A real-time luggage tracking system based on image recognition
By setting up RFID tags and image acquisition components on the airport suitcase, combined with RFID and image recognition technology, the full process tracking of the suitcase is realized, solving the problems of low accuracy of automatic luggage recognition and insufficient data capacity in the existing technology. Improved tracking efficiency and accuracy.
Patent Information
- Application Number
- CN202410554734.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-07
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-05-07
AI Technical Summary
In the prior art, the airport luggage is not recognized by the low recognition rate of one-dimensional barcodes in the process of checking, sorting, transportation, etc., resulting in low automatic identification accuracy, which increases labor costs, and the barcode data capacity is limited, so it cannot meet luggage. The requirements tracked throughout the box.
Using a real-time luggage tracking system based on image recognition, combined with RFID tags and image acquisition components, the entire process of the luggage is tracked by identifying the color and texture characteristics of the suitcase, and real-time images are collected when the tag is left behind for recognition, combined with RFID Technology and image recognition technology, update luggage location information.
It improves the accuracy of automatic identification of suitcases, realizes full-process tracking of suitcases, reduces manual identification costs, expands data capacity, and meets the practical application needs of airport suitcase tracking.
Smart Images

Figure CN118447490B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and in particular to a real-time luggage tracking system based on image recognition. Background Art
[0002] With the continuous increase in domestic and international civil aviation transportation demand, airport luggage operation security faces many technical difficulties and cannot adapt to the ever-increasing application needs of airport security and services. As an industry unit, it is urgent to establish a full-process data collection and tracking solution for luggage from consignment to delivery to meet application needs such as effective luggage security and precise services. As a civil aviation transportation customer, after the luggage is delivered to the carrier for consignment, it is necessary to promptly obtain the circulation status of your checked luggage through convenient and efficient means to achieve satisfactory and smooth civil aviation travel.
[0003] The flow data information generated by luggage during the luggage handling process at the airport is mostly stored in multiple information systems. The software architecture, data standards, and external interface protocols of each information system are different. The flow data of each link of the luggage cannot form a unified and complete data record, which is not conducive to the operational data analysis of the full-process business of the luggage and the development of multi-party collaborative support work based on the luggage process. The luggage security information software widely used in airports is generally oriented to single-link applications such as consignment, sorting, and loading. The software functions, data flow, ease of use, and extensibility cannot well support the business needs of the full-process security of luggage. Many core processing functions of the luggage processes that tend to be procedural and applied are missing, and the core data cannot be intuitively processed and presented by the software, which brings a lot of communication query and processing analysis costs.
[0004] In the prior art, airport luggage processing mostly adopts the technical solution of laser scanning and recognition combined with one-dimensional barcode. This technical solution requires the laser to establish optical contact with the luggage barcode to complete data reading. In the current airport luggage check-in, sorting, transportation, loading and unloading, and extraction confirmation processes, there are common phenomena such as barcode damage, wrinkles, pollution or occlusion, resulting in low recognition rate and high equipment maintenance costs in the automatic barcode recognition process of the luggage in the above-mentioned processing process. The manual recognition process causes a large amount of redundant labor costs, which is not conducive to the development of airport luggage security services. In addition, the data that can be carried by the one-dimensional barcode of the luggage is very limited, corresponding to only 10 digits of digital information, including basic data such as the basic status of the luggage, the three-character code of the luggage carrier, and the 6-digit serial number of the luggage. This part of data is far from meeting the actual application needs of airport luggage tracking.
[0005] Therefore, it is necessary to provide a method and system for tracking luggage throughout the entire process at an airport, so as to solve the technical problems in the prior art of low accuracy in automatic luggage identification and inability to achieve full-process luggage tracking. Summary of the invention
[0006] The present invention provides a real-time luggage tracking system based on image recognition, comprising: a label setting module, comprising an RFID label set on a luggage to be tracked, wherein the RFID label is used to record passenger information, flight information and luggage processing link information corresponding to the luggage to be tracked; an image acquisition module, comprising image acquisition components set at a plurality of luggage processing nodes, wherein the plurality of luggage processing nodes at least include a check-in node, a security inspection node, a sorting node, a loading node, an installation node, an arrival node, a pickup node and a transfer node; a label reading module, comprising label reading components set at the plurality of luggage processing nodes; and a luggage tracking module, used to perform real-time luggage tracking based on label information read by label reading components set at the plurality of luggage processing nodes and real-time images of the luggage acquired by image acquisition components set at the plurality of luggage processing nodes.
[0007] Furthermore, the baggage tracking module performs real-time baggage tracking based on the tag information read by the tag reading components arranged at the multiple baggage handling nodes and the real-time images of the luggage acquired by the image acquisition components arranged at the multiple baggage handling nodes, including: for any baggage handling node, judging whether the RFID tag on the luggage to be tracked is lost based on the tag reading component of the luggage handling node; when judging that the RFID tag on the luggage to be tracked is not lost, the tag reading component of the luggage handling node receives the response information of the RFID tag arranged on the luggage to be tracked; based on the response information, updating the location information of the luggage to be tracked; when judging that the RFID tag on the luggage to be tracked is lost, the image acquisition component of the luggage handling node acquires the real-time image of the luggage to be tracked; based on the real-time image of the luggage to be tracked, identifying the luggage to be tracked, and updating the location information of the luggage to be tracked.
[0008] Furthermore, the luggage tracking module identifies the luggage to be tracked based on the real-time luggage image of the luggage to be tracked, including: preprocessing the real-time luggage image of the luggage to be tracked to obtain a preprocessed real-time luggage image; extracting color features of multiple first target areas of the preprocessed real-time luggage image; extracting texture features of multiple second target areas of the preprocessed real-time luggage image; obtaining a preprocessed luggage image to be compared; extracting color features of multiple first target areas of the preprocessed luggage image to be compared; extracting texture features of multiple second target areas of the preprocessed luggage image to be compared; identifying the luggage to be tracked based on the color features of multiple first target areas and multiple second target areas of the preprocessed real-time luggage image and the color features of multiple first target areas and multiple second target areas of the preprocessed luggage image to be compared.
[0009] Furthermore, the luggage tracking module obtains the pre-processed luggage image to be compared, including: obtaining flight information of the luggage to be tracked and the luggage to be tracked; based on the flight information of the luggage to be tracked and the luggage to be tracked and the luggage to be tracked, obtaining the pre-processed luggage image to be compared.
[0010] Furthermore, the luggage tracking module identifies the luggage to be tracked based on the color features of the multiple first target areas and the texture features of the multiple second target areas of the preprocessed luggage real-time image and the color features of the multiple first target areas and the texture features of the multiple second target areas of the preprocessed luggage image to be compared, including: determining the luggage color representative value corresponding to the preprocessed luggage real-time image based on the color features of the multiple first target areas of the preprocessed luggage real-time image; determining the luggage pixel space arrangement representative value corresponding to the preprocessed luggage real-time image based on the texture features of the multiple second target areas of the preprocessed luggage real-time image; determining the luggage color representative value and the luggage pixel space arrangement representative value corresponding to the preprocessed luggage real-time image based on the luggage color representative value and the luggage pixel space arrangement representative value corresponding to the preprocessed luggage real-time image. a locking identification value corresponding to the preprocessed real-time image of the suitcase; determining a suitcase color representative value corresponding to the preprocessed suitcase image to be compared based on color features of multiple first target areas of the preprocessed suitcase image to be compared; determining a suitcase pixel space arrangement representative value corresponding to the preprocessed suitcase image to be compared based on texture features of multiple second target areas of the preprocessed suitcase image to be compared; determining a locking identification value corresponding to the preprocessed suitcase image to be compared based on the suitcase color representative value and the suitcase pixel space arrangement representative value corresponding to the preprocessed suitcase image to be compared; identifying the suitcase to be tracked based on the locking identification value corresponding to the preprocessed real-time image of the suitcase and the locking identification value corresponding to the preprocessed suitcase image to be compared.
[0011] Furthermore, the baggage tracking module determines the multiple first target areas, including: acquiring sample images of multiple sample suitcases; for each of the sample suitcases, determining the size of the sample suitcase based on the sample image of the sample suitcase; grouping the multiple sample suitcases based on the size of each of the sample suitcases to determine multiple sample suitcase groups; for each of the sample suitcase groups, extracting color features of multiple candidate first target areas of the sample image of each of the sample suitcases, and determining the multiple first target areas based on the color feature similarity between the color features of the multiple candidate first target areas of the sample images of any two of the sample suitcases.
[0012] Furthermore, the baggage tracking module determines the multiple second target areas, including: for each of the sample suitcase groups, extracting texture features of multiple candidate second target areas of the sample image of each of the sample suitcases, and determining the multiple second target areas based on texture feature similarity between the texture features of the multiple candidate second target areas of the sample images of any two of the sample suitcases.
[0013] Furthermore, the system also includes: a status monitoring module, including a status monitoring component set on the suitcase to be tracked; and a tag retrieval module, used to determine the retrieval position of the RFID tag set on the suitcase to be tracked based on the suitcase status information collected by the status monitoring component set on the suitcase to be tracked.
[0014] Furthermore, the tag retrieval module determines the retrieval position of the RFID tag set on the suitcase to be tracked based on the suitcase status information collected by the status monitoring component set on the suitcase to be tracked, including: acquiring the suitcase historical vibration information collected by the status monitoring component set on the associated suitcase to be tracked of the suitcase to be tracked; and determining the retrieval position of the RFID tag set on the suitcase to be tracked based on the suitcase historical vibration information collected by the status monitoring components set on multiple associated suitcases to be tracked of the suitcase to be tracked.
[0015] Furthermore, the tag retrieval module determines the retrieval position of the RFID tag set on the suitcase to be tracked based on the historical vibration information of the suitcase collected by the status monitoring component set on the suitcase to be tracked and the historical vibration information of the suitcase collected by the status monitoring component set on the suitcase to be tracked associated with the suitcase to be tracked, including: generating a vibration change curve based on the historical vibration information of the suitcase collected by the status monitoring component set on the suitcase to be tracked associated with the suitcase to be tracked; performing empirical mode decomposition on the vibration change curve to determine the intrinsic modal components and residuals corresponding to the vibration change curve; and determining the retrieval position of the RFID tag set on the suitcase to be tracked based on the intrinsic modal components and residuals corresponding to the vibration change curve through a position determination model.
[0016] Compared with the prior art, the real-time luggage tracking system based on image recognition provided in this specification has at least the following beneficial effects:
[0017] 1. First, an RFID tag is set on the luggage to be tracked, and a tag reading component is set at multiple baggage handling nodes. The luggage to be tracked is located by using the RFID technology. When the RFID tag on the luggage to be tracked falls off, the image acquisition components set at multiple baggage handling nodes are used to collect the real-time image of the luggage. The luggage is identified based on the real-time image of the luggage. The RFID technology is combined with the image recognition technology to solve the problem of the accuracy of automatic luggage identification in the existing technology and realize the full-process tracking of the luggage.
[0018] 2. Tracking of untagged luggage by color and texture positioning on the luggage, and accurate identification of untagged luggage based on the locked identification value of the untagged luggage;
[0019] 3. The historical vibration information of the suitcases collected by the status monitoring component set on the suitcases associated with the suitcases to be tracked provides data support for the recovery of the dropped tags. Furthermore, the accurate recovery position of the RFID tag set on the suitcase to be tracked is determined by the position determination model based on the intrinsic modal components and residuals corresponding to the vibration change curve. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:
[0021] Figure 1 It is a module diagram of a real-time luggage tracking system based on image recognition shown in an embodiment of the present application;
[0022] Figure 2 FIG. 1 is a flow chart of identifying a suitcase to be tracked according to an embodiment of the present application. DETAILED DESCRIPTION
[0023] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for describing the embodiments.
[0024] Figure 1 is a module diagram of a baggage real-time tracking system based on image recognition shown in an embodiment of the present application, such as Figure 1 As shown, a real-time luggage tracking system based on image recognition may include a label setting module, an image acquisition module, a label reading module, a luggage tracking module, a status monitoring module and a label retrieval module.
[0025] The tag setting module may include an RFID tag set on the luggage to be tracked.
[0026] Among them, the RFID tag is used to record the passenger information, flight information and luggage processing information corresponding to the luggage to be tracked. The luggage processing information may include relevant information of the check-in node, security check node, sorting node, loading node, installation node, arrival node, collection node and transfer node, such as number, location, etc.
[0027] The image acquisition module may include image acquisition components arranged at a plurality of baggage handling nodes.
[0028] Among them, the multiple baggage handling nodes include at least a check-in node, a security check node, a sorting node, a loading node, an installation node, an arrival node, a pick-up node and a transfer node.
[0029] The tag reading module may include tag reading components disposed at a plurality of baggage handling nodes.
[0030] The working process of RFID tags and RFID readers is as follows:
[0031] RFID tags are mainly composed of chips and antennas. When an RFID tag enters the working area of the RFID reader's transmitting antenna, the tag's antenna will receive the wireless radio frequency signal emitted by the reader and convert it into electrical energy, thereby activating the chip. After activation, the chip starts working and converts the received electrical energy into working electrical energy through the receiving device. The memory in the tag stores the passenger information, flight information and luggage processing link information corresponding to the luggage to be tracked. The chip prepares this information in the memory. The chip converts the passenger information, flight information and luggage processing link information corresponding to the luggage to be tracked into radio frequency signals through the modulation circuit and transmits it through the tag's antenna. The RFID reader transmits the radio carrier signal outward through the transmitting antenna to form a certain working area. When the RFID tag enters this working area, the reader's receiving antenna will receive the carrier signal emitted by the tag. This signal is then transmitted to the RFID reader through the antenna's regulator. The reader demodulates and decodes the received signal and prepares to send it to the computer control system in the background.
[0032] The baggage tracking module can be used to track baggage in real time based on tag information read by tag reading components arranged at multiple baggage handling nodes and real-time images of suitcases acquired by image acquisition components arranged at multiple baggage handling nodes.
[0033] Specifically include:
[0034] For any baggage handling node, based on the tag reading component of the baggage handling node, it is determined whether the RFID tag on the luggage to be tracked is lost. Specifically, when the tag reading component can receive the response information of the RFID tag on the luggage to be tracked, it is determined that the RFID tag on the luggage to be tracked is not lost. Otherwise, it is determined that the RFID tag on the luggage to be tracked is lost.
[0035] When it is determined that the RFID tag on the luggage to be tracked is not lost, the tag reading component of the luggage handling node receives the response information of the RFID tag set on the luggage to be tracked, wherein the response information of the RFID tag includes the passenger information, flight information and luggage handling link information corresponding to the luggage to be tracked;
[0036] Based on the response information, the location information of the luggage to be tracked is updated, that is, the location of the luggage to be tracked is updated to the location of the current baggage handling node;
[0037] When it is determined that the RFID tag on the luggage to be tracked is lost, the image acquisition component of the baggage handling node acquires a real-time image of the luggage to be tracked;
[0038] Based on the real-time image of the suitcase to be tracked, the suitcase to be tracked is identified, and the position information of the suitcase to be tracked is updated.
[0039] In some embodiments, the luggage tracking module identifies the luggage to be tracked based on the real-time image of the luggage to be tracked, including:
[0040] Preprocessing the real-time image of the suitcase to be tracked to obtain the preprocessed real-time image of the suitcase;
[0041] Extracting color features of a plurality of first target areas of the preprocessed real-time image of the luggage compartment, wherein the color features of the first target area may include RGB values of each pixel in the first target area;
[0042] Extracting texture features of a plurality of second target areas of the preprocessed real-time image of the suitcase, wherein the color features of the first target area may include grayscale values of each pixel in the first target area;
[0043] Acquire a preprocessed suitcase image to be compared, wherein the suitcase image to be compared may be a suitcase image acquired at a check-in node;
[0044] Extracting color features of a plurality of first target areas of the preprocessed luggage image to be compared;
[0045] Extracting texture features of a plurality of second target areas of the preprocessed luggage image to be compared;
[0046] The luggage to be tracked is identified based on the color features of the multiple first target areas and the texture features of the multiple second target areas of the preprocessed real-time image of the luggage and the color features of the multiple first target areas and the texture features of the multiple second target areas of the preprocessed image of the luggage to be compared.
[0047] In some embodiments, the luggage tracking module obtains the pre-processed luggage image to be compared, including:
[0048] Acquiring flight information of a luggage to be tracked that is associated with the luggage to be tracked, wherein the luggage to be tracked that is associated with the luggage to be tracked may be a luggage that has been identified within a period of time before the luggage to be tracked is identified;
[0049] Based on the flight information of the luggage to be tracked, the pre-processed luggage image to be compared is obtained, for example, the image of the luggage corresponding to the flight to which the luggage to be tracked belongs but has not been identified at the current baggage handling node is associated.
[0050] Figure 2 is a flowchart of identifying a luggage to be tracked shown in an embodiment of the present application. Figure 2 As shown, in some embodiments, the luggage tracking module identifies the luggage to be tracked based on the color features of multiple first target areas and the texture features of multiple second target areas of the pre-processed real-time image of the luggage and the color features of multiple first target areas and the texture features of multiple second target areas of the pre-processed image of the luggage to be compared, including:
[0051] Determining a luggage box color representative value corresponding to the preprocessed luggage box real-time image based on color features of a plurality of first target areas of the preprocessed luggage box real-time image;
[0052] Determine, based on the texture features of the plurality of second target areas of the preprocessed real-time image of the suitcase, a representative value of the spatial arrangement of pixels of the suitcase corresponding to the preprocessed real-time image of the suitcase;
[0053] Determine a lock identification value corresponding to the preprocessed real-time image of the suitcase based on a representative value of the color of the suitcase and a representative value of the spatial arrangement of pixels of the suitcase corresponding to the preprocessed real-time image of the suitcase;
[0054] Determining a luggage box color representative value corresponding to the preprocessed luggage box image to be compared based on color features of a plurality of first target areas of the preprocessed luggage box image to be compared;
[0055] Determining a representative value of a space arrangement of luggage pixels corresponding to the preprocessed luggage image to be compared based on texture features of a plurality of second target areas of the preprocessed luggage image to be compared;
[0056] Determine a lock identification value corresponding to the preprocessed luggage image to be compared based on a luggage color representative value and a luggage pixel space arrangement representative value corresponding to the preprocessed luggage image to be compared;
[0057] The luggage to be tracked is identified based on the locking identification value corresponding to the preprocessed real-time image of the luggage and the locking identification value corresponding to the preprocessed image of the luggage to be compared.
[0058] For example, the luggage color representative value can be calculated based on the following formula:
[0059]
[0060]
[0061] Among them, P RGB is the representative value of the suitcase color, V (i,RGB) is the pixel color change parameter of the i-th first target area, I is the total number of first target areas, RGB (i,j) is the RGB value of the jth pixel in the ith first target area, and J is the total number of pixels included in the ith first target area.
[0062] The representative value of the luggage pixel spatial arrangement can be calculated based on the following formula:
[0063]
[0064]
[0065] Among them, P Gray The representative value of the luggage pixel space arrangement, V (i,Gray) is the pixel gray value change parameter of the e-th second target area, E is the total number of second target areas, F is the total number of pixels included in the e-th second target area, G (e,f) is the gray value of the f-th pixel in the e-th second target area.
[0066] The lock identification value can be calculated based on the following formula:
[0067] P=a1×P RGB +a2×P Gray
[0068] Wherein, P is the locking identification value, a1 and a2 are preset weights, and a1 and a2 are both greater than 0, and a1+a2=1.
[0069] Specifically, when the locking identification value corresponding to the preprocessed real-time image of the suitcase is consistent with the locking identification value corresponding to the preprocessed image of the suitcase to be compared, the suitcase processing link information corresponding to the suitcase image to be compared is obtained as the suitcase processing link information of the suitcase to be tracked, and the position information of the suitcase to be tracked is updated.
[0070] In some embodiments, the baggage tracking module determines a plurality of first target areas, including:
[0071] Obtain sample images of a plurality of sample suitcases;
[0072] For each sample suitcase, determining a size of the sample suitcase based on a sample image of the sample suitcase;
[0073] Based on the size of each sample suitcase, group the multiple sample suitcases to determine multiple sample suitcase groups, for example, grouping sample suitcases with similar sizes into the same sample suitcase group;
[0074] For each sample suitcase group, color features of multiple candidate first target areas of the sample image of each sample suitcase are extracted, and multiple first target areas are determined based on the color feature similarity between the color features of the multiple candidate first target areas of the sample images of any two sample suitcases.
[0075] Specifically, for the sample images of any two sample suitcases, the color feature similarity of the sample images of the two sample suitcases in the candidate first target area is determined based on the difference between the RGB values of the pixels of the candidate first target area of the sample image of one sample suitcase and the RGB values of the pixels of the candidate first target area corresponding to the position of the sample image of the other sample suitcase, and then the color feature similarity fluctuation parameter corresponding to each candidate first target area is calculated, and the candidate first target area whose color feature similarity fluctuation parameter is greater than a preset color feature similarity fluctuation parameter threshold is selected as the first target area.
[0076] For example, the color feature similarity fluctuation parameter can be calculated based on the following formula:
[0077]
[0078] Among them, V RGB is the color feature similarity fluctuation parameter, M is the total number of sample suitcases, is the color feature similarity between the m1th sample suitcase and the m2th sample suitcase in the candidate first target area.
[0079] In some embodiments, the baggage tracking module determines a plurality of second target areas, including:
[0080] For each sample suitcase group, texture features of multiple candidate second target areas of the sample image of each sample suitcase are extracted, and multiple second target areas are determined based on the texture feature similarity between the texture features of the multiple candidate second target areas of the sample images of any two sample suitcases.
[0081] Specifically, for the sample images of any two sample suitcases, based on the difference between the grayscale values of the pixels of the candidate second target area of the sample image of one sample suitcase and the grayscale values of the pixels of the candidate second target area corresponding to the position of the sample image of the other sample suitcase, the texture feature similarity of the sample images of the two sample suitcases in the candidate second target area is determined, and then the texture feature similarity fluctuation parameter corresponding to each candidate second target area is calculated, and the candidate second target area whose texture feature similarity fluctuation parameter is greater than a preset texture feature similarity fluctuation parameter threshold is selected as the second target area.
[0082] For example, the texture feature similarity fluctuation parameter can be calculated based on the following formula:
[0083]
[0084] Among them, V Gray is the texture feature similarity fluctuation parameter, is the color feature similarity between the n1th sample suitcase and the n2th sample suitcase in the candidate second target area.
[0085] The status monitoring module may include a status monitoring component disposed on the luggage to be tracked, wherein the status monitoring component may include a vibration sensor.
[0086] The tag retrieval module may be used to determine a retrieval position of an RFID tag disposed on the suitcase to be tracked based on the suitcase status information collected by the status monitoring component disposed on the suitcase to be tracked.
[0087] In some embodiments, the tag retrieval module determines the retrieval position of the RFID tag set on the luggage to be tracked based on the luggage status information collected by the status monitoring component set on the luggage to be tracked, including:
[0088] Acquire historical vibration information of the luggage collected by a status monitoring component disposed on an associated luggage to be tracked luggage;
[0089] Based on the historical vibration information of the luggage collected by the status monitoring components arranged on the multiple associated luggage cases to be tracked, the recovery position of the RFID tag arranged on the luggage case to be tracked is determined.
[0090] In some embodiments, the tag retrieval module determines the retrieval position of the RFID tag set on the luggage to be tracked based on the luggage historical vibration information collected by the status monitoring component set on the luggage to be tracked and the luggage historical vibration information collected by the status monitoring component set on the luggage to be tracked associated with the luggage to be tracked, including:
[0091] generating a vibration change curve based on historical vibration information of the luggage collected by a status monitoring component disposed on an associated luggage to be tracked luggage;
[0092] Perform empirical mode decomposition on the vibration change curve to determine the intrinsic mode components and residuals corresponding to the vibration change curve;
[0093] The retrieval position of the RFID tag set on the suitcase to be tracked is determined by a position determination model based on the intrinsic modal components and residuals corresponding to the vibration change curve, wherein the position determination model can be a Convolutional Neural Network (CNN) model.
[0094] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.
Claims
1. A real-time luggage tracking system based on image recognition, characterized in that: include: A tag setting module, comprising an RFID tag set on the luggage to be tracked, wherein the RFID tag is used to record passenger information, flight information and luggage processing link information corresponding to the luggage to be tracked; An image acquisition module, comprising image acquisition components arranged at a plurality of baggage handling nodes, wherein the plurality of baggage handling nodes at least include a check-in node, a security check node, a sorting node, a loading node, a machine installation node, an arrival node, a pick-up node and a transfer node; A tag reading module, comprising a tag reading component arranged at the plurality of baggage handling nodes; A baggage tracking module, configured to perform real-time baggage tracking based on tag information read by tag reading components arranged at the plurality of baggage handling nodes and real-time images of the luggage acquired by image acquisition components arranged at the plurality of baggage handling nodes; The baggage tracking module performs real-time baggage tracking based on the tag information read by the tag reading components arranged at the multiple baggage handling nodes and the real-time images of the luggage acquired by the image acquisition components arranged at the multiple baggage handling nodes, including: For any baggage handling node, based on the tag reading component of the baggage handling node, determining whether the RFID tag on the luggage to be tracked is lost; When it is determined that the RFID tag on the luggage to be tracked is not lost, the tag reading component of the baggage handling node receives the response information of the RFID tag set on the luggage to be tracked; Based on the response information, updating the location information of the suitcase to be tracked; When it is determined that the RFID tag on the luggage to be tracked is lost, the image acquisition component of the luggage handling node acquires a real-time image of the luggage to be tracked; Based on the real-time image of the luggage to be tracked, identifying the luggage to be tracked, and updating the location information of the luggage to be tracked; The luggage tracking module identifies the luggage to be tracked based on the real-time image of the luggage to be tracked, including: Preprocessing the real-time image of the suitcase to be tracked to obtain the preprocessed real-time image of the suitcase; Extracting color features of a plurality of first target areas of the preprocessed real-time image of the luggage compartment; Extracting texture features of a plurality of second target areas of the preprocessed real-time image of the luggage compartment; Obtaining the preprocessed luggage image to be compared; Extracting color features of a plurality of first target areas of the preprocessed luggage image to be compared; Extracting texture features of a plurality of second target areas of the preprocessed suitcase image to be compared; Identify the suitcase to be tracked based on the color features of the multiple first target areas and the texture features of the multiple second target areas of the preprocessed real-time image of the suitcase and the color features of the multiple first target areas and the texture features of the multiple second target areas of the preprocessed image of the suitcase to be compared; The luggage tracking module obtains the pre-processed luggage image to be compared, including: Obtaining flight information of the luggage to be tracked that is associated with the luggage to be tracked; Based on the flight information of the luggage to be tracked and the luggage to be tracked associated with the luggage to be tracked, obtaining the pre-processed luggage image to be compared; The luggage tracking module identifies the luggage to be tracked based on the color features of the plurality of first target areas and the texture features of the plurality of second target areas of the preprocessed real-time image of the luggage and the color features of the plurality of first target areas and the texture features of the plurality of second target areas of the preprocessed image of the luggage to be compared, including: Determining a luggage box color representative value corresponding to the preprocessed real-time image of the luggage box based on color features of a plurality of first target areas of the preprocessed real-time image of the luggage box; Determining a representative value of a space arrangement of pixels of the luggage compartment corresponding to the preprocessed real-time image of the luggage compartment based on texture features of a plurality of second target areas of the preprocessed real-time image of the luggage compartment; Determining a locking identification value corresponding to the preprocessed real-time image of the suitcase based on a representative value of a suitcase color and a representative value of a space arrangement of pixels of the suitcase corresponding to the preprocessed real-time image of the suitcase; Determining a luggage box color representative value corresponding to the preprocessed luggage box image to be compared based on color features of a plurality of first target areas of the preprocessed luggage box image to be compared; Determining a representative value of a space arrangement of luggage pixels corresponding to the preprocessed luggage image to be compared based on texture features of a plurality of second target areas of the preprocessed luggage image to be compared; Determining a locking identification value corresponding to the preprocessed luggage image to be compared based on the luggage color representative value and the luggage pixel space arrangement representative value corresponding to the preprocessed luggage image to be compared; Identifying the suitcase to be tracked based on the lock identification value corresponding to the preprocessed real-time image of the suitcase and the lock identification value corresponding to the preprocessed image of the suitcase to be compared; The baggage tracking module determines the plurality of first target areas, including: Obtain sample images of a plurality of sample suitcases; For each of the sample suitcases, determining a size of the sample suitcase based on a sample image of the sample suitcase; Based on the size of each of the sample suitcases, grouping the plurality of sample suitcases to determine a plurality of sample suitcase groups; For each of the sample suitcase groups, extract color features of multiple candidate first target areas of the sample image of each of the sample suitcases, and determine the multiple first target areas based on the color feature similarity between the color features of the multiple candidate first target areas of the sample images of any two of the sample suitcases; The baggage tracking module determines the plurality of second target areas, including: For each of the sample suitcase groups, extracting texture features of a plurality of candidate second target regions of the sample image of each of the sample suitcases, and determining the plurality of second target regions based on texture feature similarities between the texture features of the plurality of candidate second target regions of the sample images of any two of the sample suitcases; The luggage color representative value is calculated based on the following formula: Among them, P RGB is the representative value of the suitcase color, V (i,RGB) is the pixel color change parameter of the i-th first target area, I is the total number of first target areas, RGB (i,j) is the RGB value of the jth pixel in the ith first target area, and J is the total number of pixels included in the ith first target area; The representative value of the luggage pixel spatial arrangement is calculated based on the following formula: Among them, P Gray The representative value of the luggage pixel space arrangement, V (e,Gray) is the pixel gray value change parameter of the e-th second target area, E is the total number of second target areas, F is the total number of pixels included in the e-th second target area, G (e,f) is the gray value of the f-th pixel in the e-th second target area; The lock identification value is calculated based on the following formula: P=a1×P RGB +a2×P Gray Wherein, P is the lock identification value, a1 and a2 are preset weights, and a1 and a2 are both greater than 0, a1+a2=1; When the locking identification value corresponding to the preprocessed real-time image of the suitcase is consistent with the locking identification value corresponding to the preprocessed image of the suitcase to be compared, the suitcase processing link information corresponding to the suitcase image to be compared is obtained as the suitcase processing link information of the suitcase to be tracked, and the position information of the suitcase to be tracked is updated; For the sample images of any two sample suitcases, based on the difference between the grayscale value of the pixel of the candidate second target area of the sample image of one sample suitcase and the grayscale value of the pixel of the candidate second target area corresponding to the position of the sample image of the other sample suitcase, determine the texture feature similarity of the sample images of the two sample suitcases in the candidate second target area, then calculate the texture feature similarity fluctuation parameter corresponding to each candidate second target area, and select the candidate second target area whose texture feature similarity fluctuation parameter is greater than a preset texture feature similarity fluctuation parameter threshold as the second target area; The color feature similarity fluctuation parameter is calculated based on the following formula: Among them, V RGB is the color feature similarity fluctuation parameter, M is the total number of sample suitcases, is the color feature similarity between the m1th sample suitcase and the m2th sample suitcase in the candidate first target area; The texture feature similarity fluctuation parameter is calculated based on the following formula: Among them, V Gray is the texture feature similarity fluctuation parameter, is the color feature similarity between the n1th sample suitcase and the n2th sample suitcase in the candidate second target area.
2. A real-time luggage tracking system based on image recognition according to claim 1, characterized in that: Also includes: A status monitoring module, comprising a status monitoring component disposed on the luggage to be tracked; The tag retrieval module is used to determine the retrieval position of the RFID tag set on the luggage to be tracked based on the luggage status information collected by the status monitoring component set on the luggage to be tracked.
3. A real-time luggage tracking system based on image recognition according to claim 2, characterized in that: The tag retrieval module determines the retrieval position of the RFID tag set on the luggage to be tracked based on the luggage status information collected by the status monitoring component set on the luggage to be tracked, including: Acquire historical vibration information of the luggage collected by a status monitoring component disposed on a luggage to be tracked that is associated with the luggage to be tracked; Based on the historical vibration information of the luggage collected by the status monitoring components arranged on the multiple associated luggage cases to be tracked, the recovery position of the RFID tag arranged on the luggage case to be tracked is determined.
4. A real-time luggage tracking system based on image recognition according to claim 3, characterized in that: The tag retrieval module determines the retrieval position of the RFID tag set on the suitcase to be tracked based on the historical vibration information of the suitcase collected by the status monitoring component set on the suitcase to be tracked and the historical vibration information of the suitcase collected by the status monitoring component set on the suitcase to be tracked associated with the suitcase to be tracked, including: generating a vibration change curve based on historical vibration information of the luggage collected by a status monitoring component disposed on a luggage to be tracked associated with the luggage to be tracked; Performing empirical mode decomposition on the vibration change curve to determine the intrinsic modal components and residuals corresponding to the vibration change curve; The retrieval position of the RFID tag set on the suitcase to be tracked is determined by a position determination model based on the intrinsic modal components and residuals corresponding to the vibration change curve.
