain the parcel's position information, the position information of the swing wheel device, and the position information of the chute; based on a preset target tracking method, trajectory tracking processing is performed on the parcel's position information to obtain the parcel's motion trajectory information; based on the parcel's motion trajectory information, the position information of the swing wheel device, and the position information of the chute, at least one dropped parcel that has moved from the swing wheel device to the chute is determined; based on the number of dropped parcels and the swing wheel working signal, the sorting anomaly detection result of the dropped parcels is determined, wherein the swing wheel working signal is the signal received by the sorting system to control the swing wheel to work. This solution performs target detection processing on video data of package sorting based on a preset target detection method to obtain the location information of the packages, the position information of the swing wheel device, and the position information of the chute. This facilitates the accurate positioning of each target object in the package sorting area, avoiding omissions and misjudgments due to manual observation. It also performs trajectory tracking processing on the package location information based on a preset target tracking method to obtain the package's movement trajectory information, which is beneficial for continuously tracking the package's movement process. By determining at least one package that has moved from the swing wheel device to the chute based on the package's movement trajectory information, the position information of the swing wheel device, and the position information of the chute, it helps to accurately screen packages that have moved from the swing wheel device to the chute, reducing interference from irrelevant packages. Finally, by determining the sorting anomaly detection results of the sorted packages based on the number of sorted packages and the swing wheel's operating signal, it facilitates the automated judgment of whether there are anomalies in package sorting, thereby improving the efficiency of package sorting anomaly detection and reducing the workload and time cost of manual inspection.

[0073] In one exemplary embodiment, reference is made to Figure 2 In step S104 above, the sorting anomaly detection result of the dropped packages is determined based on the number of dropped packages and the working signal of the balance wheel, specifically including the following:

[0074] Step S201: When there is only one unloaded package, determine the sorting anomaly detection result of the unloaded package based on the detection result of the preset balance wheel working signal.

[0075] Step S202: When there are multiple undelivered packages, determine the sorting anomaly detection results of multiple undelivered packages based on the overlapping information of the location areas between the multiple undelivered packages.

[0076] Among them, the location area overlap information includes the cross-union ratio information between multiple dropped packages, which is calculated based on the boundary box information of multiple dropped packages in the balance wheel area.

[0077] The number of dropped packages is 1, which can refer to the case where only 1 package is dropped from the balance wheel device to the slide within 3 seconds.

[0078] Among them, the preset balance wheel working signal can refer to the balance wheel dropping signal, that is, the signal received by the sorting system to control the balance wheel to work, which is used to indicate whether the balance wheel device swings according to normal instructions and guides the package to the target compartment.

[0079] The detection result of the preset balance wheel working signal can refer to the detection result of whether there is a balance wheel working signal within 3 seconds to instruct the balance wheel to drop the object.

[0080] The phrase "multiple packages" can refer to the situation where multiple packages move from the balance wheel device to the chute within 3 seconds.

[0081] Among them, the location area overlap information between multiple parcels can refer to the degree of overlap between multiple parcels on the balance wheel device, which includes crossover ratio information.

[0082] The intersection-union ratio (IOU) information can refer to the IOU value calculated based on the boundary box information of multiple parcels in the balance wheel area, which is used to quantify the degree of overlap between multiple parcels on the balance wheel device.

[0083] Among them, the sorting anomaly detection result of multiple undelivered packages can refer to the sorting status result determined by comparing the crossover ratio information between multiple express packages with a preset threshold.

[0084] Optionally, the sorting system determines the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the working signal of the balance wheel. Specifically, when a dropped package exists, it determines whether other dropped packages exist within 3 seconds. If there is only one dropped package, the system determines the sorting anomaly detection result based on the detection result of the preset balance wheel working signal. Specifically, if there is only one dropped package, it determines whether the balance wheel working signal instructs the balance wheel to drop the package within 3 seconds. If the balance wheel working signal instructs the balance wheel to drop the package within 3 seconds, the sorting anomaly detection result of the dropped package is determined to be a normal dropped package; otherwise, the sorting anomaly detection result of the dropped package is determined to be a single mis-sorted package, and the relevant staff are notified to handle the single mis-sorted express package. If there are multiple dropped packages, the system determines the sorting anomaly detection result based on the position area between the multiple dropped packages. The system uses overlapping information to determine the sorting anomaly detection results of multiple dropped packages. The overlapping information includes the intersection-overlap ratio (IOU) information between the multiple dropped packages. The IOU information is calculated based on the boundary box information of the multiple dropped packages in the balance wheel area. Specifically, when there are multiple dropped packages, the boundary box information of the multiple dropped packages on the balance wheel device is obtained through the motion trajectory information of the multiple express packages. The boundary box information includes the coordinate information of the multiple express packages on the balance wheel device. Then, the IOU information between the multiple express packages on the balance wheel device is calculated based on the boundary box information, and a preset threshold T is set. If the IOU information is greater than the preset threshold T, the sorting anomaly detection result of the multiple dropped packages is determined to be multiple mis-sorted packages, and the relevant staff are notified to handle the multiple mis-sorted express packages. Otherwise, the sorting anomaly detection result of the multiple dropped packages is determined to be normal dropped packages.

[0085] The technical solution provided in this embodiment determines the sorting anomaly detection result of a single package based on the detection result of a preset balance wheel working signal, which is beneficial for determining whether a single package was sorted under normal balance wheel working conditions. When there are multiple packages, the sorting anomaly detection result of multiple packages is determined based on the overlap information of the positional areas between the multiple packages (including the intersection-union ratio information calculated based on the bounding box information), which is beneficial for quantitatively determining whether multiple packages are missorted due to stacking or overlapping. This allows for the use of different detection logics for different numbers of packages, improving the accuracy and specificity of sorting anomaly detection.

[0086] In an exemplary embodiment, the above steps, determining the sorting anomaly detection result of the dropped package based on the detection result of the preset balance wheel working signal, specifically includes the following:

[0087] If the balance wheel is detected to be working within the preset time window, indicating that the balance wheel should drop the package, the sorting anomaly detection result of the dropped package is determined to be normal.

[0088] If no balance wheel working signal is detected within the preset time window to instruct the balance wheel to drop the package, the sorting anomaly detection result of the dropped package is determined to be a sorting anomaly.

[0089] The preset time window can be a 3-second time range, used to determine whether there is a balance wheel working signal within this time range to instruct the balance wheel to drop the package, so as to determine the sorting anomaly detection result of the dropped package.

[0090] The detection of a balance wheel working signal within a preset time window to instruct the balance wheel to drop the item can mean that a balance wheel working signal is present within 3 seconds and the signal instructs the balance wheel to drop the item, indicating that the balance wheel device swings according to normal instructions and guides the package to the target compartment.

[0091] The result of the sorting anomaly detection for the dropped package being "sorting normally" means that the sorting anomaly detection result for the dropped package is "normal," indicating that the single package was sorted under normal working conditions of the balance wheel, and there is no case of missorting of a single package.

[0092] The statement that no balance wheel working signal was detected within the preset time window to instruct the balance wheel to drop the item can mean that there is no balance wheel working signal or the balance wheel working signal does not instruct the balance wheel to drop the item within 3 seconds, indicating that the balance wheel device does not swing and guide the package to the target compartment according to the normal instructions.

[0093] Among them, the sorting anomaly detection result of the dropped package is sorting anomaly, which can mean that the sorting anomaly detection result of the dropped package is a single item missorted, indicating that the single package entered the chute when the balance wheel did not swing according to the normal instructions.

[0094] Optionally, the sorting system determines the sorting anomaly detection result of the dropped package based on the detection result of the preset balance wheel working signal. Specifically, when the number of dropped packages is 1, the balance wheel working signal is detected within a preset time window, which can be a 3-second time range. The balance wheel working signal is the signal received by the sorting system that controls the balance wheel to work (also known as the balance wheel dropping signal). This signal is used to instruct the balance wheel device to swing according to normal instructions and guide the package to the target compartment. If the balance wheel working signal is detected within the preset time window, instructing the balance wheel to drop the package, the abnormality of the dropped package is determined. If the sorting anomaly detection result is "sorting normal," it means the package was sorted normally, indicating that the package was sorted under normal operating conditions and there was no missorting. If no operating signal is detected from the balance wheel within the preset time window, the package is determined to be missorted, indicating that the package entered the chute without the balance wheel swinging as instructed, resulting in missorting. The relevant staff will then be notified to handle the missorted package.

[0095] The technical solution provided in this embodiment detects whether the balance wheel working signal instructs the balance wheel to drop the package within a preset time window, and determines the sorting anomaly detection result of the dropped package accordingly. This is beneficial for judging whether a single package is sorted normally by combining the actual working signal of the sorting system, effectively identifying missorting caused by the balance wheel device not working properly, thereby improving the accuracy of single package sorting anomaly detection.

[0096] In an exemplary embodiment, the steps described above, determining the sorting anomaly detection results of multiple dropped packages based on the overlapping location areas between them, specifically include the following:

[0097] If the crossover ratio is greater than a preset threshold, the sorting anomaly detection results of multiple dropped packages are determined as sorting anomalies;

[0098] If the crossover ratio is less than or equal to a preset threshold, the sorting anomaly detection result of multiple dropped packages is determined to be normal.

[0099] The intersection-union ratio (IOU) information can refer to the IOU value calculated based on the boundary box information of multiple parcels in the balance wheel area, which is used to quantify the degree of overlap between multiple parcels on the balance wheel device.

[0100] The preset threshold can refer to threshold T, which is used to compare with the cross-combination ratio information to determine the sorting anomaly detection results of multiple dropped packages.

[0101] Among them, the sorting anomaly detection result of multiple dropped packages is sorting anomaly, which means that the sorting anomaly detection result of multiple dropped packages is determined to be multiple missorted, indicating that multiple express packages are stacked or overlapped on the swing wheel device.

[0102] Among them, the sorting anomaly detection result of multiple unloaded packages is normal sorting, which means that the sorting anomaly detection result of multiple unloaded packages is normal unloading, indicating that multiple express packages are not stacked or overlapped on the swing wheel device, and it is normal multi-item sorting.

[0103] Optionally, the sorting system determines the sorting anomaly detection result of multiple dropped packages based on the overlapping information of their location areas. Specifically, when there are multiple dropped packages, the system calculates the intersection-union ratio (IOU) information based on the boundary box information of the multiple dropped packages in the balance wheel area. The boundary box information includes the coordinate information of the multiple express packages on the balance wheel device. The IOU information between the multiple express packages on the balance wheel device is calculated using the boundary box information of the multiple express packages on the balance wheel device. The IOU is used to represent the degree of overlap between the multiple express packages on the balance wheel device. If the IOU information is greater than a preset threshold, multiple dropped packages are determined to be... If the parcel sorting anomaly detection result is "sorting anomaly," it means that multiple parcels have been missorted, indicating that the overlap between the parcels on the balance wheel device is high, and there is a situation where stacking or overlapping has led to missorting. The relevant staff will be notified to handle the multiple missorted parcels. If the crossover ratio is less than or equal to the preset threshold, the parcel sorting anomaly detection result is "sorting normal," meaning that the parcel sorting anomaly detection result is "normal parcel sorting," indicating that the overlap between the parcels on the balance wheel device is low, and there is no missorting due to stacking or overlapping. It is a normal situation of multiple parcel sorting.

[0104] The technical solution provided in this embodiment calculates the cross-union ratio information based on the boundary box information of multiple dropped packages in the balance wheel area, and compares the cross-union ratio information with a preset threshold to determine the sorting anomaly detection result of multiple dropped packages. This is beneficial for quantitatively judging the degree of overlap between multiple packages, accurately identifying multiple missorting situations caused by package stacking or overlap, and thus improving the accuracy of multi-package sorting anomaly detection.

[0105] In an exemplary embodiment, step S103 above, determining at least one package that has moved from the balance wheel device to the chute based on the package's movement trajectory information, the position information of the balance wheel device, and the position information of the chute, specifically includes the following:

[0106] Based on a preset time window, the movement trajectory information of the package is analyzed to obtain the trajectory analysis information of the package.

[0107] Based on trajectory analysis information, the position information of the balance wheel device, and the position information of the chute, at least one package that has moved from the balance wheel device to the chute is identified.

[0108] The trajectory analysis and processing can be a process of analyzing and judging the movement trajectory information of the package within a preset time window, used to determine whether the package has moved from the balance wheel device to the slide.

[0109] Among them, the trajectory analysis information of the package can be the result information obtained after analyzing and judging the movement trajectory information of the package within a preset time window.

[0110] Optionally, the sorting system determines at least one unloading package that has moved from the balance wheel device to the chute based on the package's movement trajectory information, the position information of the balance wheel device, and the position information of the chute. Specifically, based on a preset time window, the system performs trajectory analysis processing on the package's movement trajectory information to obtain the package's trajectory analysis information. The preset time window is a 3-second time range. Within 3 seconds, the system analyzes and judges the package's movement trajectory information to determine whether the package has moved from the balance wheel device to the chute within 3 seconds. The trajectory analysis information includes the judgment result information of whether the package has moved from the balance wheel device to the chute within 3 seconds. Based on the trajectory analysis information, the position information of the balance wheel device, and the position information of the chute, the system determines at least one unloading package that has moved from the balance wheel device to the chute. Specifically, based on the trajectory analysis information, the system judges whether the package has moved from the balance wheel device to the chute within 3 seconds. If the package has moved from the balance wheel device to the chute within 3 seconds, then the package is determined to be an unloading package.

[0111] The technical solution provided in this embodiment analyzes the trajectory information of the package based on a preset time window, and determines at least one package that has moved from the balance wheel to the slide based on the trajectory analysis information, the position information of the balance wheel device and the position information of the slide. This helps to accurately identify the package that has completed the dropping action within a limited time range, reduces interference from irrelevant packages, and thus improves the accuracy and timeliness of package dropping identification.

[0112] In an exemplary embodiment, in step S101 above, based on a preset target detection method, target detection processing is performed on the video data of package sorting to obtain the position information of the package, the position information of the swing wheel device, and the position information of the chute, specifically including the following:

[0113] Based on a preset time interval, image extraction processing is performed on the video data to obtain image data of package sorting;

[0114] Image preprocessing is performed on the image data to obtain the target image data;

[0115] Based on the target detection method, the target image data is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute.

[0116] The preset time interval can refer to a time interval of 100 milliseconds.

[0117] Image extraction processing can refer to the process of extracting images from video data at preset time intervals.

[0118] Among them, the image data for package sorting can refer to image data extracted from real-time video data at preset time intervals of 100 milliseconds.

[0119] Image preprocessing can refer to preprocessing operations such as standardization, image compensation, and image format conversion of image data.

[0120] The target image data can refer to the image data obtained after image preprocessing.

[0121] Optionally, the sorting system performs target detection processing on the video data of package sorting based on a preset target detection method to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute. Specifically, based on a preset time interval of 100 milliseconds, image extraction processing is performed on the video data to obtain image data of package sorting. Real-time video data with specifications of 25FPS and 1920×1080 is acquired by a camera device set above the balance wheel. The camera device is set at a position that shines vertically downward with the balance wheel as the center. To reduce the hardware burden, images are extracted from the video data at a time interval of 100 milliseconds to obtain package sorting image data with a size of 640×640. Image preprocessing is performed on the image data to obtain target image data. Image preprocessing includes preprocessing operations such as standardization, image compensation, and image format conversion to improve the effect of subsequent processing. Based on the target detection method, target detection processing is performed on the target image data to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute.

[0122] The technical solution provided in this embodiment performs image extraction processing on video data based on a preset time interval, and performs target detection processing on the extracted image data after preprocessing. This helps to reduce the hardware burden and improve processing efficiency while ensuring the detection effect, thereby facilitating the accurate acquisition of the position information of the package, the swing wheel device and the chute.

[0123] In an exemplary embodiment, after determining the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the balance wheel working signal, step S104 above further includes the following:

[0124] If the sorting anomaly detection result is a sorting anomaly, a sorting anomaly warning message for the lost package will be generated based on the sorting anomaly detection result.

[0125] The sorting anomaly warning message will be sent to the preset monitoring terminal.

[0126] Among them, the sorting anomaly detection result can refer to the sorting anomaly detection result of a single missorted package or multiple missorted packages.

[0127] Among them, the sorting anomaly warning information for undelivered packages can refer to the warning information generated based on the sorting anomaly detection results.

[0128] The preset monitoring terminal can refer to a terminal device used to receive sorting abnormality warning information. The preset monitoring terminal can be a staff mobile terminal device or a display terminal device in the monitoring center.

[0129] Optionally, after determining the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the working signal of the balance wheel, the sorting system generates a sorting anomaly warning message for the dropped packages if the sorting anomaly detection result is a sorting anomaly. Specifically, when the sorting anomaly detection result is a single missorted package, a sorting anomaly warning message is generated, including the single missorted package type, the location information of the dropped package, the time information of the single missorted package, and the motion trajectory information of the dropped package from the balance wheel device to the chute. When the sorting anomaly detection result is multiple missorted packages, a sorting anomaly warning message is generated, including the multiple missorted package type, the location information of multiple dropped packages, the time information of the multiple missorted packages, and the crossover ratio information between multiple dropped packages. The sorting anomaly warning message is then sent to a preset monitoring terminal, which can be a mobile terminal device of the staff or a display terminal device of the monitoring center. After receiving the sorting anomaly warning message, the preset monitoring terminal displays the sorting anomaly warning message to the staff.

[0130] The technical solution provided in this embodiment generates a sorting anomaly warning message for the missorted package based on the sorting anomaly detection result when the sorting anomaly is detected, and sends the warning message to a preset monitoring terminal. This helps to promptly notify relevant staff to handle the missorted express package, prevent the missorted package from entering the downstream logistics process, and thus help reduce transportation costs and manual error correction costs.

[0131] The following example illustrates the parcel sorting anomaly detection method provided in this application. This example demonstrates the application of this method to a terminal.

[0132] In automated logistics sorting systems, swing wheel sorters are widely used in express delivery, e-commerce warehousing, and other scenarios due to their high efficiency and flexibility. Their core working principle is to guide packages to the target compartment (sorting destination) using a rotating swing wheel (or pulley). However, due to mechanical failures, identification errors, or system scheduling problems, packages may be incorrectly sorted to non-target compartments, leading to chaos in subsequent logistics processes and increasing the cost of manual error correction.

[0133] Currently, the main methods to reduce misclassification are as follows, but significant shortcomings still exist:

[0134] 1. Single identification technology based on barcode / RFID (Radio Frequency Identification) identifies the unique identifier of a package (such as a barcode or QR code) by installing a fixed barcode scanner or RFID reader on the sorting line to determine the target compartment. However, it has problems such as high missed / false scanning rate (especially in high-speed sorting, package stacking, or label damage), inability to detect missorting caused by mechanical failure (such as incorrect rotation of the balance wheel), and lack of redundant verification mechanism and real-time error correction capability.

[0135] 2. Weight / volume detection auxiliary technology compares the actual parameters of the package with the database records through dynamic weighing or 3D volume measurement to detect anomalies and trigger an alarm. However, it has the problem that it is difficult to detect light and small packages (such as envelopes and soft bags) because the weight difference is not significant, and it cannot distinguish packages of the same weight class but with different destinations.

[0136] 3. Visual sorting systems use cameras to capture images of packages and use OCR (Optical Character Recognition) or deep learning models to identify the label information. However, the recognition accuracy will decrease under complex lighting conditions (such as reflections and shadows), and visual technology alone cannot solve the physical missorting problem caused by the mechanical malfunction of the balance wheel.

[0137] 4. Manual verification and subsequent error correction: This involves setting up manual inspection stations at the end of sorting or using subsequent logistics links (such as package scanning) to identify missorted packages. However, there are problems such as delayed error correction (causing missorted packages to enter downstream and increase reverse logistics costs) and difficulty in scaling up labor costs during peak periods.

[0138] The main problems with existing technologies are: low reliability of single-point detection (relying on a single sensor and easily affected by environmental interference), lack of real-time error correction capability (unable to dynamically adjust misaligned paths) and lack of awareness of mechanical faults (difficult to detect hardware problems such as incorrect balance wheel steering or jamming in a timely manner).

[0139] To address the above problems, this embodiment solves the deficiencies of the prior art through the following innovations:

[0140] 1. A visual data acquisition method for balance wheel misalignment with the balance wheel as the center and viewed from above; the position of the balance wheel and the slide can be accurately regressed through detection algorithms, reducing manual maintenance costs; it can also effectively handle the case of a single balance wheel with multiple slides, saving camera equipment and computing resources, and the camera has high stability, avoiding misjudgments due to equipment instability;

[0141] 2. Parcel motion status recognition based on target recognition and multi-target tracking algorithms; while offering high real-time performance, it also improves detection accuracy for small targets and dense scenes through structural optimization, making it suitable for parcel recognition in complex logistics environments; it has good adaptability to changes in lighting, occlusion, or tilted parcels, reducing missed detections and false detections. Target position is predicted using Kalman filtering, and tracking can be resumed even if the parcel is briefly occluded (e.g., due to stacking or sorting machine obstruction).

[0142] 3. Package missorting logic based on target IOU and parcel motion status combined with balance wheel oscillation signal; the missorting logic combining IOU, motion status and balance wheel signal, through multi-dimensional cross-validation, significantly improves the intelligence and reliability of the sorting system, and is especially suitable for modern logistics scenarios with high throughput and high accuracy requirements.

[0143] refer to Figure 3 The technical solution of this embodiment is as follows:

[0144] Step 1: Camera setup. Set up a camera angle that shines vertically downwards from the center of the balance wheel, and acquire real-time video data at 25 FPS and 1920×1080 resolution.

[0145] Step 2: Video data processing. The video data is preprocessed. To reduce the hardware burden, the video is extracted into 640×640 images with a time interval of 100 milliseconds. The images are then preprocessed with standardization, image compensation, and image format conversion to improve the effect of subsequent processing.

[0146] Step 3: Obtain target location information. Input the preprocessed surveillance video images into the trained target detection network to extract features and obtain the location information of the express package, balance wheel, and chute.

[0147] Step 4: Obtain the movement trajectory of the express package. The obtained package location information is used to obtain the trajectory information of the express package through a multi-target tracking algorithm.

[0148] Step 5: Parcel drop analysis. Analyze the parcel trajectory information within 3 seconds. If the parcel moves from the balance wheel area to the slide area, it is considered a dropped parcel.

[0149] Step 6: When a package is found to be missing, determine if there are any other missing packages within 3 seconds. If it is a single missing package, proceed to step 7; if it is multiple missing packages, proceed to step 8.

[0150] Step 7: When a single piece is dropped, if a drop signal is received from the balance wheel within 3 seconds, it is a normal drop; otherwise, it is a single piece misplaced.

[0151] Step 8: When multiple packages are dropped, analyze the IOU between packages on the balance wheel by the movement trajectory of multiple packages, and set a threshold T. If the IOU > T, it is a case of multiple packages being missorted; otherwise, it is a case of normal dropping.

[0152] Step 9: If a single item or multiple items are mis-packaged, report it to the relevant staff to handle the mis-packaged parcels and avoid increasing transportation costs.

[0153] The technical solution provided in this embodiment achieves the following: a visual data acquisition method for misclassification of the balance wheel centered on the balance wheel and viewed from above, which effectively saves camera and algorithm resources in scenarios with chute, and the accuracy of the balance wheel and chute position information regressed by the algorithm is high, resulting in low manual maintenance costs in the later stages; the recognition of the motion state of express parcels based on target recognition and multi-target tracking algorithms improves the real-time monitoring performance of the system; and the parcel misclassification logic based on target IOU and the motion state of the express parcel combined with the balance wheel swing signal is highly accurate and timely.

[0154] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0155] Based on the same inventive concept, this application also provides a package sorting anomaly detection device for implementing the package sorting anomaly detection method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the package sorting anomaly detection device provided below can be found in the limitations of the package sorting anomaly detection method described above, and will not be repeated here.

[0156] In one exemplary embodiment, such as Figure 4As shown, a parcel sorting anomaly detection device 400 is provided, which is applied to a sorting system. The sorting system includes a swing wheel device and a chute connected to the swing wheel device. Parcels move to the chute under the control of the swing wheel device to complete the unloading process. The parcel sorting anomaly detection device 400 may include:

[0157] The target detection module 401 is used to perform target detection processing on the video data of package sorting based on a preset target detection method to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel;

[0158] The trajectory tracking module 402 is used to perform trajectory tracking processing on the position information of the package based on a preset target tracking method to obtain the motion trajectory information of the package.

[0159] The package determination module 403 is used to determine at least one package that has moved from the balance wheel device to the slide based on the movement trajectory information of the package, the position information of the balance wheel device and the position information of the slide.

[0160] The result determination module 404 is used to determine the sorting anomaly detection result of the dropped package based on the number of dropped packages and the balance wheel working signal, wherein the balance wheel working signal is the signal received by the sorting system to control the balance wheel to work.

[0161] In an exemplary embodiment, the result determination module 404 is further configured to, when the number of the dropped packages is one, determine the sorting anomaly detection result of the dropped package based on the detection result of a preset balance wheel working signal; and when the number of the dropped packages is multiple, determine the sorting anomaly detection result of multiple dropped packages based on the positional area overlap information between the multiple dropped packages; wherein, the positional area overlap information includes the intersection-overlap ratio information between the multiple dropped packages, and the intersection-overlap ratio information is calculated based on the boundary box information of the multiple dropped packages in the balance wheel area.

[0162] In an exemplary embodiment, the result determination module 404 is further configured to determine the sorting anomaly detection result of the dropped package as sorting normal if the balance wheel working signal is detected within a preset time window to instruct the balance wheel to drop the package; and to determine the sorting anomaly detection result of the dropped package as sorting abnormal if the balance wheel working signal is not detected within the preset time window to instruct the balance wheel to drop the package.

[0163] In an exemplary embodiment, the result determination module 404 is further configured to determine the sorting anomaly detection results of the multiple dropped packages as sorting anomalies when the cross-combination ratio information is greater than a preset threshold; and to determine the sorting anomaly detection results of the multiple dropped packages as sorting normal when the cross-combination ratio information is less than or equal to the preset threshold.

[0164] In an exemplary embodiment, the package determination module 403 is further configured to perform trajectory analysis processing on the movement trajectory information of the package based on a preset time window to obtain trajectory analysis information of the package; and determine the at least one package that has moved from the balance wheel device to the chute based on the trajectory analysis information, the position information of the balance wheel device and the position information of the chute.

[0165] In an exemplary embodiment, the target detection module 401 is further configured to perform image extraction processing on the video data based on a preset time interval to obtain image data of package sorting; perform image preprocessing on the image data to obtain target image data; and perform target detection processing on the target image data based on the target detection method to obtain the position information of the package, the position information of the swing wheel device, and the position information of the chute.

[0166] In an exemplary embodiment, the device 400 further includes: an information sending module, configured to generate a sorting abnormality warning message for the dropped package based on the sorting abnormality detection result when the sorting abnormality detection result is a sorting abnormality; and send the sorting abnormality warning message to a preset monitoring terminal.

[0167] Each module in the aforementioned package sorting anomaly detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0168] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for detecting package sorting anomalies. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0169] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0170] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0171] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0172] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0173] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0174] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0175] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for detecting anomalies in package sorting, characterized in that, The method is applied to a sorting system, which includes a swing wheel device and a chute connected to the swing wheel device. The package moves to the chute under the control of the swing wheel device to complete the unloading. The method includes: Based on a preset target detection method, the video data of package sorting is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute; the video data is collected by a camera device set above the balance wheel. Based on a preset target tracking method, the location information of the package is processed to obtain the movement trajectory information of the package; Based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the chute, at least one drop package that has moved from the balance wheel device to the chute is determined. The sorting anomaly detection result of the dropped packages is determined based on the number of dropped packages and the working signal of the balance wheel, wherein the working signal of the balance wheel is the signal received by the sorting system to control the balance wheel to work.

2. The method according to claim 1, characterized in that, The step of determining the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the balance wheel working signal includes: When the number of the unloaded packages is 1, the sorting anomaly detection result of the unloaded packages is determined based on the detection result of the preset balance wheel working signal. When there are multiple unloaded packages, the sorting anomaly detection results of the multiple unloaded packages are determined based on the overlapping information of the positional areas between the multiple unloaded packages. The location region overlap information includes the intersection-union ratio information among the multiple dropped packages, which is calculated based on the boundary box information of the multiple dropped packages in the balance wheel region.

3. The method according to claim 2, characterized in that, The step of determining the sorting anomaly detection result of the dropped package based on the detection result of the preset balance wheel working signal includes: If the balance wheel is detected to be working within a preset time window, indicating that the balance wheel is to drop the package, the sorting anomaly detection result of the dropped package is determined to be normal. If no working signal of the balance wheel is detected within the preset time window to instruct the balance wheel to drop the package, the sorting anomaly detection result of the dropped package is determined to be a sorting anomaly.

4. The method according to claim 2, characterized in that, The step of determining the sorting anomaly detection results of multiple dropped packages based on the overlapping location areas between the multiple dropped packages includes: If the crossover ratio is greater than a preset threshold, the sorting anomaly detection results of multiple unloaded packages are determined to be sorting anomalies. If the crossover ratio information is less than or equal to the preset threshold, the sorting anomaly detection results of multiple dropped packages are determined to be normal.

5. The method according to claim 1, characterized in that, The step of determining at least one package to be dropped from the balance wheel device to the chute based on the movement trajectory information of the package, the position information of the balance wheel device, and the position information of the chute includes: Based on a preset time window, the movement trajectory information of the package is analyzed to obtain the trajectory analysis information of the package. Based on the trajectory analysis information, the position information of the balance wheel device, and the position information of the chute, the at least one drop package that moves from the balance wheel device to the chute is determined.

6. The method according to claim 1, characterized in that, The preset target detection method performs target detection processing on the video data of package sorting to obtain the position information of the package, the position information of the swing wheel device, and the position information of the chute, including: Based on a preset time interval, image extraction processing is performed on the video data to obtain image data of package sorting; The image data is preprocessed to obtain the target image data; Based on the target detection method, the target image data is processed to obtain the position information of the package, the position information of the balance wheel device, and the position information of the chute.

7. The method according to claim 1, characterized in that, After determining the sorting anomaly detection result of the dropped packages based on the number of dropped packages and the balance wheel working signal, the process also includes: If the sorting anomaly detection result is a sorting anomaly, a sorting anomaly warning message for the dropped package is generated based on the sorting anomaly detection result. The sorting anomaly warning information is sent to a preset monitoring terminal.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images