Automatic sorting machine luggage falling detection method and mechanism and electronic equipment

Abstract

The invention provides an automatic sorting machine luggage falling detection method, which is characterized in that a luggage detection model is obtained based on deep learning training, so that target luggage can be analyzed, whether an object is luggage or not can be identified, and the spatial position of the luggage can be positioned; and whether the luggage falls off from the sorting machine is judged based on time and space dimensions. According to the luggage falling detection method, key technologies such as deep learning, target detection and an attention mechanism are integrated, the luggage falling from the sorting machine is monitored in real time based on time and space dimensions, an alarm is given out, automatic, intelligent and informatization detection of the luggage is achieved, a traditional manual visual method is replaced, and the detection efficiency is improved. The luggage delayed delivery rate can be effectively reduced, and the operation management efficiency is improved. The invention further provides a luggage falling detection mechanism of the automatic sorting machine and electronic equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of baggage detection, and in particular to a baggage swap detection method and mechanism for an automatic sorting machine and an electronic device. Background Art

[0002] In the operation and management of the airport baggage conveying system, luggage often falls off the sorting machine due to the centrifugal motion at the bend, the vibration of the sorting machine during operation, and the sliding of luggage on the surface of the sorting machine, which may cause the luggage to be delivered late, affecting the luggage handling efficiency and passenger satisfaction. In order to ensure that the checked luggage at the airport can be delivered on time, it is necessary to detect and handle the problem of luggage substitution in real time.

[0003] In the actual application scenarios of baggage swap detection, due to the constraints of uncontrollable factors such as the external environment, the following problems may exist when performing target detection on baggage images taken in real scenes: the movement of the sorting machine, the shaking of the protective net, the reflection of light by the luggage, etc., cause the environment to change continuously, resulting in blurred images or overexposure; at the same time, since the actual detection environment is relatively complex, the captured background often contains a large amount of interference information, the luggage may be blocked by other objects in the environment, or dense luggage may block each other due to angle problems, resulting in incomplete baggage images and problems such as image quality degradation; in addition, the position and angle of the luggage swap are highly random, which may lead to large differences in the shape and size of the luggage.

[0004] Therefore, how to improve the accuracy of luggage swap detection at airports is a key issue in the current research and application of intelligent baggage detection. Summary of the invention

[0005] The present invention provides a baggage swap detection method for an automatic sorting machine based on deep learning, which solves the above-mentioned technical problems and is conducive to improving the baggage detection accuracy in complex environments. The baggage swap detection has high reliability and good practicality.

[0006] The technical solution adopted by the present invention is as follows: An embodiment of the present invention provides a method for detecting baggage substitution in an automatic sorting machine, which comprises the following steps: S101, collecting historical videos of the baggage conveying system, constructing an original baggage dataset, performing data enhancement, and obtaining an enhanced baggage dataset; S102, setting a spatial attention module after the feature extraction network and before the detection head, creating a baggage detection model in combination with the attention mechanism, and optimizing the baggage detection model using the enhanced baggage data set obtained in S101; S103, pre-processing the video stream, combining the luggage detection model optimized in S102, adopting a luggage swap detection method based on time and space dimensions, realizing luggage swap detection and issuing an alarm.

[0007] Optionally, the constructing of the original baggage dataset in S101 comprises the following steps: extracting frames from the historical video stream, cleaning out luggage images that meet format requirements, are non-repetitive, and have balanced categories, and marking the cleaned luggage images with selection boxes to obtain the original baggage dataset.

[0008] Optionally, the data enhancement means described in S101 include any one or more of the following: Data enhancement method 1: If the number of luggage images blocked by the protective net screened out in S101 is less than a preset threshold, a Mixup method is used for data enhancement, that is, the cropped protective net and luggage images are linearly mixed in a certain proportion to generate a new training sample, and their labels are linearly mixed accordingly; Data enhancement method 2: If the number of luggage images blocked by the fixed support screened out in S101 is less than a preset threshold, a Mixup method is used for data enhancement, that is, the cropped fixed support and luggage images are linearly mixed in a certain proportion to generate a new training sample, and their labels are linearly mixed accordingly; Data enhancement method three: If the number of luggage images with luggage objects blocking each other or multiple luggage objects selected in S101 is less than the preset threshold, the CutMix method is used for data enhancement, that is, the cropped luggage images are randomly pasted onto other luggage images, and the corresponding labels are adjusted to generate new training samples.

[0009] Optionally, the spatial attention module includes at least the following modules: a separation module, a fusion module and a selection module. The workflow of the spatial attention module includes: S401, separation: performing convolution operations on the extracted feature maps with multiple convolution kernels of different sizes respectively to obtain feature maps with different receptive field sizes; S402, fusion: the feature maps obtained by the separation module are fused, and the channel-based global average pooling and global maximum pooling operations are performed on the fused feature maps in turn to obtain feature maps that count feature information of each spatial position, and the channel splicing operation is performed on the two feature maps; S403, selection: perform a convolution operation on the feature map obtained by the fusion module and pass it into the Sigmoid function to obtain multiple feature maps, which are multiplied with the feature maps obtained by the separation module and fused and outputted to obtain a feature map with enhanced feature information.

[0010] Optionally, the baggage detection model is created by combining the attention mechanism in S102, and the baggage detection model at least includes a feature extraction network, a feature fusion network, and a detection head, wherein: The feature extraction network includes a first convolutional layer, a second convolutional layer, a first C2f module, a third convolutional layer, a second C2f module, a fourth convolutional layer, a third C2f module, a fifth convolutional layer, a fourth C2f module and an SPPF module, which are arranged in sequence; The feature fusion network includes a first upsampling layer, a first Concat fusion layer, a fifth C2f module, a second upsampling layer, a second Concat fusion layer, a sixth C2f module, a sixth convolutional layer, a third Concat fusion layer, a seventh C2f module, a seventh convolutional layer, a fourth Concat fusion layer, and an eighth C2f module, and spatial attention modules are added after the sixth C2f module, the seventh C2f module, and the eighth C2f module respectively; The detection head includes a first decoupling head detection layer, a second decoupling head detection layer and a third decoupling head detection layer, which are respectively arranged after the spatial attention module.

[0011] Optionally, the preprocessing of the video stream described in S103 specifically includes the following steps: S601: For the current video stream, an interactive ROI selection method is used to select a baggage swap detection area according to the camera's field of view and the area where the protective net is located; S602: Extract frames from the video stream to obtain the current image to be detected.

[0012] Optionally, the baggage substitution detection method based on time and space dimensions described in S103 specifically includes the following steps: S701, for the first image to be detected, initialize a counter; S702, using the baggage detection model optimized in S102 to infer the image to be detected, and calculating the size of the overlapping part of the baggage bounding box obtained by inference with the area to be detected, as well as the size occupied by the bounding box. If the calculated result is greater than or equal to the area proportion threshold, the baggage is deemed to be detected; S703: For the image of the checked luggage, calculate whether the checked luggage meets the time stability requirement, that is, whether the time stamp difference between two consecutive images of the checked luggage is less than or equal to the time threshold; S704: For images that meet the requirements of temporal stability, calculate whether the detected luggage meets the requirements of spatial stability, that is, whether the luggage bounding box IOUs of two consecutive images of detected luggage are greater than or equal to the IOU threshold. If so, add one to the counter. S705. For images that satisfy both temporal stability and spatial stability, determine whether the counter is greater than or equal to a counting threshold. If the counter is greater than or equal to the counting threshold, determine that luggage substitution has occurred and issue an alarm. Otherwise, re-execute S702.

[0013] The embodiment of the present invention further provides a baggage switching detection mechanism for an automatic sorting machine, which is applied to the baggage switching detection method for an automatic sorting machine to perform real-time detection and analysis of baggage images. The mechanism includes: Baggage image acquisition module, used to collect historical videos of the baggage conveying system, construct the original baggage dataset, and perform data enhancement; A model creation module, for creating a baggage detection model in combination with an attention mechanism, and optimizing the baggage detection model using an enhanced baggage dataset; The data processing module is used to process the acquired data to determine whether luggage substitution has occurred.

[0014] An embodiment of the present invention further provides an electronic device, including a processor and a memory, wherein the memory stores executable instructions that can be executed by the processor, and the processor executes the executable instructions to implement the automatic sorting machine luggage substitution detection method.

[0015] The baggage swap detection method for an automatic sorting machine provided by an embodiment of the present invention combines a data enhancement strategy to enhance the diversity of baggage data sets, thereby improving the robustness and accuracy of the model when processing complex scenarios; adopts a deep learning method to detect baggage targets, selectively highlights important spatial areas through an attention module, and improves the model's ability to detect small targets; and combines time and space stability to detect baggage swapping, which can effectively alleviate the interference caused by the complex and changeable environment of the baggage conveying system, reduce false alarms and improve the detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 It is a schematic diagram of the method flow of the present invention; Figure 2 It is a schematic diagram of the structure of the baggage detection model of the present invention; Figure 3 Schematic diagram of the spatial attention module structure of the present invention; Figure 4The figure is a flow chart of a baggage substitution detection method based on time and space dimensions of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0021] In the description of the present invention, it should be noted that if the terms "upper", "lower", "inside", "outside", etc. appear to indicate an orientation or position relationship, they are based on the orientation or position relationship shown in the accompanying drawings, or are the orientation or position relationship in which the product of the invention is usually placed when used. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0022] In addition, the terms “first”, “second”, etc., if used, are merely used to distinguish between the descriptions and should not be understood as indicating or implying relative importance.

[0023] It should be noted that, in the absence of conflict, the features in the embodiments of the present invention may be combined with each other.

[0024] Please refer to Figure 1 This embodiment provides a baggage substitution detection method for an automatic sorting machine, comprising the following steps: S101. Collect historical videos of the baggage conveying system, construct an original baggage dataset, and perform data enhancement.

[0025] This embodiment extracts frames from the historical video stream of the baggage conveying system of Huanghua Airport, and cleans out the baggage pictures that meet the format requirements, have no duplication, and have balanced categories. Then, the Labelme tool is used to select and mark the cleaned baggage pictures to obtain the original baggage data set, which contains a total of 1323 pictures. Based on the original baggage dataset, the following data enhancement methods are adopted to obtain an enhanced dataset containing a total of 4553 images: Data enhancement method 1: If the number of luggage images blocked by the protective net screened out in S101 is less than the preset number threshold, the Mixup method is used for data enhancement, that is, the cropped protective net and luggage images are linearly mixed in a certain proportion to generate a new training sample, and their labels are linearly mixed accordingly, so as to increase the diversity and generalization ability of the training data, thereby improving the robustness and accuracy of the model in processing complex scenes; Data enhancement method 2: If the number of luggage images blocked by the fixed brackets screened out in S101 is less than the preset threshold, the Mixup method is used for data enhancement, that is, the cropped fixed brackets and luggage images are linearly mixed in a certain proportion to generate a new training sample, and their labels are linearly mixed accordingly, so as to increase the diversity and generalization ability of the training data, thereby improving the robustness and accuracy of the model in processing complex scenes; Data enhancement method three: If the number of luggage images with luggage targets occluding each other or with multiple luggage targets screened out in S101 is less than the preset threshold, the CutMix method is used for data enhancement, that is, the cropped luggage images are randomly pasted onto other luggage images, and the corresponding labels are adjusted to generate new training samples. In this way, the diversity and richness of the data are increased, thereby improving the robustness of the model to occlusion and the recognition accuracy.

[0026] S102. Set a spatial attention module after the feature extraction network and before the detection head, create a baggage detection model in combination with the attention mechanism, and optimize the model using the enhanced baggage dataset obtained in S101.

[0027] The baggage detection model includes a feature extraction network, a feature fusion network, and a detection head, where: The feature extraction network includes a first convolution layer Conv1, a second convolution layer Conv2, a first C2f module C2f1, a third convolution layer Conv3, a second C2f module C2f2, a fourth convolution layer Conv4, a third C2f module C2f3, a fifth convolution layer Conv5, a fourth C2f module C2f4, and an SPPF module, which are arranged in sequence; The feature fusion network includes a first upsampling layer Upsample1, a first Concat fusion layer Concat1, a fifth C2f module C2f5, a second upsampling layer Upsample2, a second Concat fusion layer Concat2, a sixth C2f module C2f6, a sixth convolutional layer Conv6, a fourth Concat fusion layer Concat4, a seventh C2f module C2f7, a seventh convolutional layer Conv7, a fifth Concat fusion layer Concat5, and an eighth C2f module C2f8, and a spatial attention module SA is added after the sixth C2f module C2f6, the seventh C2f module C2f7, and the eighth C2f module C2f8 respectively; The detection head includes a first decoupling head detection layer Detect1, a second decoupling head detection layer Detect2 and a third decoupling head detection layer Detect3, which are respectively arranged after the spatial attention module SA.

[0028] Based on the existing YOLOv8 model, this embodiment adds a spatial attention module SA after the feature extraction network of this model and before the detection head to improve the small target detection capability of the baggage detection model. Figure 2 The spatial attention module SA specifically includes the following modules: separation module, fusion module and selection module. Figure 3 As shown in Figure 2, the workflow of the spatial attention module includes: S401, separation: extract the feature map Perform a separation operation, that is, perform 3×3 convolution and 5×5 hole convolution respectively, to obtain two different feature maps and , where R, H, W, and C are the height, width, and number of channels of the feature atlas and feature map, respectively; S402, fusion: the feature map obtained by the separation module and Fusion, that is , and perform channel-based global average pooling and global maximum pooling operations on the fused feature map in turn to obtain a feature map that counts the feature information of each spatial position and , and perform channel concatenation on the two feature maps to obtain the feature map ; S403, selection: the feature map obtained by the fusion module After 7×7 convolution, pass it into the Sigmoid function to get two feature maps and ,Will and Respectively and Multiply and fuse the output feature maps to obtain the spatial attention feature map after feature information enhancement. .

[0029] This example uses an enhanced data set to optimize the baggage detection model. During the training process, the initial learning rate is set to 0.01, the weight decay coefficient is set to 0.0005, and the total number of iterations is set to 100. Experimental results show that by adding a spatial attention mechanism, the important spatial regions are selectively highlighted, which effectively improves the detection performance of the original YOLOv8 model and achieves an average accuracy improvement of 0.82%.

[0030] S103, pre-processing the video stream, combining with the baggage detection model optimized in S102, adopting a baggage swap detection method based on time and space dimensions, realizing baggage swap detection and issuing an alarm.

[0031] The video stream preprocessing method adopted by the present invention comprises the following steps: S601: For the current video stream, an interactive ROI selection method is used to select a baggage swap detection area according to the camera's field of view and the area where the protective net is located; S602: Extract frames from the video stream at intervals of 25 frames to obtain the current image to be detected.

[0032] The luggage and vehicles moving under the detection area will also interfere with the detection results. The luggage swap detection method based on time and space dimensions adopted in this embodiment can effectively alleviate the impact caused by the above factors. The method flow chart is as follows: Figure 4 As shown, the following steps are included: S701, for the first image to be detected, initialize a counter; S702: Set the area ratio threshold to 0.4, use the baggage detection model optimized in S102 to infer the image to be detected, and calculate the size of the overlapping part of the baggage boundary box obtained by inference with the area to be detected. , and the size of the bounding box ,like If it is greater than or equal to 0.4, it is considered as checked-in luggage; S703, setting the time threshold to 1 second, and calculating whether the detected luggage image meets the time stability requirement, that is, whether the time stamp difference between two consecutive images of the detected luggage is less than or equal to 1 second; S704, setting the IOU threshold to 0.9, and for images that meet the requirements of temporal stability, calculating whether the detected luggage meets the requirements of spatial stability, that is, whether the luggage bounding box IOUs of two consecutive images of detected luggage are greater than or equal to 0.9, and if so, adding one to the counter; S705. Set the counter threshold to 10. For images that meet the requirements of temporal and spatial stability, determine whether the counter is greater than or equal to the counting threshold. When the counter is greater than or equal to 10, it is determined that luggage substitution has occurred and an alarm is issued. Otherwise, re-execute S702.

[0033] It can be understood that in other embodiments, the area ratio threshold, time threshold, IOU threshold, and counter threshold can be set to other values ​​as needed.

[0034] The present embodiment provides a baggage swap detection method for an automatic sorting machine. The data enhancement strategy is combined to enhance the diversity of the baggage data set, thereby improving the robustness and accuracy of the model when processing complex scenarios. A deep learning method is adopted to detect baggage targets. The attention module selectively highlights important spatial areas to enhance the model's ability to detect small targets. The baggage swap phenomenon is detected by combining temporal and spatial stability, which can effectively alleviate the interference caused by the complex and changeable environment of the baggage conveying system, reduce false alarms and improve the detection effect.

[0035] This embodiment also provides a baggage switching detection mechanism for an automatic sorting machine, which is applied to the baggage switching detection method for an automatic sorting machine to perform real-time detection and analysis of baggage images. The mechanism includes: Baggage image acquisition module, used to collect historical videos of the baggage conveying system, construct the original baggage dataset, and perform data enhancement; A model creation module, which is used to create a baggage detection model in combination with the attention mechanism and optimize the model using the enhanced baggage dataset; The data processing module is used to process the acquired data to determine whether luggage substitution has occurred.

[0036] An embodiment of the present invention further provides an electronic device, including a processor and a memory, wherein the memory stores executable instructions that can be executed by the processor, and the processor executes the executable instructions to implement a baggage substitution detection method for an automatic sorting machine.

[0037] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for detecting baggage substitution in an automatic sorting machine, characterized in that: The steps include: S101, collecting historical videos of the baggage conveying system, constructing an original baggage dataset, performing data enhancement, and obtaining an enhanced baggage dataset; S102, setting a spatial attention module after the feature extraction network and before the detection head, creating a baggage detection model in combination with the attention mechanism, and optimizing the baggage detection model using the enhanced baggage data set obtained in S101; the spatial attention module is used to improve the small target detection capability of the baggage detection model; S103, pre-processing the video stream, combining the luggage detection model optimized in S102, adopting a luggage swap detection method based on time and space dimensions, realizing luggage swap detection and issuing an alarm.

2. The automatic sorting machine baggage substitution detection method according to claim 1, characterized in that: The construction of the original baggage data set described in S101 includes the following steps: extracting frames from the historical video stream, cleaning out luggage pictures that meet format requirements, have no duplication, and have balanced categories, and marking the cleaned luggage pictures with selection boxes to obtain the original baggage data set.

3. The automatic sorting machine baggage substitution detection method according to claim 1, characterized in that: The data enhancement means described in S101 include any one or more of the following: Data enhancement method 1: If the number of luggage images blocked by the protective net screened out in S101 is less than a preset threshold, a Mixup method is used for data enhancement, that is, the cropped protective net and luggage images are linearly mixed in a certain proportion to generate a new training sample, and their labels are linearly mixed accordingly; Data enhancement method 2: If the number of luggage images blocked by the fixed support screened out in S101 is less than a preset threshold, a Mixup method is used for data enhancement, that is, the cropped fixed support and luggage images are linearly mixed in a certain proportion to generate a new training sample, and their labels are linearly mixed accordingly; Data enhancement method three: If the number of baggage images with baggage targets blocking each other or with multiple baggage targets screened out in S101 is less than a preset threshold, the CutMix method is used for data enhancement, that is, the cropped baggage images are randomly pasted onto other baggage images, and the corresponding labels are adjusted to generate new training samples.

4. The automatic sorting machine baggage substitution detection method according to claim 1, characterized in that: The spatial attention module at least includes the following modules: a separation module, a fusion module and a selection module. The workflow of the spatial attention module includes: S401, separation: performing convolution operations on the extracted feature maps with multiple convolution kernels of different sizes respectively to obtain feature maps with different receptive field sizes; S402, fusion: the feature maps obtained by the separation module are fused, and the channel-based global average pooling and global maximum pooling operations are performed on the fused feature maps in turn to obtain feature maps that count feature information of each spatial position, and the channel splicing operation is performed on the two feature maps; S403, selection: perform a convolution operation on the feature map obtained by the fusion module and pass it into the Sigmoid function to obtain multiple feature maps, which are multiplied with the feature maps obtained by the separation module and fused and outputted to obtain a feature map with enhanced feature information.

5. The method for detecting baggage substitution in an automatic sorting machine according to claim 1, characterized in that: The baggage detection model is created by combining the attention mechanism in S102, and the baggage detection model at least includes a feature extraction network, a feature fusion network, and a detection head, wherein: The feature extraction network includes a first convolutional layer, a second convolutional layer, a first C2f module, a third convolutional layer, a second C2f module, a fourth convolutional layer, a third C2f module, a fifth convolutional layer, a fourth C2f module and an SPPF module, which are arranged in sequence; The feature fusion network includes a first upsampling layer, a first Concat fusion layer, a fifth C2f module, a second upsampling layer, a second Concat fusion layer, a sixth C2f module, a sixth convolutional layer, a third Concat fusion layer, a seventh C2f module, a seventh convolutional layer, a fourth Concat fusion layer, and an eighth C2f module, and spatial attention modules are added after the sixth C2f module, the seventh C2f module, and the eighth C2f module respectively; The detection head includes a first decoupling head detection layer, a second decoupling head detection layer and a third decoupling head detection layer, which are respectively arranged after the spatial attention module.

6. The method for detecting baggage substitution in an automatic sorting machine according to claim 1, characterized in that: The pre-processing of the video stream described in S103 specifically includes the following steps: S601: For the current video stream, an interactive ROI selection method is used to select a baggage swap detection area according to the camera's field of view and the area where the protective net is located; S602: Extract frames from the video stream to obtain the current image to be detected.

7. The method for detecting baggage substitution in an automatic sorting machine according to claim 1, characterized in that: The baggage substitution detection method based on time and space dimensions described in S103 specifically comprises the following steps: S701, for the first image to be detected, initialize a counter; S702, using the baggage detection model optimized in S102 to infer the image to be detected, and calculating the size of the overlapping part of the baggage bounding box obtained by inference with the area to be detected, as well as the size occupied by the bounding box. If the calculated result is greater than or equal to the area proportion threshold, the baggage is deemed to be detected; S703: For the image of the checked luggage, calculate whether the checked luggage meets the time stability requirement, that is, whether the time stamp difference between two consecutive images of the checked luggage is less than or equal to the time threshold; S704: For images that meet the requirements of temporal stability, calculate whether the detected luggage meets the requirements of spatial stability, that is, whether the luggage bounding box IOUs of two consecutive images of detected luggage are greater than or equal to the IOU threshold. If so, add one to the counter. S705. For images that satisfy both temporal stability and spatial stability, determine whether the counter is greater than or equal to a counting threshold. If the counter is greater than or equal to the counting threshold, determine that luggage substitution has occurred and issue an alarm. Otherwise, re-execute S702.

8. A baggage substitution detection mechanism for an automatic sorting machine, characterized in that: The method for detecting baggage substitution in an automatic sorting machine as claimed in any one of claims 1 to 7 is used for real-time detection and analysis of baggage images, and the mechanism comprises: Baggage image acquisition module, used to collect historical videos of the baggage conveying system, construct the original baggage dataset, and perform data enhancement; A model creation module, for creating a baggage detection model in combination with an attention mechanism, and optimizing the baggage detection model using an enhanced baggage dataset; The data processing module is used to process the acquired data to determine whether luggage substitution has occurred.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores executable instructions that can be executed by the processor, and the processor executes the executable instructions to implement the baggage substitution detection method for an automatic sorting machine as described in any one of claims 1 to 7.

Images