Logistics accurate tracking system combining Internet of Things and visual identification

Abstract

The invention relates to a logistics accurate tracking system combining the Internet of Things and visual identification. The system comprises a management cloud server, a data processing host, a plurality of terminal devices, a gantry terminal, a handheld terminal and a vehicle-mounted terminal. Each terminal device carries out image acquisition on a transported piece in a logistics link, and extracts the characteristics of a logistics sheet and the shape of the transported piece through a visual identification technology. And the data processing host receives the image data, performs shape transformation and data compression, and then uploads the image data to the management cloud server. And the management cloud server compares the image feature information of different timestamps to judge whether the transport piece is deformed or damaged, and informs related personnel of a result. In addition, the system further comprises a client terminal, and a client is allowed to remotely obtain transportation whole-process data of logistics transportation. According to the system, through integration of the Internet of Things technology and the visual identification technology, accurate tracking and state monitoring of logistics transportation are realized, and the efficiency and accuracy of logistics management are improved.

Description

Technical Field

[0001] The present invention relates to the technical fields of logistics and Internet of Things (IoT), and particularly to a precise logistics tracking system that combines IoT and visual recognition. Background Art

[0002] With the rapid development of e-commerce and global trade, the logistics industry is facing unprecedented challenges and opportunities. As a key technology to improve logistics efficiency, reduce costs, and enhance customer satisfaction, the precise logistics tracking system has received extensive attention. Traditional logistics tracking systems mainly rely on barcodes or RFID tags. These technologies have achieved the identification and tracking of goods to a certain extent, but there are some limitations, such as being unable to obtain the real-time status of goods, and it is difficult to detect the deformation or damage of goods during transportation.

[0003] In recent years, the development of IoT technology and visual recognition technology has provided new solutions for logistics tracking systems. IoT technology enables data exchange and communication between devices through various sensors and network connections, providing a rich data source for logistics tracking. Visual recognition technology, especially image recognition technology based on deep learning, can automatically identify and analyze the object features in images, providing a powerful tool for the identification and status monitoring of logistics goods.

[0004] However, there are still some deficiencies in the existing logistics tracking systems in integrating IoT and visual recognition technologies. For example, the images captured by different terminal devices (such as gantry cranes, handheld scanners, vehicle-mounted cameras, etc.) may vary due to factors such as angles and lighting conditions, resulting in the accuracy and consistency of image recognition being affected. In addition, the transmission and storage of a large amount of image data also require a large amount of bandwidth and storage resources, increasing the operating cost of the system. Summary of the Invention

[0005] To solve the above problems, the present invention provides a precise logistics tracking system that combines IoT and visual recognition, including:

[0006] A management cloud server for remote management and data storage;

[0007] A data processing host, remotely connected to the management cloud server, for receiving and processing data from terminal devices;

[0008] The terminal devices include gantry terminals, handheld terminals, and vehicle-mounted terminals;

[0009] The gantry terminal is in the form of a gantry crane, and is used to collect images of the logistics waybill and the outer shape of the shipped item when the shipped item is being transported on the logistics conveyor belt, and obtain the images of the logistics waybill and the outer shape of the shipped item;

[0010] The handheld terminal is in the form of a handheld scanner gun, which is used for workers to collect images of the logistics list and the outer shape of the logistics shipment when handling the logistics shipment, and obtain the images of the logistics list and the outer shape of the logistics shipment;

[0011] The vehicle-mounted terminal is in a vehicle-mounted form, which is used to collect images of the logistics list and the outer shape of the logistics shipment when loading and unloading the logistics shipment, and obtain the images of the logistics list and the outer shape of the logistics shipment;

[0012] The gantry terminal, the handheld terminal and the vehicle-mounted terminal communicate with the data processing host through wireless or wired means to realize real-time transmission and processing of image data; the data processing host communicates with the management cloud server through wireless or wired network to realize data upload and instruction reception;

[0013] After the data processing host obtains the data of the gantry terminal, the handheld terminal or the vehicle-mounted terminal, it compresses the data and sends the compressed data to the management cloud server.

[0014] In some embodiments, the structures of the gantry terminal, the handheld terminal and the vehicle-mounted terminal are as follows:

[0015] The gantry terminal includes:

[0016] At least one top camera, installed on the top of the gantry, for capturing images of the top of the logistics shipment; at least one side camera, installed on the side of the gantry, for capturing images of the side of the logistics shipment; an image processing unit, for preliminarily processing the images obtained by the top camera and the side camera; a communication module, for transmitting the processed image data to the data processing host through wireless or wired means;

[0017] The handheld terminal includes:

[0018] A single camera, installed on the shell of the handheld terminal, for handheld capturing images of the logistics list and the outer shape of the logistics shipment; an image processing unit, for preliminarily processing the images obtained by the single camera; a communication module, for transmitting the processed image data to the data processing host through wireless or wired means; a handheld shell, for facilitating workers to carry and operate;

[0019] The vehicle-mounted terminal includes:

[0020] At least one left camera, installed at the upper left corner of the vehicle cargo compartment, for obliquely shooting to capture images of the upper left part of the logistics shipment; at least one right camera, installed at the upper right corner of the vehicle cargo compartment, for obliquely shooting to capture images of the upper right part of the logistics shipment; an image processing unit, for preliminarily processing the images obtained by the left and right cameras; a communication module, for wirelessly or wiredly transmitting the processed image data to the data processing host; a vehicle-mounted installation structure, for fixing the vehicle-mounted terminal on the vehicle;

[0021] Among them, the top camera, side camera, single camera, left camera and right camera all have visual recognition functions, and can recognize barcodes, two-dimensional code information on the logistics order, and the external shape characteristics of the shipment.

[0022] In some embodiments, the process of the data processing host processing data from the gantry terminal, handheld terminal and vehicle-mounted terminal includes the following steps:

[0023] Receiving image data from the gantry terminal, handheld terminal and vehicle-mounted terminal; performing preprocessing on the received image data, including color calibration, denoising, and enhancing contrast;

[0024] Segmenting the image area of the logistics shipment from the preprocessed image, and accurately separating the logistics shipment from the background by applying background subtraction, edge detection or machine learning algorithms;

[0025] Identifying the two-dimensional code and barcode on the logistics order in the segmented logistics shipment image, and using visual recognition technology to extract the shape features of the two-dimensional code and barcode as the shape transformation reference;

[0026] According to the result of the reference shape recognition, performing shape transformation on the segmented logistics shipment image, applying affine transformation or perspective transformation to adjust the shape of the logistics shipment image to align it with the reference shape;

[0027] Converting the image data after shape transformation into a unified data format, compressing the image data, and defining a unified data structure, including the image data of the logistics shipment, the recognized logistics order information, and the timestamp, to ensure the consistency of the data formats of different terminals and facilitate subsequent data processing and analysis;

[0028] Uploading the data to the management cloud server through wireless or wired network.

[0029] In some embodiments, according to the result of the reference shape recognition, performing shape transformation on the segmented logistics shipment image, specifically:

[0030] Using visual recognition technology, identify and extract the shape features of the QR code and barcode on the logistics waybill from the segmented logistics shipment image, and determine the four corner point coordinates of the QR code and barcode as the vertices of the reference shape;

[0031] Calculate the transformation parameters required to transform each reference shape;

[0032] The transformation parameters include the rotation angle (θ), the scaling factor (s), and the translation vector (t x ,t y ), and the calculation formulas are as follows:

[0033]

[0034] t x = x3 - s·x1cos(θ) - s·y1sin(θ)

[0035] t y = y3 - s·x1sin(θ) + s·y1cos(θ)

[0036] Where (x1, y1) and (x2, y2) are two corner points of the reference shape, and (x3, y3) and (x4, y4) are two corner points of the target shape;

[0037] Use the calculated transformation parameters to perform shape transformation on the images obtained by each terminal device. The transformation formula is as follows:

[0038]

[0039] Where (x, y) are the point coordinates in the original image, and (x′, y′) are the point coordinates after transformation.

[0040] In some embodiments, the process of compressing image data in the data processing host includes the following steps:

[0041] Feature line extraction:

[0042] For the image after shape transformation, extract the feature lines of the logistics shipment. The feature lines include the contour line, the packing line, and the joint line; Contour line: the outer edge line of the logistics shipment; Packing line: the shape line of the packing material of the logistics shipment; Joint line: the line at the joint position on the package of the logistics shipment;

[0043] Segmented block color extraction:

[0044] Segment the transformed image into multiple blocks, extract the representative color for each block, and store the representative color of each block and its position information in the image by calculating the average color value of all pixels within the block;

[0045] Compress the extracted feature line coordinate data and segmentation block color data, apply data compression algorithms, including Huffman coding, arithmetic coding, or LZW coding, to further reduce the data volume;

[0046] Store the compressed feature lines and color data in a unified data structure, and define the data structure to include the compressed image feature information, the identified logistics order information, and the timestamp.

[0047] In some embodiments, the management cloud server includes the following functions:

[0048] Receive the compressed image feature information from multiple data processing hosts, including the feature lines and segmentation block color data;

[0049] Sort the data corresponding to the same logistics order according to the timestamp of the image feature information to ensure that the image feature information is compared in chronological order;

[0050] Compare the image feature information with different timestamps for the same logistics order information. The comparison includes changes in feature lines, changes in segmentation block colors, and the correspondence between feature lines and segmentation block colors;

[0051] Use machine learning algorithms or rule engines to analyze the comparison results to determine whether the logistics shipment has been deformed or damaged. Deformation detection is achieved by comparing changes in the geometric shape of the feature lines, and damage detection is achieved by comparing changes in the segmentation block colors and the integrity of the feature lines;

[0052] Record the detection results, including the type, degree, and time point of deformation or damage; According to the system settings, send notifications to relevant personnel or systems via email, text message, or in-system notification;

[0053] Store the processed data and detection results on the cloud server to ensure data security and traceability.

[0054] In some embodiments, a standard color card is set on the waybill for color calibration;

[0055] Each data processing host is connected to at least one of the gantry terminal, handheld terminal, and vehicle terminal. Among the gantry terminal, handheld terminal, and vehicle terminal, each terminal can send at least one image including the waybill to the data processing host.

[0056] In some embodiments, a customer terminal is further included. The customer terminal is remotely connected to the management cloud server and can obtain the full transportation process data of the logistics shipment from the management cloud server, including the type, degree, and time point of deformation or damage.

[0057] In some embodiments, the data processing host can also directly send the original captured images to the management cloud server; the management cloud service area deletes the uploaded original images after storing them for a certain period of time to save storage space.

[0058] In some embodiments, the management cloud service area deletes the uploaded original images after storing them for 90 days.

[0059] The beneficial effects of the present invention are as follows:

[0060] The present invention proposes a logistics precise tracking system that combines the Internet of Things and visual recognition. The system collects images of logistics shipments during the logistics conveyor belt, the process of workers handling, and the loading and unloading of goods, and uses visual recognition technology to identify and analyze the logistics orders and the shapes of the shipments. By performing shape transformation on the images captured by different terminals based on the QR codes and barcodes on the logistics orders, effective comparison and analysis between different images are achieved. In addition, the system also adopts data compression technologies such as feature line extraction and segmented block color extraction, significantly reducing the requirements for data transmission and storage, and improving the efficiency and accuracy of logistics tracking.

[0061] The logistics precise tracking system of the present invention can not only monitor the status of logistics shipments in real time, discover and handle problems such as deformation or damage in a timely manner, but also provide customers with data on the entire transportation process, improving the transparency of logistics services and customer satisfaction. By integrating the Internet of Things and visual recognition technologies, the present invention provides an innovative solution for the logistics industry, with broad application prospects and market value. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0063] Attached Figure 1 is a schematic diagram of the overall architecture of the system of the present invention;

[0064] Attached Figure 2 is a schematic diagram of the structure of the gantry terminal of the present invention;

[0065] Attached Figure 3 is a schematic diagram of the structure of the vehicle-mounted terminal of the present invention;

[0066] Attached Figure 4 is a schematic diagram of the shape of the waybill of the present invention.

[0067] Among them: 1 top camera, 2 side cameras, 3 left camera, 4 right camera. Detailed implementation mode

[0068] Embodiment 1:

[0069] Refer to Figures 1 to 4 , the present invention provides a logistics precise tracking system combining the Internet of Things and visual recognition, including:

[0070] A management cloud server for remote management and data storage;

[0071] A data processing host, remotely connected to the management cloud server, for receiving and processing data from terminal devices;

[0072] The terminal devices include a gantry terminal, a handheld terminal, and a vehicle-mounted terminal;

[0073] The gantry terminal is in the form of a gantry, and is used for collecting images of the logistics waybill and the shape of the shipped item when the shipped item is transported on the logistics conveyor belt, and obtaining images of the logistics waybill and the shape of the shipped item;

[0074] The handheld terminal is in the form of a handheld scanner gun, and is used for collecting images of the logistics waybill and the shape of the shipped item when the worker transports the shipped item, and obtaining images of the logistics waybill and the shape of the shipped item;

[0075] The vehicle-mounted terminal is in the form of a vehicle, and is used for collecting images of the logistics waybill and the shape of the shipped item when loading and unloading the shipped item, and obtaining images of the logistics waybill and the shape of the shipped item;

[0076] The gantry terminal, the handheld terminal, and the vehicle-mounted terminal communicate with the data processing host wirelessly or by wire to realize real-time transmission and processing of image data; the data processing host communicates with the management cloud server wirelessly or by wire network to realize data upload and instruction reception;

[0077] After the data processing host obtains the data of the gantry terminal, the handheld terminal or the vehicle-mounted terminal, it compresses the data and sends the compressed data to the management cloud server.

[0078] In some embodiments, the structures of the gantry terminal, the handheld terminal, and the vehicle-mounted terminal are as follows:

[0079] The gantry terminal includes:

[0080] At least one top camera, installed on the top of the gantry, for capturing images of the top of the logistics shipment; at least one side camera, installed on the side of the gantry, for capturing images of the side of the logistics shipment; an image processing unit, for preliminarily processing the images obtained by the top camera and the side camera; a communication module, for transmitting the processed image data to the data processing host wirelessly or by wire;

[0081] The handheld terminal includes:

[0082] A single camera, installed on the housing of the handheld terminal, for hand-held capturing of images of the logistics label and the shape of the shipment of the logistics shipment; an image processing unit, for preliminarily processing the images obtained by the single camera; a communication module, for transmitting the processed image data to the data processing host wirelessly or by wire; a handheld housing, for facilitating the carrying and operation by workers;

[0083] The vehicle-mounted terminal includes:

[0084] At least one left camera, installed at the upper left corner of the vehicle cargo hold, for obliquely shooting and capturing images of the upper left of the logistics shipment; at least one right camera, installed at the upper right corner of the vehicle cargo hold, for obliquely shooting and capturing images of the upper right of the logistics shipment; an image processing unit, for preliminarily processing the images obtained by the left camera and the right camera; a communication module, for transmitting the processed image data to the data processing host wirelessly or by wire; a vehicle-mounted installation structure, for fixing the vehicle-mounted terminal on the vehicle;

[0085] Wherein, the top camera, the side camera, the single camera, the left camera and the right camera all have a visual recognition function, and can recognize barcodes, two-dimensional code information on the logistics label, and the shape characteristics of the shipment.

[0086] In some embodiments, the process of the data processing host processing data from the gantry terminal, the handheld terminal and the vehicle-mounted terminal includes the following steps:

[0087] Receiving image data from the gantry terminal, the handheld terminal and the vehicle-mounted terminal; preprocessing the received image data, including color calibration, denoising, and enhancing contrast;

[0088] Segmenting the image area of the logistics shipment from the preprocessed image, and accurately separating the logistics shipment from the background by applying background subtraction, edge detection or machine learning algorithms;

[0089] Identifying two-dimensional codes and barcodes on the logistics label in the segmented logistics shipment image, and using visual recognition technology to extract the shape characteristics of the two-dimensional codes and barcodes as the shape transformation reference;

[0090] According to the result of the reference shape recognition, perform a shape transformation on the segmented logistics shipment image, apply an affine transformation or a perspective transformation to adjust the shape of the logistics shipment image to align it with the reference shape;

[0091] Convert the image data after the shape transformation into a unified data format, compress the image data, and define a unified data structure, including the image data of the logistics shipment, the recognized logistics order information, and the timestamp, to ensure that the data from different terminals is consistent in format, facilitating subsequent data processing and analysis;

[0092] Upload the data to the management cloud server through a wireless or wired network.

[0093] In some embodiments, according to the result of the reference shape recognition, perform a shape transformation on the segmented logistics shipment image. Specifically:

[0094] Use visual recognition technology to recognize and extract the shape features of the QR code and barcode on the logistics order from the segmented logistics shipment image, and determine the four corner coordinates of the QR code and barcode as the vertices of the reference shape;

[0095] Calculate the transformation parameters required for transforming each reference shape;

[0096] The transformation parameters include the rotation angle (θ), the scaling factor (s), and the translation vector (t x ,t y ), and the calculation formulas are as follows:

[0097]

[0098] t x = x3 - s·x1cos(θ) - s·y1sin(θ)

[0099] t y = y3 - s·x1sin(θ) + s·y1cos(θ)

[0100] where (x1,y1) and (x2,y2) are two corner points of the reference shape, and (x3,y3) and (x4,y4) are two corner points of the target shape;

[0101] Use the calculated transformation parameters to perform a shape transformation on the images obtained by each terminal device. The transformation formula is as follows:

[0102]

[0103] where (x,y) are the point coordinates in the original image, and (x′,y′) are the point coordinates after the transformation.

[0104] In some embodiments, the process of compressing image data in the data processing host includes the following steps:

[0105] Feature line extraction:

[0106] For the image after shape transformation, extract the feature lines of the logistics shipment. The feature lines include the contour line, the packing line, and the joint line; Contour line: The outer edge line of the logistics shipment; Packing line: The shape line of the packing material of the logistics shipment; Joint line: The line at the joint position on the packaging of the logistics shipment.

[0107] Segment block color extraction:

[0108] Divide the transformed image into multiple blocks, extract the representative color for each block, calculate the average color value of all pixels within the block, and store the representative color of each block and its position information in the image.

[0109] Compress the extracted feature line coordinate data and segment block color data, apply data compression algorithms, including Huffman coding, arithmetic coding, or LZW coding, to further reduce the data volume.

[0110] Store the compressed feature lines and color data in a unified data structure, and define the data structure to include the compressed image feature information, the identified logistics order information, and the timestamp.

[0111] In some embodiments, the management cloud server includes the following functions:

[0112] Receive the compressed image feature information from multiple data processing hosts, including the feature lines and segment block color data.

[0113] Sort the data corresponding to the same logistics order according to the timestamp of the image feature information to ensure that the image feature information is compared in chronological order.

[0114] Compare the image feature information with different timestamps for the same logistics order information. The comparison includes changes in feature lines, changes in segment block colors, and the correspondence between feature lines and segment block colors.

[0115] Use machine learning algorithms or rule engines to analyze the comparison results to determine whether the logistics shipment has undergone deformation or damage. Deformation detection is achieved by comparing changes in the geometric shape of the feature lines, and damage detection is achieved by comparing changes in segment block colors and the integrity of the feature lines.

[0116] Record the detection results, including the type, degree, and time point of deformation or damage; According to system settings, send notifications to relevant personnel or systems via email, text message, or in-system notification.

[0117] Store the processed data and detection results on the cloud server to ensure data security and traceability.

[0118] In some embodiments, a standard color card is set on the waybill for color calibration.

[0119] Each data processing host is connected to at least one of the gantry terminal, the handheld terminal, and the vehicle-mounted terminal. Among the gantry terminal, the handheld terminal, and the vehicle-mounted terminal, each terminal can send at least one image including the waybill to the data processing host.

[0120] In some embodiments, a customer terminal is further included. The customer terminal is remotely connected to the management cloud server and can obtain the entire transportation process data of the logistics shipment from the management cloud server, including the type, degree, and time point of deformation or damage.

[0121] In some embodiments, the data processing host can also directly send the original image taken to the management cloud server; the management cloud server deletes the uploaded original image after storing it for a certain period of time to save storage space.

[0122] In some embodiments, the management cloud server deletes the uploaded original image after storing it for 90 days.

[0123] Embodiment 2:

[0124] The following is a specific embodiment, which describes the working steps of a logistics precise tracking system combining the Internet of Things and visual recognition:

[0125] Image acquisition during the logistics transportation process:

[0126] The gantry terminal automatically performs top view and side view image acquisition on the passing logistics shipments on the logistics conveyor belt.

[0127] The handheld terminal is manually triggered by the worker when handling the logistics shipment to capture the images of the waybill and the shape of the shipment.

[0128] The vehicle-mounted terminal performs image acquisition on the logistics shipment during loading and unloading, including the perspectives of the upper left corner and the upper right corner.

[0129] Initial image processing:

[0130] The image processing unit of each terminal performs initial processing on the acquired images, such as color calibration, denoising, and contrast enhancement.

[0131] Image transmission:

[0132] The processed image data is transmitted to the data processing host in real time via wireless or wired means.

[0133] Image segmentation and feature extraction:

[0134] The data processing host divides the received image and separates the logistics shipment from the background.

[0135] Identify and extract the QR code and barcode on the logistics label as the benchmark for shape transformation.

[0136] Shape transformation:

[0137] According to the reference shape, perform affine transformation or perspective transformation on the segmented logistics shipment image to align it with the reference shape.

[0138] Data compression:

[0139] Extract the characteristic lines (contour lines, packing lines, joint lines) of the transformed image and the representative colors of the segmented blocks.

[0140] Compress the coordinate data and color data of the characteristic lines, such as using Huffman coding.

[0141] Data upload:

[0142] Upload the compressed data, including image feature information, logistics label information, and timestamp, to the management cloud server.

[0143] Data analysis and monitoring:

[0144] The management cloud server compares and analyzes the image feature information of the same logistics label received at different timestamps.

[0145] Use machine learning algorithms or rule engines to determine whether the logistics shipment has been deformed or damaged.

[0146] Result notification:

[0147] Record the detection results and send notifications to relevant personnel according to the system settings.

[0148] Customer access:

[0149] Customers remotely connect to the management cloud server through the customer terminal to obtain the entire transportation process data of the logistics shipment, including deformation or damage information.

[0150] Data management:

[0151] The management cloud server periodically deletes the original uploaded images to save storage space, such as deleting them after 90 days of storage.

[0152] To this end, the above-described embodiments are provided for purposes of illustration and description. It is not intended to be exhaustive or to limit the disclosure. The individual elements or features of a particular embodiment are generally not limited to that particular embodiment, but, where applicable, may be interchanged and used in selected embodiments, even if not specifically shown or described. In many respects, the same elements or features may also vary. Such variations are not to be regarded as a departure from the disclosure, and all such modifications are intended to be included within the scope of the disclosure.

[0153] Example embodiments are provided so that this disclosure will be thorough, and will fully convey the scope to those skilled in the art. To thoroughly understand the embodiments of this disclosure, numerous specific details are set forth, such as examples of specific components, devices, and methods. Obviously, for those skilled in the art, specific details are not required, and the example embodiments may be implemented in many different forms, and neither should be construed as limiting the scope of the disclosure. In some example embodiments, well-known processes, well-known device structures, and well-known technologies are not described in detail.

[0154] Herein, specific terminology is used for the sole purpose of describing particular example embodiments and is not intended to be limiting. Unless the context clearly dictates otherwise, the singular forms "a" and "the" used herein may also be intended to include the plural forms. The terms "comprising" and "having" are inclusive and thus specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or combinations thereof. Unless specifically indicated the order of performance, the method steps, processes, and operations described herein are not to be construed as necessarily requiring to be performed in the particular order discussed and shown. It should also be understood that additional or alternative steps may be employed.

Claims

1. A logistics precise tracking system combining the Internet of Things and visual recognition, characterized in that, Including: A management cloud server for remote management and data storage; A data processing host, remotely connected to the management cloud server, for receiving and processing data from terminal devices; The terminal devices include a gantry terminal, a handheld terminal, and a vehicle-mounted terminal; The gantry terminal is in the form of a gantry, and is used for collecting images of the logistics waybill and the shape of the shipped item when the shipped item is transported on the logistics conveyor belt, and obtaining images of the logistics waybill and the shape of the shipped item; The handheld terminal is in the form of a handheld scanner gun, and is used for collecting images of the logistics waybill and the shape of the shipped item when a worker transports the shipped item, and obtaining images of the logistics waybill and the shape of the shipped item; The vehicle-mounted terminal is in a vehicle-mounted form, and is used for collecting images of the logistics waybill and the shape of the shipped item when loading and unloading the shipped item, and obtaining images of the logistics waybill and the shape of the shipped item; The gantry terminal, the handheld terminal, and the vehicle-mounted terminal communicate with the data processing host through wireless or wired means to realize real-time transmission and processing of image data; The data processing host communicates with the management cloud server through wireless or wired network to realize data upload and instruction reception; After the data processing host obtains the data of the gantry terminal, the handheld terminal or the vehicle-mounted terminal, it compresses the data and sends the compressed data to the management cloud server.

2. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 1, characterized in that: The structures of the gantry terminal, the handheld terminal, and the vehicle-mounted terminal are as follows: The gantry terminal includes: At least one top camera, installed on the top of the gantry, for capturing images of the top of the shipped item; at least one side camera, installed on the side of the gantry, for capturing images of the side of the shipped item; an image processing unit, for preliminarily processing the images obtained by the top camera and the side camera; a communication module, for transmitting the processed image data to the data processing host through wireless or wired means; The handheld terminal includes: A single camera, installed on the housing of the handheld terminal, for hand-held capturing of images of the logistics waybill and the shape of the shipped item; an image processing unit, for preliminarily processing the images obtained by the single camera; a communication module, for transmitting the processed image data to the data processing host through wireless or wired means; a handheld housing, for facilitating the worker to carry and operate; The vehicle-mounted terminal includes: At least one left camera, installed in the upper left corner of the vehicle cargo hold, for tilted shooting to capture images of the upper left of the shipped item; at least one right camera, installed in the upper right corner of the vehicle cargo hold, for tilted shooting to capture images of the upper right of the shipped item; an image processing unit, for preliminarily processing the images obtained by the left camera and the right camera; a communication module, for transmitting the processed image data to the data processing host through wireless or wired means; a vehicle-mounted installation structure, for fixing the vehicle-mounted terminal on the vehicle; Wherein, the top camera, the side camera, the single camera, the left camera, and the right camera all have a visual recognition function, and can recognize barcodes, two-dimensional code information on the logistics waybill, and the shape characteristics of the shipped item.

3. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 2, characterized in that: The process of the data processing host processing the data from the gantry terminal, the handheld terminal and the vehicle-mounted terminal includes the following steps: Receiving image data from the gantry terminal, the handheld terminal and the vehicle-mounted terminal; preprocessing the received image data, including color calibration, denoising, and enhancing contrast; Segmenting the image area of the logistics shipment from the preprocessed image, and accurately separating the logistics shipment from the background by applying background subtraction, edge detection or machine learning algorithms; Identifying the two-dimensional code and bar code on the logistics waybill in the segmented logistics shipment image, and using visual recognition technology to extract the shape features of the two-dimensional code and bar code as the shape transformation benchmark; According to the result of the benchmark shape recognition, performing shape transformation on the segmented logistics shipment image, applying affine transformation or perspective transformation to adjust the shape of the logistics shipment image to align it with the benchmark shape; Converting the image data after shape transformation into a unified data format, compressing the image data, and defining a unified data structure, including the image data of the logistics shipment, the identified logistics waybill information, and the timestamp, to ensure the consistency of the data formats of different terminals and facilitate subsequent data processing and analysis; Uploading the data to the management cloud server through a wireless or wired network.

4. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 3, characterized in that: According to the result of the benchmark shape recognition, performing shape transformation on the segmented logistics shipment image, specifically: Using visual recognition technology to identify and extract the shape features of the two-dimensional code and bar code on the logistics waybill from the segmented logistics shipment image, and determining the four corner coordinates of the two-dimensional code and bar code as the vertices of the benchmark shape; Calculating the transformation parameters required for transforming each benchmark shape; The transformation parameters include the rotation angle (θ), the scaling factor (s), and the translation vector (t x , t y ), and the calculation formula is as follows: t x = x3 - s·x1cos(θ) - s·y1sin(θ) t y = y3 - s·x1sin(θ) + s·y1cos(θ) where, (x1, y1) and (x2, y2) are two corner points of the benchmark shape, and (x3, y3) and (x4, y4) are two corner points of the target shape; Using the calculated transformation parameters to perform shape transformation on the images obtained by each terminal device, and the transformation formula is as follows: where, (x, y) is the point coordinate in the original image, and (x′, y′) is the point coordinate after transformation.

5. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 4, characterized in that: The process of compressing the image data in the data processing host includes the following steps: Feature line extraction: For the image after shape transformation, extracting the feature lines of the logistics shipment, and the feature lines include the contour line, the packing line and the joint line; Contour line: the outer edge line of the logistics shipment; Packing line: the shape line of the packing material of the logistics shipment; Joint line: the line at the joint position on the package of the logistics shipment; Segmented block color extraction: Segmenting the transformed image into multiple blocks, extracting the representative color for each block, and storing the representative color of each block and its position information in the image by calculating the average color value of all pixels in the block; Compress the extracted feature line coordinate data and segmented block color data, and apply data compression algorithms, including Huffman coding, arithmetic coding or LZW coding, to further reduce the data volume; Store the compressed feature lines and color data in a unified data structure, and define the data structure to include the compressed image feature information, the identified logistics order information, and the timestamp.

6. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 5, characterized in that: The management cloud server includes the following functions: Receive the compressed image feature information from multiple data processing hosts, including feature line and segmented block color data; Sort the data corresponding to the same logistics order according to the timestamp of the image feature information to ensure that the image feature information is compared in chronological order; Compare the image feature information with different timestamps of the same logistics order information. The comparison includes the change of feature lines, the change of segmented block colors, and the corresponding relationship between feature lines and segmented block colors; Use machine learning algorithms or rule engines to analyze the comparison results to determine whether the logistics shipment is deformed or damaged. Deformation detection is achieved by comparing the geometric shape changes of feature lines, and damage detection is achieved by comparing the changes of segmented block colors and the integrity of feature lines; Record the detection results, including the type, degree and occurrence time point of deformation or damage; According to the system settings, send notifications to relevant personnel or systems via email, text message or in-system notification; Store the processed data and detection results on the cloud server to ensure the security and traceability of the data.

7. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 1, characterized in that: A standard color card is set on the waybill for color calibration; Each data processing host is at least connected to one of the gantry terminal, handheld terminal and vehicle-mounted terminal. Among the gantry terminal, handheld terminal and vehicle-mounted terminal, each terminal can at least send one image including the waybill to the data processing host.

8. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 1, characterized in that: It further includes a customer terminal, which is remotely connected to the management cloud server and can obtain the full-process transportation data of the logistics shipment from the management cloud server, including the type, degree and occurrence time point of deformation or damage.

9. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 1, characterized in that: The data processing host can also directly send the original image taken to the management cloud server; the management cloud service area deletes the uploaded original image after storing it for a certain period of time to save storage space.

10. The logistics precise tracking system combining the Internet of Things and visual recognition according to claim 9, characterized in that: The management cloud service area deletes the uploaded original image after storing it for 90 days.

Images