Article component detection method and device, computer equipment and storage medium

By building an image database and computing components to connect weight matrix, identifying and defending disassembled contraband, the problem of not being able to automatically identify disassembled contraband in the prior art is solved, and the security and efficiency of the security check system are improved.

CN119007094APending Publication Date: 2024-11-22ZHONGKE HONGTUO (SUZHOU) INTELLIGENT TECH CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202410934077.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The prior art cannot automatically correlate and analyze and accurately identify dismantled contraband, especially dismantled contraband with specific mechanical structures, such as guns and police equipment, which makes it difficult to identify and defend during security checks.

Method used

By constructing an image database, recording the component connection relationship and weight matrix of contraband, obtaining real-time image information, calculating the disassembled synthetic value based on the object detection strategy and the component connection weight matrix, and issuing early warning information to identify and defend against disassembled contraband.

Benefits of technology

It realizes rapid and accurate detection and identification of disassembled contraband, improves the overall security of the security inspection system, overcomes the limitations of traditional security inspection relying on manual map judgment, and improves detection efficiency and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119007094A_ABST
    Figure CN119007094A_ABST
Patent Text Reader

Abstract

The invention relates to an article component detection method and device, computer equipment and a storage medium, and belongs to the technical field of artificial intelligence. Acquiring real-time image information; based on a target detection strategy and an image database, determining a target article component in the real-time image information and a confidence coefficient corresponding to the target article component; according to the confidence coefficient and the component connection weight matrix of the target article component, calculating a disassembly synthesis value of the target article component; when it is detected that the confidence coefficient is larger than a first preset value and / or the disassembling synthesis value is larger than a second preset value, early warning information is sent out. According to the invention, some contrabands which have specific mechanical structures and can be disassembled to successively pass through the security inspection equipment can be automatically, quickly and accurately detected, identified, analyzed and judged, so that all similar detachable contrabands can be effectively defended and early warned, and the overall security of the security inspection system is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an object component detection method, device, computer equipment and storage medium. Background Art

[0002] Nowadays, target detection algorithms and semantic segmentation technologies based on X-ray image input can detect and identify common fixed categories of contraband relatively well. However, despite the continuous advancement of technology, there are still many problems and challenges that have not been well resolved. One of them is the detection and identification of detachable contraband. Some contraband with specific mechanical structures and detachable, such as firearms and police equipment, may be manually disassembled and scattered in multiple different luggage packages to pass security checks one after another, and then reassembled. Since such contraband is disassembled and scattered, its X-ray image features may change significantly, making it difficult to identify. How to automatically associate and analyze, reconstruct matching and accurately identify these disassembled contraband is an urgent problem to be solved. Summary of the invention

[0003] Based on this, it is necessary to provide an object component detection method, device, computer equipment and storage medium that can improve detection and identification efficiency and accuracy in response to the above technical problems.

[0004] In one aspect, a method for detecting an item component is provided, the method comprising:

[0005] Constructing an image database, wherein the image database includes component connection weight matrices of multiple object components;

[0006] Obtain real-time image information;

[0007] Determine, based on the target detection strategy and the image database, a target object component in the real-time image information and its corresponding confidence level, wherein the target object component is a single component and / or a combination of components;

[0008] Calculating a disassembly synthesis value of the target item component according to the confidence and a component connection weight matrix of the target item component;

[0009] In response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, a warning message is issued.

[0010] Optionally, the method for constructing the image database includes:

[0011] Disassemble the components of the target object to obtain multiple single components and component combinations obtained by connecting multiple adjacent components, and record the component connection relationship;

[0012] Based on the single component and the combination of components, generating first image information under multiple viewing angles;

[0013] Marking the first image information based on the attribute of the target object to obtain second image information;

[0014] The second image information and the component connection relationship are saved in a database to generate the image database.

[0015] Optionally, based on the component connection relationship, constructing the component connection weight matrix of the target item includes:

[0016] The target object is disassembled into N components, and the N components are represented as P1, P2, P3, ..., P N ;

[0017] Define component P i and component P j The component connection weight between them is w ij , where i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N);

[0018] In response to detecting component P i and component P j When there is no connection relationship between them, w ij =0;

[0019] In response to detecting component P i and component P j When there is a connection relationship between them, w ij ∈(0,1];

[0020] And, component P i and component P j The component connection weight between them is w ij Need to meet:

[0021]

[0022] Based on multiple component connection weights, the component connection weight matrix of the target item is constructed as follows:

[0023]

[0024] Where W represents the component connection weight matrix.

[0025] Optionally, based on the target detection strategy and the image database, determining the target object component in the real-time image information and its corresponding confidence level includes:

[0026] Based on a component detection model, detecting whether there is a target object component in the real-time image information, wherein the component detection model is generated by training image information in an image database;

[0027] and / or, matching the real-time image information with the image information in the image database based on an image feature matching mechanism to determine whether a target object component exists in the real-time image information;

[0028] In response to detecting the presence of a target item component in the real-time image information, the confidence level of the target item component is determined based on the component detection model and / or the image feature matching mechanism.

[0029] Optionally, calculating the disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component includes:

[0030] Based on the confidence, determining a confidence vector of the target item component;

[0031] Calculating a disassembly synthesis value of the target item component based on the confidence vector and a component connection weight matrix of the target item component includes:

[0032]

[0033] Among them, S represents the disassembly synthesis value, C represents the confidence vector, and c i 、c j =c1,c2,c3,…,c N , c i 、c j ∈[0,1],c i 、c j Both represent elements in the confidence vector, w ij Represents the component connection weight, W represents the component connection weight matrix, N represents the number of components, i and j represent the i-th component and the j-th component respectively, i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N).

[0034] Optionally, based on the confidence, determining the confidence vector of the target item component includes:

[0035] Construct a confidence vector, including:

[0036] C=[c1 c2 c3…c N ]

[0037] In response to detecting the confidence d∈(0,1] of the target item component, assigning values ​​to elements in the confidence vector according to the number of components of the target item component;

[0038] The confidence vector after the assignment is defined as the confidence vector of the target item component.

[0039] Optionally, after calculating the disassembly synthesis value of the target item component, the method further includes:

[0040] Saving the confidence vector of the target item component;

[0041] In response to detecting a component of an item corresponding to the target item component again within a preset time period, if the component is the same as the target item component, selecting the confidence of the component with a higher confidence to determine the element assignment in the confidence vector, and recalculating the disassembly synthesis value based on the confidence vector after adjusting the element assignment;

[0042] If the component is not the same as the target item component, the element assignment in the confidence vector is determined based on the confidence of the component, the confidence vector is adjusted based on the element assignment, and the disassembly synthesis value is recalculated using the confidence vector.

[0043] In another aspect, an article component detection device is provided, the device comprising:

[0044] A database construction module, used to construct an image database, wherein the image database includes component connection weight matrices of multiple object components;

[0045] An information acquisition module, used for acquiring real-time image information;

[0046] A determination module, configured to determine a target object component in the real-time image information and its corresponding confidence level based on a target detection strategy and the image database, wherein the target object component is a single component and / or a combination of components;

[0047] A calculation module, configured to calculate a disassembly synthesis value of the target item component according to the confidence and a component connection weight matrix of the target item component;

[0048] The early warning module is used to issue an early warning message when it is detected that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value.

[0049] In another aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:

[0050] Constructing an image database, wherein the image database includes component connection weight matrices of multiple object components;

[0051] Obtain real-time image information;

[0052] Determine, based on the target detection strategy and the image database, a target object component in the real-time image information and its corresponding confidence level, wherein the target object component is a single component and / or a combination of components;

[0053] Calculating a disassembly synthesis value of the target item component according to the confidence and a component connection weight matrix of the target item component;

[0054] In response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, a warning message is issued.

[0055] In another aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0056] Constructing an image database, wherein the image database includes component connection weight matrices of multiple object components;

[0057] Obtain real-time image information;

[0058] Determine, based on the target detection strategy and the image database, a target object component in the real-time image information and its corresponding confidence level, wherein the target object component is a single component and / or a combination of components;

[0059] Calculating a disassembly synthesis value of the target item component according to the confidence and a component connection weight matrix of the target item component;

[0060] In response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, a warning message is issued.

[0061] The above-mentioned object component detection method, device, computer equipment and storage medium, the method includes: constructing an image database, the image database includes a component connection weight matrix of multiple object components; obtaining real-time image information; based on the target detection strategy and the image database, determining the target object component and its corresponding confidence in the real-time image information, the target object component is a single component and / or a component combination; according to the confidence and the component connection weight matrix of the target object component, calculating the disassembly synthesis value of the target object component; in response to detecting that the confidence is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, issuing a warning message, the present application can automatically perform rapid detection, identification and analysis and judgment on certain contraband items with a specific mechanical structure and that can be disassembled and passed through security inspection equipment one after another, so as to provide effective defense and warning for all similar detachable contraband items, thereby improving the overall security of the security inspection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A diagram showing an application environment of an object component detection method in an embodiment;

[0063] Figure 2 is a schematic diagram of a flow chart of an object component detection method in one embodiment;

[0064] Figure 3 is another schematic diagram of a flow chart of an object component detection method in one embodiment;

[0065] Figure 4 A diagram showing the connection relationship of gun and police equipment components in an object component detection method in one embodiment;

[0066] Figure 5 is a structural block diagram of an object component detection device in one embodiment;

[0067] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0069] It should be understood that in the description of the present application, unless the context clearly requires otherwise, words such as "include", "comprises", and the like throughout the specification should be interpreted as including rather than being exclusive or exhaustive; that is, as including but not limited to.

[0070] It should also be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, "plurality" means two or more.

[0071] It should be noted that the terms "S1", "S2", etc. are only used for the purpose of describing the steps, and do not specifically refer to the order or sequence, nor are they used to limit the present application. They are only for the convenience of describing the method of the present application, and cannot be understood as indicating the order of the steps. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present application.

[0072] Micro-dose X-ray security inspection equipment, commonly known as X-ray machine, is an X-ray device used for security inspection. It uses micro-dose X-ray radiation technology and can perform effective inspections while ensuring safety. This equipment is usually used in security inspections, logistics, medical imaging and other fields. It has the characteristics of low radiation dose, high safety, and clear imaging. The use of micro-dose X-ray security inspection equipment can quickly and accurately detect the internal structure of the target object, effectively identify contraband, and help improve the level of security precautions in the public domain. Generally speaking, the detection and identification of contraband requires security inspectors to use their naked eyes to judge the X-ray machine image. This process is called image judgment. Traditional security inspection image judgment is completely dependent on manual labor, and the recognition effect depends on the experience of the security inspector. The ability, energy, and attitude of the staff will affect the missed detection rate. Security inspectors generally work continuously for about 1 hour before changing shifts to rest. Frequent shift changes will also affect the overall security inspection efficiency. Check-in security inspectors judge every day. Figures 1-2 Tens of thousands of pieces of paper, with high intensity of work, high concentration of attention and high work pressure. Traditional security inspection is a labor-intensive industry with a large base of personnel, frequent turnover and few experienced check-in personnel. All these factors make the manual image judgment method in the traditional security inspection process unstable. If the prohibited items in the X-ray machine images can be automatically identified through computer technology, the security inspection process can be simplified, the security inspection efficiency can be improved, the labor cost and labor intensity can be reduced, and public safety can be effectively guaranteed. At the same time, it can also overcome the various defects of manual image judgment. According to the background technology, the existing technology cannot automatically associate and analyze, reconstruct the match and accurately identify these disassembled prohibited items.

[0073] In order to solve the above technical problems, the present application provides an object component detection method, device, equipment and storage medium, which can automatically and quickly detect, identify and analyze certain prohibited items that have a specific mechanical structure and can be disassembled to pass security inspections, so as to establish an effective defense warning for all similar disassembled prohibited items, and further improve the overall security of the security inspection system.

[0074] like Figure 1 FIG. 1 is a schematic diagram of a scene of an object component detection system provided in an embodiment of the present application. The object component detection system may include a computer device 100, in which an object component detection device is integrated. Figure 1 Computer equipment in.

[0075] In the embodiment of the present application, the computer device 100 is mainly used for: constructing an image database, which includes a component connection weight matrix of multiple item components; acquiring real-time image information; determining the target item component in the real-time image information and its corresponding confidence based on the target detection strategy and the image database, wherein the target item component is a single component and / or a component combination; calculating the disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component; and issuing a warning message in response to detecting that the confidence is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value.

[0076] In the embodiment of the present application, the computer device 100 may be an independent server, or a server network or server cluster composed of servers. For example, the computer device 100 described in the embodiment of the present application includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets or a cloud server composed of a plurality of servers. The cloud server is composed of a large number of computers or network servers based on cloud computing.

[0077] It is understandable that the computer device 100 in the embodiment of the present application may also be a terminal, and the terminal may be a device including both receiving and transmitting hardware, that is, a device having receiving and transmitting hardware capable of performing two-way communication on a two-way communication link, and such a device may include: a cellular or other communication device having a single-line display or a multi-line display or a cellular or other communication device without a multi-line display. The specific terminal may be a desktop terminal or a mobile terminal, and the terminal may also be one of a tablet computer, a laptop computer, etc.

[0078] In an embodiment of the present application, the computer device 100 may be a computer device that is communicatively connected to one or more X-ray machines (also known as security inspection machines), and the images of objects scanned by one or more X-ray machines may be input into the computer device 100 as images to be detected for image detection of detachable prohibited items.

[0079] Those skilled in the art will understand that Figure 1 The application environment shown in the figure is only one application scenario of the present application solution and does not constitute a limitation on the application scenario of the present application solution. Other application environments may also include Figure 1 More or less computer equipment as shown in Figure 1 Only one computer device is shown. It can be understood that the item component detection system can also include one or more other services, which are not specifically limited here.

[0080] In addition, if Figure 1As shown, the object component detection system may also include a memory 200 for storing data in the constructed image database, such as image information related to detachable contraband scanned by an X-ray machine and its corresponding component connection weight matrix.

[0081] It should be noted that Figure 1 The scenario diagram of the item component detection system shown is merely an example. The item component detection system and scenario described in the embodiment of the present application are intended to more clearly illustrate the technical solution of the embodiment of the present application, and do not constitute a limitation on the technical solution provided in the embodiment of the present application. A person of ordinary skill in the art can appreciate that with the evolution of the item component detection system and the emergence of new business scenarios, the technical solution provided in the embodiment of the present application is equally applicable to similar technical problems.

[0082] In one embodiment, Figure 2-Figure 3 As shown, it is a schematic diagram of an embodiment of the method for detecting an article component in the embodiment of the present application. The method for detecting an article component includes steps S1 to S5, which are specifically as follows:

[0083] S1: Construct an image database, wherein the image database includes component connection weight matrices of multiple object components.

[0084] It should be noted that item components refer to components of disassembled prohibited items. Disassembled prohibited items that meet the requirements usually have a certain specific mechanical structure and can be disassembled into several parts (i.e., components), and can be reassembled after disassembly without losing the original functions, such as firearms and police equipment; the image database is constructed based on the three-dimensional structural image information of one or more components of disassembled prohibited items and their component combinations. The image database needs to record the connection relationship between each component, and record all the main components of the corresponding disassembled prohibited items and their corresponding typical three-dimensional structural image information; the three-dimensional structural image information can be an X-ray scan image obtained by scanning the components and their component combinations with an X-ray machine; the main components include a single component and a combination of components connected to each other by multiple adjacent components, and some of the components or component combinations can be marked as core key components or component combinations; in addition, based on the connection relationship of each component, a component connection weight matrix corresponding to the disassembled prohibited items can be constructed.

[0085] In some specific embodiments, the method for constructing an image database includes:

[0086] Disassemble the components of the target item to obtain multiple single components and component combinations obtained by connecting multiple adjacent components, and record the component connection relationship, where a single component refers to a single component used to assemble the target item, such as magazines, bolts, sights and other parts in firearms and police equipment; a component combination refers to a component combination consisting of two or more parts, such as magazines and bolts, etc. An item may include multiple single components and multiple component combinations; a component connection relationship refers to whether two components have a connection relationship or not, such as a connection relationship between a bolt and a sight, or between a magazine and a bolt, etc.;

[0087] Based on the single component and the component combination, first image information under multiple viewing angles is generated, wherein the first image information is an image of the component or the component combination scanned by an X-ray machine in different placement postures, that is, the image database needs to record the image of each single component and component combination after being scanned by an X-ray machine, and each component and component combination needs to be scanned by an X-ray machine multiple times in different placement postures to include X-ray scan images of multiple different viewing angles;

[0088] The first image information is marked based on the attribute of the target item to obtain the second image information, wherein the attribute of the target item is the basic information of the item. For example, if the contraband is a firearm or police device, its attribute is the item type, such as a 97-type explosion-proof gun. If the image information is a bolt image of a firearm or police device, the mark may be the bolt of the 97-type explosion-proof gun, or a set proprietary number, such as 97-1, where 97 indicates the item type and 1 indicates the bolt. By marking the image, the recognition efficiency is improved.

[0089] The second image information and the component connection relationship are saved in a database to generate the image database, wherein the multiple image information corresponding to each object and the corresponding component connection relationship are clustered and saved to improve detection efficiency.

[0090] In some specific implementations, based on the component connection relationship, a component connection weight matrix of the target object is constructed, that is, according to the internal component interconnection relationship corresponding to each type of contraband recorded in the image database of the disassembled contraband, the component connection weight matrix is ​​constructed respectively, including:

[0091] Disassemble the target item into N components, that is, assume that a certain disassembled contraband can be disassembled into at most N main parts, and represent the N components as P1, P2, P3, ..., P N ;

[0092] Define component P i and component P j The component connection weight between them is w ij, where i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N);

[0093] In response to detecting component P i and component P j When there is no connection relationship between them, that is, component P i and component P j There is no interconnection between them. ij =0;

[0094] In response to detecting component P i and component P j When there is a connection relationship between components P i and component P j There is a mutual connection relationship between ij ∈(0,1];

[0095] And, component P i and component P j The component connection weight between them is w ij , that is, the overall connection weight of an item must satisfy:

[0096]

[0097] Among them, each non-zero component connection weight w ij The specific value of needs to be pre-set according to the criticality of the interconnection between the internal components of the disassembled contraband. If the criticality of the interconnection relationship is low, then w ij The value is close to 0. If the criticality of the interconnection relationship is high, then w ij The value is close to 1, and its value can be assigned by industry experts. For example, if the criticality of the connection between the sight and the bolt is high, the value can be 0.7, and if the criticality of the connection between the bolt and the buttstock is low, the value can be 0.1, and so on, which will not be repeated here.

[0098] Based on multiple component connection weights, the component connection weight matrix of the target item is constructed as follows:

[0099]

[0100] Where W represents the component connection weight matrix, which is an N×N square matrix with all elements on the diagonal and below zero;

[0101] Associate the component connection weight matrix of the target item with the corresponding item for quick subsequent call to calculate the disassembly synthesis value.

[0102] S2: Acquire real-time image information.

[0103] It should be noted that the real-time image information is the image information after being scanned by an X-ray machine in scenarios such as security inspection and logistics.

[0104] S3: Based on the target detection strategy and the image database, determine the target object component in the real-time image information and its corresponding confidence level, wherein the target object component is a single component and / or a combination of components.

[0105] It should be noted that the target item component is a suspected contraband component or a combination of components. The target detection strategy includes detecting items in real-time image information based on a pre-trained component detection model, detecting and locating disassembled contraband components, and calculating the image feature descriptors of all X-ray scanned image samples in the disassembled contraband image database by image feature matching, performing image feature matching in the X-ray scanned images, and matching and locating the disassembled contraband components that may appear therein.

[0106] In some specific embodiments, the training method of the component detection model includes: collecting and annotating a batch of data sets based on the physical objects corresponding to the main components and combinations of the disassembled prohibited items in the above-mentioned image database, where the data sets can be divided into X-ray scan images of ordinary passenger luggage containing disassembled prohibited items components and X-ray scan images of ordinary passenger luggage not containing disassembled prohibited items components, annotating the types of disassembled prohibited items components that appear and their positions in the X-ray scan images, and pre-training a component detection model, such as YOLO v8, that can be used to detect whether there are disassembled prohibited items components in the X-ray scan images of passenger luggage based on the data sets through algorithms such as back propagation in deep learning. The specific training method of the model is a common method and will not be repeated here.

[0107] In some specific embodiments, a component detection model or an image feature matching mechanism may be selected alone to determine whether there is a detachable contraband component in the real-time image information, or the two detection methods may be selected in combination to determine whether there is a detachable contraband component in the real-time image information. The specific selection rules may be determined according to the real-time image information acquisition scenario. For example, for scenarios with low detection accuracy requirements or low computing power, an image feature matching mechanism may be selected. For scenarios with high detection accuracy requirements or high computing power, a component detection model or a combination of the two detection methods may be selected. If a combination of the two detection methods is selected, and if there is a time requirement, that is, a lower latency is required, the component detection model alone may be selected for detection.

[0108] In some specific implementations, based on the target detection strategy and the image database, determining the target object component in the real-time image information and its corresponding confidence level includes:

[0109] Based on the component detection model, detect whether there is a target item component in the real-time image information, the component detection model is generated by training the image information in the image database, and the specific process has been described in the above steps and will not be repeated here;

[0110] and / or, matching the real-time image information with the image information in the image database based on an image feature matching mechanism to determine whether a target object component exists in the real-time image information;

[0111] In response to detecting the presence of a target item component in the real-time image information, the confidence level of the target item component is determined based on the component detection model and / or the image feature matching mechanism.

[0112] Specifically, the method for selecting a specific detection strategy has been described in the above steps and will not be repeated here. Among them, the detection confidence of the suspected contraband component can be directly obtained by using the component detection model, and the detection confidence is used as the confidence of the corresponding component or component combination. The confidence determination method using the image feature matching method is: calculate the proportion of the number of image feature descriptors that match the X-ray scan image sample of the disassembled contraband component in the passenger luggage X-ray scan image to the number of all image feature descriptors, and use the proportion score as the confidence of the corresponding component or component combination. When the confidence is determined by combining the component detection model and the image feature matching, the final confidence calculation formula is: y=ax+bz, wherein a and b are weights, which are set according to the detection accuracy or other standards. For example, if the accuracy requirement is high, a accounts for a large proportion and b accounts for a small proportion. If the delay requirement is high, b is 0. x and z are the confidence output by the component detection model and the confidence obtained based on the image feature matching method, respectively.

[0113] At the same time, the identity and itinerary information of the passenger whose luggage package contains suspected contraband components shall be recorded.

[0114] S4: Calculate a disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component.

[0115] It should be noted that the component connection weight matrix is ​​the product of the component connection weight matrix of all interconnected relationships of the disassembled contraband and its confidence vector.

[0116] In some specific implementations, determining the confidence vector of the target item component based on the confidence includes:

[0117] Construct a confidence vector, including:

[0118] C=[c1 c2 c3 … c N ]

[0119] Where C is a 1×N matrix vector, c i =c1,c2,c3,…,c N , c i ∈[0,1], is the main component P i The confidence level of , the initial value is 0, i∈(1,2,3…N);

[0120] In response to the confidence d∈(0,1] of the target item component being detected, the elements in the confidence vector are assigned values ​​according to the number of components of the target item component. For example, if a single component P is detected i The confidence level is d i ∈(0,1], then the element c in the confidence vector of the corresponding main component i Assign the detection match confidence d i , if the component combination P is detected i , P j The confidence level is d ij ∈(0,1], then the element c in the confidence vector of the main component i 、c j Assigned to the square root of the detection confidence If the component combination P is detected i , P j , P k The confidence level is d ijk ∈(0,1], then the element c in the confidence vector of the main component i 、c j 、c k The values ​​are respectively assigned as the cube root of the detection confidence If a component combination consisting of more parts is detected, the same procedure will be followed and no further description will be given here;

[0121] The confidence vector after definition assignment is the confidence vector of the target item component, wherein the number of elements in the confidence vector of an item is fixed, and the element values ​​in the original confidence vector are replaced according to the element assignment of the component or component combination determined by the detection. For example, if the element value of c1 is detected to be 0.3, then C = [0.3 0 0…0], and so on, which will not be repeated here.

[0122] Calculating the disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component includes:

[0123] Based on the confidence, determine the confidence vector of the target item component, wherein the confidence vector is calculated and determined according to the above steps, which will not be repeated here;

[0124] Based on the confidence vector and the component connection weight matrix of the target item component, the disassembly synthesis value of the target item component is calculated, wherein the component connection weight matrix of the target item component is a component connection weight matrix corresponding to the item directly called from the database, and the specific calculation formula includes:

[0125]

[0126] Among them, S represents the disassembly synthesis value, C represents the confidence vector, and c i 、c j =c1,c2,c3,…,c N , c i 、c j ∈[0,1],c i 、c j Both represent elements in the confidence vector, w ij Represents the component connection weight, W represents the component connection weight matrix, N represents the number of components, i and j represent the i-th component and the j-th component respectively, i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N).

[0127] Among them, S∈[0,1] is the sum of the confidence weights of all interconnected relationships corresponding to the disassembled contraband.

[0128] S5: In response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, issuing a warning message.

[0129] It should be noted that the first preset value and the second preset value can be set according to actual needs, such as the first preset value is 0.7, the second preset value is 0.8, etc.

[0130] In some specific implementations, if the confidence level of any core key component or component combination is detected to exceed a first preset value, the system issues a warning message and reports the identity and itinerary information of all passengers of the luggage packages containing suspected contraband components. At this time, the disassembly synthesis value may no longer be calculated;

[0131] Furthermore, if it does not exceed the first preset value, the disassembly synthesis value of the disassembled prohibited items is calculated. When the disassembly synthesis value S exceeds the second preset value, the system issues a warning message and reports the identity and itinerary information of all passengers whose luggage packages contain suspected prohibited items.

[0132] In some specific implementations, after calculating the disassembly synthesis value of the target item component, the method further includes:

[0133] Saving the confidence vector of the target item component;

[0134] In response to detecting a component of an item corresponding to the target item component again within a preset time period, if the component is the same as the target item component, selecting the confidence of the component with a higher confidence to determine the element assignment in the confidence vector, and recalculating the disassembly synthesis value based on the confidence vector after adjusting the element assignment;

[0135] If the component is not the same as the target item component, the element assignment in the confidence vector is determined based on the confidence of the component, the confidence vector is adjusted based on the element assignment, and the disassembly synthesis value is recalculated using the confidence vector.

[0136] Specifically, the preset time period can be set according to actual needs, and can be 0.5 days or 1 day, that is, the confidence vector of the target item component is saved for 0.5 days or 1 day, and all elements c in the confidence vector C are saved after the preset time period. i The purpose of resetting to zero is to save the different parts of the disassembled contraband placed in different luggage packages through the X-ray machine. There will be a certain interval between the time when they pass through the X-ray machine. If this happens, the disassembly synthesis value is recalculated based on the saved confidence vector. If a suspected contraband component is detected again after a period of time, if the suspected contraband component and the marked item component belong to the same item, determine whether the two components are of the same type. If so, determine the size of the confidence values ​​of the two components. If the confidence value of the current suspected contraband component is larger, then the above confidence vector is used to recalculate the disassembly synthesis value. The confidence is assigned by the above method, and the element value in the original confidence vector is replaced based on the assignment, and the disassembly synthesis value is recalculated. If the confidence value of the target item component is larger, there is no need to recalculate. If the two components are not of the same type, and the element value corresponding to this component in the original confidence vector is the initial value, the confidence is assigned by the above method, and the element value in the original confidence vector is replaced based on the assignment, and the disassembly synthesis value is recalculated. If the element value is not the initial value, the confidence size is compared first, and then it is determined whether to adjust it. It will not be repeated here.

[0137] In the above-mentioned object component detection method, the method includes: constructing an image database, wherein the image database includes a component connection weight matrix of multiple object components; obtaining real-time image information; determining the target object component in the real-time image information and its corresponding confidence based on the target detection strategy and the image database, wherein the target object component is a single component and / or a component combination; calculating the disassembly synthesis value of the target object component according to the confidence and the component connection weight matrix of the target object component; in response to detecting that the confidence is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, issuing an early warning message. The method for detecting and identifying disassembled contraband proposed in the present application can overcome the limitations of the existing contraband detection method for a single target X-ray machine to a certain extent, and realize the detection and identification of disassembled contraband within a time window; the present application not only considers a single main component, but also considers the situation where multiple components may be combined, which helps to establish a more three-dimensional early warning defense; the present application can not only be used for the detection and identification of firearms and police equipment, but can also be extended to other disassembled contraband of the same type, with a wide range of application scenarios, based on which the overall security of the security inspection system is improved.

[0138] It should be understood that although Figure 2-Figure 3 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-Figure 3 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0139] In one embodiment, in order to verify and illustrate the technical solution adopted in this application, this embodiment adopts the following Figure 4 The firearms and police equipment shown are used as examples for verification, and the real effect of this application is verified by scientific argumentation.

[0140] Specifically, Figure 4 As shown, the typical structure of a firearm police device can be broken down into the following main parts:

[0141] (1) The bolt, including the barrel, bolt housing, bolt safety pin, etc. The bolt is the main body of the gun and the core part of the sub-component assembly;

[0142] (2) The buttstock, which supports the arm and provides better stability and grip;

[0143] (3) Magazines, which are used to store ammunition and can be replaced individually;

[0144] (4) Sights, including mechanical sights and optical sights, used to assist shooting;

[0145] (5) Gun catch, including trigger, trigger spring, barrel release button, etc., used to control shooting;

[0146] By combining these parts, various types of guns can be assembled, such as pistols, rifles, shotguns, etc.

[0147] Step 1: Construct the interconnection relationship between the main components of firearms and police equipment and the X-ray scanning image database;

[0148] Among them, the database is for a variety of firearms and police equipment represented by pistols. Firearms and police equipment that meet the requirements usually have a certain specific mechanical mechanism, can be disassembled into several parts, and can be reassembled after disassembly without losing the original function; the database needs to record the typical three-dimensional structure of the corresponding firearms and police equipment, including the connection relationship between the various parts, and record all the main parts of the corresponding firearms and police equipment, including single parts (such as bolts, buttstocks, magazines, sights, gun buckles, etc.) and several component combinations connected by multiple adjacent parts (such as bolts-buttstocks, bolts-magazines, bolts-sights-gun buckles, etc.), and at the same time, some of the parts or component combinations can be marked as core key parts or component combinations (such as bolts, etc.); the database needs to record the image of each single part and component combination after being scanned by an X-ray machine, and each part and component combination needs to be scanned by an X-ray machine multiple times in different placement postures to include X-ray scan images of multiple different perspectives;

[0149] Step 2: Determine the target object components and their corresponding confidences in the real-time image information based on the target detection strategy and the image database;

[0150] Using the target detection model: Based on the physical objects corresponding to the main parts and components of firearms and police equipment in the aforementioned database, a batch of data sets are collected and annotated. The data set is divided into 8,000 X-ray scan images of ordinary passenger luggage containing firearms and police equipment components and 5,000 X-ray scan images of ordinary passenger luggage without firearms and police equipment components. The types of firearms and police equipment components that appear and their locations in the X-ray scan images are annotated. Based on the data set, a target detection model YOLO v8 that can be used to detect whether there are firearms and police equipment components in passenger luggage X-ray scan images is pre-trained through algorithms such as back propagation in deep learning, and the firearms and police equipment components are detected and located;

[0151] Adopting the image feature matching method: directly calculating the SURF (Speeded Up Robust Features) image feature descriptors of all X-ray scanned image samples in the database of main components of firearms and police equipment, and performing image feature matching in the X-ray scanned images of passenger luggage, and matching and locating the possible firearms and police equipment components in them;

[0152] By adopting the target detection model, the detection confidence of the suspected firearm and police equipment component detected is used as the confidence of the corresponding component or component combination, and / or, by adopting the image feature matching method, the proportion of the number of SURF image feature descriptors that match the X-ray scanned image samples of the firearm and police equipment component in the X-ray scanned image of the passenger's luggage to the number of all SURF image feature descriptors is calculated as the confidence of the corresponding component or component combination, and at the same time, the identity and itinerary information of the passenger to whom the luggage package containing the suspected firearm and police equipment component belongs is recorded;

[0153] Step 3: Calculate the disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component;

[0154] A component connection weight matrix corresponding to the firearms and police equipment is extracted from the database. The method for generating the component connection weight matrix includes: constructing component connection weight matrices respectively according to the internal component interconnection relationship corresponding to each firearm and police equipment recorded in the main component database of the firearms and police equipment:

[0155] Assume that a certain type of police gun can be divided into at most five main parts, represented by P1 (bolt), P2 (stock), P3 (magazine), P4 (sight), and P5 (gun buckle). i With P j The component connection weight between them is w ij , where i∈(1,2,3,4) and j∈(i+1,i+2,i+3…5);

[0156] If P i With P j There is no interconnection between them, then w ij =0,

[0157] If P i With P j If there is a mutual connection between ij ∈(0,1],

[0158] Every non-zero w ijThe specific value of needs to be pre-set according to the criticality of the interconnection between the internal components of the disassembled contraband. If the criticality of the interconnection relationship is low, w ij The value is close to 0. If the criticality of the interconnection relationship is high, w ij The value is close to 1, and the overall condition needs to meet the following requirements:

[0159]

[0160] Then the component connection weight matrix corresponding to the gun police equipment is:

[0161]

[0162] Where W is a 5×5 square matrix, and all elements on the diagonal and below are zero;

[0163] The confidence vector of the main components of the gun police equipment is:

[0164] C=[c1 c2 c3 c4 c5]

[0165] Where C is a 1×5 matrix vector, where c i ∈[0,1], corresponding to the main component P i The confidence level of , the initial value is 0, i∈(1,2,3,4,5);

[0166] According to the results of the X-ray scanning image detection of the gun and police equipment components, the confidence vector of the main components is updated in real time. i The confidence level is d i ∈(0,1], then the element c in the confidence vector of the corresponding main component i Assign the detection match confidence d i (For example, if the confidence of a single component P1 gun is 0.7, the element c1 in the corresponding main component confidence vector is assigned the detection matching confidence of 0.7). i , P j The confidence level is d ij ∈(0,1], then the element c in the confidence vector of the main component i 、c j Assign the values ​​to the square root of the detection confidence respectively (For example, if the confidence level of the detected component combination P1 bolt-P3 magazine is 0.64, the elements c1 and c3 in the confidence vector of the main components are assigned the square root of the detection confidence level 0.8 respectively). i , P j , P k The confidence level is d ijk∈(0,1], then the element c in the confidence vector of the main component i 、c j 、c k Assign the values ​​to the cube root of the detection confidence respectively (For example, if the confidence level of the detected component combination P1 bolt-P4 sight-P5 gun button is 0.729, the elements c1, c4, and c5 in the confidence vector of the main components are assigned the cube root of the detection confidence level 0.9 respectively. If a component combination consisting of more components is detected, the same is applied and will not be described in detail here.

[0167] Based on the above content, the disassembly synthesis value of the firearm police equipment is calculated:

[0168]

[0169] We get S∈[0,1], which is the sum of the confidence weights of all interconnected relationships of the guns and police equipment.

[0170] Step 4: in response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, issuing a warning message;

[0171] A warning threshold of 0.7 and 0.8 is set in advance. If the calculated disassembly composite value S of the firearm is greater than 0.8, or the confidence of any core key component or component combination is detected to be greater than 0.7, the system will issue a warning message and report the identity and itinerary information of all passengers whose luggage packages contain suspected firearm components. In addition, every 12 hours, all elements c in the confidence vector C of the main components will be i Zero reset.

[0172] This example verifies the effectiveness and feasibility of the solution of the present application through relevant real data.

[0173] In one embodiment, Figure 5 As shown, an article component detection device is provided, comprising: a database construction module, an information acquisition module, a determination module, a calculation module and an early warning module, wherein:

[0174] A database construction module, used to construct an image database, wherein the image database includes component connection weight matrices of multiple object components;

[0175] An information acquisition module, used for acquiring real-time image information;

[0176] A determination module, configured to determine a target object component in the real-time image information and its corresponding confidence level based on a target detection strategy and the image database, wherein the target object component is a single component and / or a combination of components;

[0177] A calculation module, configured to calculate a disassembly synthesis value of the target item component according to the confidence and a component connection weight matrix of the target item component;

[0178] The early warning module is used to issue an early warning message when it is detected that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value.

[0179] As a preferred implementation, in the embodiment of the present invention, the database construction module is specifically used for:

[0180] Disassemble the components of the target object to obtain multiple single components and component combinations obtained by connecting multiple adjacent components, and record the component connection relationship;

[0181] Based on the single component and the combination of components, generating first image information under multiple viewing angles;

[0182] Marking the first image information based on the attribute of the target object to obtain second image information;

[0183] The second image information and the component connection relationship are saved in a database to generate the image database.

[0184] As a preferred implementation, in the embodiment of the present invention, the database construction module is further used for:

[0185] The target object is disassembled into N components, and the N components are represented as P1, P2, P3, ..., P N ;

[0186] Define component P i and component P j The component connection weight between them is w ij , where i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N);

[0187] In response to detecting component P i and component P j When there is no connection relationship between them, w ij =0;

[0188] In response to detecting component P i and component P j When there is a connection relationship between them, w ij ∈(0,1];

[0189] And, component P i and component P j The component connection weight between them is w ij Need to meet:

[0190]

[0191] Based on multiple component connection weights, the component connection weight matrix of the target item is constructed as follows:

[0192]

[0193] Where W represents the component connection weight matrix.

[0194] As a preferred implementation, in the embodiment of the present invention, the determination module is specifically used for:

[0195] Based on a component detection model, detecting whether there is a target object component in the real-time image information, wherein the component detection model is generated by training image information in an image database;

[0196] and / or, matching the real-time image information with the image information in the image database based on an image feature matching mechanism to determine whether a target object component exists in the real-time image information;

[0197] In response to detecting the presence of a target item component in the real-time image information, the confidence level of the target item component is determined based on the component detection model and / or the image feature matching mechanism.

[0198] As a preferred implementation, in the embodiment of the present invention, the calculation module is specifically used for:

[0199] Based on the confidence, determining a confidence vector of the target item component;

[0200] Calculating a disassembly synthesis value of the target item component based on the confidence vector and a component connection weight matrix of the target item component includes:

[0201]

[0202] Among them, S represents the disassembly synthesis value, C represents the confidence vector, and c i 、c j =c1,c2,c3,…,c N , c i 、c j ∈[0,1],c i 、c j Both represent elements in the confidence vector, w ij Represents the component connection weight, W represents the component connection weight matrix, N represents the number of components, i and j represent the i-th component and the j-th component respectively, i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N).

[0203] As a preferred implementation, in the embodiment of the present invention, the calculation module is further used for:

[0204] Construct a confidence vector, including:

[0205] C=[c1 c2 c3 … c N ]

[0206] In response to detecting the confidence d∈(0,1] of the target item component, assigning values ​​to elements in the confidence vector according to the number of components of the target item component;

[0207] The confidence vector after the assignment is defined as the confidence vector of the target item component.

[0208] As a preferred implementation, in the embodiment of the present invention, the device further includes an adjustment module, and the adjustment module is specifically used to:

[0209] Saving the confidence vector of the target item component;

[0210] In response to detecting a component of an item corresponding to the target item component again within a preset time period, if the component is the same as the target item component, selecting the confidence of the component with a higher confidence to determine the element assignment in the confidence vector, and recalculating the disassembly synthesis value based on the confidence vector after adjusting the element assignment;

[0211] If the component is not the same as the target item component, the element assignment in the confidence vector is determined based on the confidence of the component, the confidence vector is adjusted based on the element assignment, and the disassembly synthesis value is recalculated using the confidence vector.

[0212] For the specific definition of the article component detection device, please refer to the definition of the article component detection method above, which will not be repeated here. Each module in the above-mentioned article component detection device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0213] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for detecting an item component is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0214] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0215] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0216] S1: constructing an image database, wherein the image database includes component connection weight matrices of multiple object components;

[0217] S2: Acquire real-time image information;

[0218] S3: Determine the target object component in the real-time image information and its corresponding confidence level based on the target detection strategy and the image database, wherein the target object component is a single component and / or a combination of components;

[0219] S4: calculating a disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component;

[0220] S5: In response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, issuing a warning message.

[0221] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0222] Disassemble the components of the target object to obtain multiple single components and component combinations obtained by connecting multiple adjacent components, and record the component connection relationship;

[0223] Based on the single component and the combination of components, generating first image information under multiple viewing angles;

[0224] Marking the first image information based on the attribute of the target object to obtain second image information;

[0225] The second image information and the component connection relationship are saved in a database to generate the image database.

[0226] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0227] The target object is disassembled into N components, and the N components are represented as P1, P2, P3, ..., P N ;

[0228] Define component P i and component P j The component connection weight between them is w ij , where i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N);

[0229] In response to detecting component P i and component P j When there is no connection relationship between them, w ij =0;

[0230] In response to detecting component P i and component P j When there is a connection relationship between them, w ij ∈(0,1];

[0231] And, component P i and component P j The component connection weight between them is w ij Need to meet:

[0232]

[0233] Based on multiple component connection weights, the component connection weight matrix of the target item is constructed as follows:

[0234]

[0235] Where W represents the component connection weight matrix.

[0236] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0237] Based on a component detection model, detecting whether there is a target object component in the real-time image information, wherein the component detection model is generated by training image information in an image database;

[0238] and / or, matching the real-time image information with the image information in the image database based on an image feature matching mechanism to determine whether a target object component exists in the real-time image information;

[0239] In response to detecting the presence of a target item component in the real-time image information, the confidence level of the target item component is determined based on the component detection model and / or the image feature matching mechanism.

[0240] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0241] Based on the confidence, determining a confidence vector of the target item component;

[0242] Calculating a disassembly synthesis value of the target item component based on the confidence vector and a component connection weight matrix of the target item component includes:

[0243]

[0244] Among them, S represents the disassembly synthesis value, C represents the confidence vector, and c i 、c j =c1,c2,c3,…,c N , c i 、c j ∈[0,1],c i 、c j Both represent elements in the confidence vector, w ij Represents the component connection weight, W represents the component connection weight matrix, N represents the number of components, i and j represent the i-th component and the j-th component respectively, i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N).

[0245] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0246] Construct a confidence vector, including:

[0247] C=[c1 c2 c3…c N ]

[0248] In response to detecting the confidence d∈(0,1] of the target item component, assigning values ​​to elements in the confidence vector according to the number of components of the target item component;

[0249] The confidence vector after the assignment is defined as the confidence vector of the target item component.

[0250] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0251] Saving the confidence vector of the target item component;

[0252] In response to detecting a component of an item corresponding to the target item component again within a preset time period, if the component is the same as the target item component, selecting the confidence of the component with a higher confidence to determine the element assignment in the confidence vector, and recalculating the disassembly synthesis value based on the confidence vector after adjusting the element assignment;

[0253] If the component is not the same as the target item component, the element assignment in the confidence vector is determined based on the confidence of the component, the confidence vector is adjusted based on the element assignment, and the disassembly synthesis value is recalculated using the confidence vector.

[0254] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0255] S1: constructing an image database, wherein the image database includes component connection weight matrices of multiple object components;

[0256] S2: Acquire real-time image information;

[0257] S3: Determine the target object component in the real-time image information and its corresponding confidence level based on the target detection strategy and the image database, wherein the target object component is a single component and / or a combination of components;

[0258] S4: calculating a disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component;

[0259] S5: In response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, issuing a warning message.

[0260] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0261] Disassemble the components of the target object to obtain multiple single components and component combinations obtained by connecting multiple adjacent components, and record the component connection relationship;

[0262] Based on the single component and the combination of components, generating first image information under multiple viewing angles;

[0263] Marking the first image information based on the attribute of the target object to obtain second image information;

[0264] The second image information and the component connection relationship are saved in a database to generate the image database.

[0265] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0266] The target object is disassembled into N components, and the N components are represented as P1, P2, P3, ..., P N ;

[0267] Define component P i and component P j The component connection weight between them is w ij , where i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N);

[0268] In response to detecting component P i and component P j When there is no connection relationship between them, w ij =0;

[0269] In response to detecting component P i and component P j When there is a connection relationship between them, w ij ∈(0,1];

[0270] And, component P i and component P j The component connection weight between them is w ij Need to meet:

[0271]

[0272] Based on multiple component connection weights, the component connection weight matrix of the target item is constructed as follows:

[0273]

[0274] Where W represents the component connection weight matrix.

[0275] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0276] Based on a component detection model, detecting whether there is a target object component in the real-time image information, wherein the component detection model is generated by training image information in an image database;

[0277] and / or, matching the real-time image information with the image information in the image database based on an image feature matching mechanism to determine whether a target object component exists in the real-time image information;

[0278] In response to detecting the presence of a target item component in the real-time image information, the confidence level of the target item component is determined based on the component detection model and / or the image feature matching mechanism.

[0279] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0280] Based on the confidence, determining a confidence vector of the target item component;

[0281] Calculating a disassembly synthesis value of the target item component based on the confidence vector and a component connection weight matrix of the target item component includes:

[0282]

[0283] Among them, S represents the disassembly synthesis value, C represents the confidence vector, and c i 、c j =c1,c2,c3,…,c N , c i 、c j ∈[0,1],c i 、c j Both represent elements in the confidence vector, w ij Represents the component connection weight, W represents the component connection weight matrix, N represents the number of components, i and j represent the i-th component and the j-th component respectively, i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N).

[0284] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0285] Construct a confidence vector, including:

[0286] C=[c1 c2 c3 … c N ]

[0287] In response to detecting the confidence d∈(0,1] of the target item component, assigning values ​​to elements in the confidence vector according to the number of components of the target item component;

[0288] The confidence vector after the assignment is defined as the confidence vector of the target item component.

[0289] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0290] Saving the confidence vector of the target item component;

[0291] In response to detecting a component of an item corresponding to the target item component again within a preset time period, if the component is the same as the target item component, selecting the confidence of the component with a higher confidence to determine the element assignment in the confidence vector, and recalculating the disassembly synthesis value based on the confidence vector after adjusting the element assignment;

[0292] If the component is not the same as the target item component, the element assignment in the confidence vector is determined based on the confidence of the component, the confidence vector is adjusted based on the element assignment, and the disassembly synthesis value is recalculated using the confidence vector.

[0293] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0294] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0295] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A method for detecting an object component, characterized in that: The method comprises: Constructing an image database, wherein the image database includes component connection weight matrices of multiple object components; Obtain real-time image information; Determine, based on the target detection strategy and the image database, a target object component in the real-time image information and its corresponding confidence level, wherein the target object component is a single component and / or a combination of components; Calculating a disassembly synthesis value of the target item component according to the confidence and a component connection weight matrix of the target item component; In response to detecting that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value, a warning message is issued.

2. The object component detection method according to claim 1, characterized in that: The method for constructing the image database comprises: Disassemble the components of the target object to obtain multiple single components and component combinations obtained by connecting multiple adjacent components, and record the component connection relationship; Based on the single component and the combination of components, generating first image information under multiple viewing angles; Marking the first image information based on the attribute of the target object to obtain second image information; The second image information and the component connection relationship are saved in a database to generate the image database.

3. The object component detection method according to claim 2, characterized in that: Based on the component connection relationship, the component connection weight matrix of the target item is constructed including: The target object is disassembled into N components, and the N components are represented as P1, P2, P3, ..., P N ; Define component P i and component P j The component connection weight between them is w ij , where i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N); In response to detecting component P i and component P j When there is no connection relationship between them, w ij =0; In response to detecting component P i and component P j When there is a connection relationship between them, w ij ∈(0,1]; And, component P i and component P j The component connection weight between them is w ij Need to meet: Based on multiple component connection weights, the component connection weight matrix of the target item is constructed as follows: Where W represents the component connection weight matrix.

4. The object component detection method according to claim 1, characterized in that: Based on the target detection strategy and the image database, determining the target object component in the real-time image information and its corresponding confidence level includes: Based on a component detection model, detecting whether there is a target object component in the real-time image information, wherein the component detection model is generated by training image information in an image database; and / or, matching the real-time image information with the image information in the image database based on an image feature matching mechanism to determine whether a target object component exists in the real-time image information; In response to detecting the presence of a target item component in the real-time image information, the confidence level of the target item component is determined based on the component detection model and / or the image feature matching mechanism.

5. The object component detection method according to claim 4, characterized in that: Calculating the disassembly synthesis value of the target item component according to the confidence and the component connection weight matrix of the target item component includes: Based on the confidence, determining a confidence vector of the target item component; Calculating a disassembly synthesis value of the target item component based on the confidence vector and a component connection weight matrix of the target item component includes: Among them, S represents the disassembly synthesis value, C represents the confidence vector, and c i 、c j =c1,c2,c3,…,c N , c i 、c j ∈[0,1],c i 、c j Both represent elements in the confidence vector, w ij Represents the component connection weight, W represents the component connection weight matrix, N represents the number of components, i and j represent the i-th component and the j-th component respectively, i∈(1,2,3…N-1) and j∈(i+1,i+2,i+3…N).

6. The object component detection method according to claim 5, characterized in that: Based on the confidence, determining a confidence vector of the target item component includes: Construct a confidence vector, including: C=[c1 c2 c3…c N ] In response to detecting the confidence d∈(0,1] of the target item component, assigning values ​​to elements in the confidence vector according to the number of components of the target item component; The confidence vector after the assignment is defined as the confidence vector of the target item component.

7. The object component detection method according to claim 6, characterized in that: After calculating the disassembly synthesis value of the target item component, the method further includes: Saving the confidence vector of the target item component; In response to detecting a component of an item corresponding to the target item component again within a preset time period, if the component is the same as the target item component, selecting the confidence of the component with a higher confidence to determine the element assignment in the confidence vector, and recalculating the disassembly synthesis value based on the confidence vector after adjusting the element assignment; If the component is not the same as the target item component, the element assignment in the confidence vector is determined based on the confidence of the component, the confidence vector is adjusted based on the element assignment, and the disassembly synthesis value is recalculated using the confidence vector.

8. An object component detection device, characterized in that: The device comprises: A database construction module, used to construct an image database, wherein the image database includes component connection weight matrices of multiple object components; An information acquisition module, used for acquiring real-time image information; A determination module, configured to determine a target object component in the real-time image information and its corresponding confidence level based on a target detection strategy and the image database, wherein the target object component is a single component and / or a combination of components; A calculation module, configured to calculate a disassembly synthesis value of the target item component according to the confidence and a component connection weight matrix of the target item component; The early warning module is used to issue an early warning message when it is detected that the confidence level is greater than a first preset value and / or the disassembly synthesis value is greater than a second preset value.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

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

Citation Information

Patent Citations

  • Contraband detection method and device

    CN110261923A

  • Complex mechanical product assembly modeling method based on complex network

    CN111310284A

  • Dangerous article identification method and device, electronic equipment and storage medium

    CN114037939A

  • Multi-part intelligent assembling method based on neural network

    CN116562804A

  • Method of registration in a contraband detection system

    US20070172129A1