Security check image discrimination method, device, system, equipment and medium

Abstract

The invention provides a security check image discrimination method. The method comprises the following steps: selecting a target working mode; if the target working mode is selected as a first mode, a second judgment result of the security check image is obtained after a first judgment result of the security check image is obtained, the first judgment result is a graph judgment result obtained based on manual graph judgment operation, and the second judgment result is a graph judgment result obtained based on image analysis and an identification algorithm; if the target working mode is selected as a second mode, acquiring the first judgment result and the second judgment result in real time; a first judgment conclusion is generated based on the first judgment result and the second judgment result, the first mode comprises a first sub-mode and a second sub-mode, and the first judgment result and the second judgment result of the first sub-mode comprise whether the security check image is suspected or not; the first judgment result and the second judgment result of the second sub-mode comprise visual marking of the suspected area of the security check image.

Description

Technical Field

[0001] The present disclosure relates to the field of security technologies, and in particular, to a method, device, system, equipment and medium for discriminating security inspection images. Background Art

[0002] In modern security inspection processes, image discrimination technology plays a crucial role. With the increasing strictness of security inspection requirements, there is a greater need to quickly and accurately judge security inspection images. However, existing security inspection image discrimination methods usually adopt fixed image judgment processes and technologies, such as manual operations or automated methods based on image recognition, and cannot flexibly adapt according to the characteristics of the images, the security inspection environment or the operation mode, etc., which affects the overall efficiency and accuracy of the security inspection system. Summary of the Invention

[0003] To solve the above technical problems, embodiments of the present disclosure provide a method, device, system, equipment and medium for discriminating security inspection images, which can switch different working modes during security inspection work, so as to meet the requirements in different scenarios.

[0004] The first aspect of the present disclosure provides a method for discriminating security inspection images, including: selecting a target working mode; if the target working mode is selected as the first mode, obtaining a second discrimination result of the security inspection image after obtaining a first discrimination result of the security inspection image, where the first discrimination result is a discrimination result obtained based on manual image judgment operations, and the second discrimination result is a discrimination result obtained based on image analysis and recognition algorithms; if the target working mode is selected as the second mode, obtaining the first discrimination result and the second discrimination result in real time; and generating a first discrimination conclusion based on the first discrimination result and the second discrimination result, where the first mode includes a first sub-mode and a second sub-mode, where the first discrimination result and the second discrimination result of the first sub-mode include that the security inspection image is suspected or the security inspection image is not suspected; the first discrimination result and the second discrimination result of the second sub-mode include visual marking of the suspected area of the security inspection image.

[0005] According to an embodiment of the present disclosure, if the target working mode is selected as the first sub-mode, a first discrimination conclusion is generated based on the first discrimination result and the second discrimination result, which specifically includes: when the first discrimination result of the first sub-mode is that the security inspection image is suspicious, the first discrimination conclusion is to manually inspect the target object in the security inspection image; when the first discrimination result of the first sub-mode is that the security inspection image is not suspicious and the second discrimination result of the first sub-mode is that the security inspection image is not suspicious, the first discrimination conclusion is to release the target object in the security inspection image; and when the first discrimination result of the first sub-mode is that the security inspection image is not suspicious and the second discrimination result of the first sub-mode is that the security inspection image is suspicious, the first discrimination conclusion is to perform a second image discrimination on the security inspection image to obtain a second discrimination conclusion.

[0006] According to an embodiment of the present disclosure, if the target working mode is selected as the second sub-mode, a first discrimination conclusion is generated based on the first discrimination result and the second discrimination result, which specifically includes: when the second discrimination result of the second sub-mode does not have a visual mark for the suspicious area of the security inspection image, the first discrimination conclusion is generated based on the first discrimination result of the second sub-mode; when the first discrimination result of the second sub-mode does not have a visual mark for the suspicious area of the security inspection image and the second discrimination result of the second sub-mode has a visual mark for the suspicious area of the security inspection image, the first discrimination conclusion is to perform a second image discrimination on the security inspection image to obtain a second discrimination conclusion; and when both the first discrimination result and the second discrimination result of the second sub-mode have visual marks for the suspicious area of the security inspection image, the first discrimination conclusion is to manually inspect the target object in the security inspection image.

[0007] According to an embodiment of the present disclosure, when both the first discrimination result and the second discrimination result of the second sub-mode have visual marks for the suspicious area of the security inspection image, the method further includes: if the areas with visual marks in the first discrimination result and the second discrimination result of the second sub-mode are the same, the second discrimination result of the second sub-mode is not displayed; and if the areas with visual marks in the first discrimination result and the second discrimination result of the second sub-mode are different, the second discrimination result of the second sub-mode is displayed.

[0008] According to an embodiment of the present disclosure, when the second discrimination result of the second sub-mode does not have a visual mark for the suspected area of the security inspection image, generating a first discrimination conclusion based on the first discrimination result of the second sub-mode specifically includes: when the first discrimination result of the second sub-mode does not have a visual mark for the suspected area of the security inspection image, the first discrimination conclusion is to release the target object in the security inspection image; and when the first discrimination result of the second sub-mode has a visual mark for the suspected area of the security inspection image, the first discrimination conclusion is to manually inspect the target object in the security inspection image.

[0009] According to an embodiment of the present disclosure, the image analysis and recognition algorithm includes a sensitivity parameter that can be dynamically adjusted, and the sensitivity parameter is used to adjust the recognition sensitivity of the image analysis and recognition algorithm to abnormal objects in the security inspection image.

[0010] According to an embodiment of the present disclosure, based on the credit level of the target object in the security inspection image, the threshold of the sensitivity parameter is adjusted.

[0011] According to an embodiment of the present disclosure, the method further includes: obtaining multi-dimensional information, where the multi-dimensional information is used to provide associated information of the target object in the security inspection image; and based on the multi-dimensional information, generating an auxiliary discrimination result, where the auxiliary discrimination result is used to correct or verify the second discrimination result.

[0012] A second aspect of the present disclosure provides a discrimination device for security inspection images, including: a mode selection module for selecting a target working mode; a dual-image discrimination module for, if the target working mode is selected as the first mode, obtaining a second discrimination result of the security inspection image after obtaining a first discrimination result of the security inspection image, where the first discrimination result is a discrimination result obtained based on a manual image discrimination operation, and the second discrimination result is a discrimination result obtained based on an image analysis and recognition algorithm, and the first mode includes a first sub-mode and a second sub-mode, where the first discrimination result and the second discrimination result of the first sub-mode include that the security inspection image has a suspicion or the security inspection image has no suspicion; the first discrimination result and the second discrimination result of the second sub-mode include visual marking of the suspected area of the security inspection image; an auxiliary image discrimination module for, if the target working mode is selected as the second mode, obtaining the first discrimination result and the second discrimination result in real time; and a first discrimination conclusion generation module for generating a first discrimination conclusion based on the first discrimination result and the second discrimination result.

[0013] A third aspect of the present disclosure provides a security inspection system, including: an image acquisition device configured to scan a target object to obtain a security inspection image; an image interpretation device provided with an image analysis and recognition algorithm for generating a second discrimination result based on the security inspection image; an image interpretation station; and a scheduling server communicatively connected to the image acquisition device, the image interpretation device, and the image interpretation station; wherein the scheduling server is configured to obtain the security inspection image uploaded by the image acquisition device and send the security inspection image to the image interpretation station and the image interpretation device respectively, wherein the image interpretation station is configured to: select a target working mode; if the target working mode is selected as the first mode, obtain the second discrimination result after obtaining a first discrimination result of the security inspection image, wherein the first discrimination result is a discrimination result obtained based on manual image interpretation operation; if the target working mode is selected as the second mode, obtain the first discrimination result and the second discrimination result in real time; and generate a first discrimination conclusion based on the first discrimination result and the second discrimination result, wherein the first mode includes a first sub-mode and a second sub-mode, wherein the first discrimination result and the second discrimination result of the first sub-mode include that the security inspection image is suspicious or the security inspection image is not suspicious; the first discrimination result and the second discrimination result of the second sub-mode include visual marking of the suspicious area of the security inspection image.

[0014] A fourth aspect of the present disclosure provides an electronic device. The electronic device includes one or more processors and a memory. The memory is configured to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to execute the above method.

[0015] A fifth aspect of the present disclosure further provides a computer-readable storage medium having executable instructions stored thereon, and when the instructions are executed by a processor, the processor is caused to execute the above method.

[0016] A sixth aspect of the present disclosure further provides a computer program product including a computer program, and when the computer program is executed by a processor, the above method is implemented.

[0017] According to the method for discriminating security inspection images of the present disclosure, different working modes can be switched during security inspection work to meet the requirements in different scenarios: in an environment where high-precision judgment is required, the on-site observation ability and intuitive judgment advantages of security inspection personnel can be fully utilized, and the AI system will not interfere with the judgment of security inspection personnel, thereby ensuring the accuracy of recognition. Further, two different discrimination methods can be provided according to the actual requirements in this environment: one is fast and efficient discrimination, which can quickly screen a large number of images and improve work efficiency; the other is accurate marking of suspicious areas, which helps security inspection personnel more accurately identify potential threats by visually marking the suspicious areas; in an environment where a large number of images need to be processed efficiently and quickly, the security inspection speed can be maximally increased while ensuring the accuracy of security inspection, ensuring that the discrimination task can be completed quickly and accurately in a busy workplace. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 FIG. schematically shows an application scenario diagram of a method, device, equipment, medium, and program product for discriminating security inspection images according to an embodiment of the present disclosure;

[0019] Figure 2 FIG. schematically shows a flowchart of a method for discriminating security inspection images according to an embodiment of the present disclosure;

[0020] Figure 3A FIG. schematically shows a flowchart of a discrimination method for a full-image discrimination mode according to an embodiment of the present disclosure;

[0021] Figure 3B FIG. schematically shows a schematic diagram of a display interface for a full-image discrimination mode according to an embodiment of the present disclosure;

[0022] Figure 4A FIG. schematically shows a flowchart of a discrimination method for a suspicious-image discrimination mode according to an embodiment of the present disclosure;

[0023] Figure 4B FIG. schematically shows a schematic diagram of a display interface for a suspicious-image discrimination mode according to an embodiment of the present disclosure;

[0024] Figure 5 FIG. schematically shows a structural block diagram of a device for discriminating security inspection images according to an embodiment of the present disclosure; and

[0025] Figure 6 FIG. schematically shows a block diagram of an electronic device suitable for implementing a method for discriminating security inspection images according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] In order to make the objectives, technical solutions, and advantages of the present disclosure more clear and understandable, the present disclosure will be further described in detail below with reference to specific embodiments and the accompanying drawings.

[0027] However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the following detailed description, for the sake of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may be practiced without these specific details. In addition, in the following description, descriptions of well-known technologies are omitted to avoid unnecessarily obscuring the concepts of the present disclosure.

[0028] The terms used herein are merely for the purpose of describing specific embodiments and are not intended to limit the present disclosure. The term "comprising" as used herein indicates the presence of features, steps, operations, but does not preclude the presence or addition of one or more other features.

[0029] In the case of using expressions such as "at least one of A, B, and C, etc.", generally, it should be interpreted according to the meaning that those skilled in the art usually understand such expressions (for example, "a system having at least one of A, B, and C" should include, but is not limited to, a system having only A, only B, only C, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.). In the case of using expressions such as "at least one of A, B, or C, etc.", generally, it should be interpreted according to the meaning that those skilled in the art usually understand such expressions (for example, "a system having at least one of A, B, or C" should include, but is not limited to, a system having only A, only B, only C, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.).

[0030] The terms used herein are merely for the purpose of describing specific embodiments and are not intended to limit the present disclosure. The terms "including", "comprising", etc. as used herein indicate the presence of features, steps, operations, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

[0031] All terms used herein (including technical and scientific terms) have the meaning commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0032] With the increasingly strict security inspection requirements, especially in high-risk, busy places and complex environments, how to quickly and accurately judge security inspection images has become the key to enhancing security and efficiency. To cope with the increasing volume of security inspections and complex security threats, more and more security inspection systems are beginning to rely on image discrimination technology, especially AI-assisted image judgment devices, to process a large number of security inspection images in a short time through automated and intelligent technologies, helping security inspection personnel quickly identify potential threat items, reducing the burden of manual operations, and improving security inspection efficiency.

[0033] However, the applicant has found through research that existing security inspection image discrimination methods usually adopt fixed image judgment processes and technologies, and are unable to adjust the image judgment method according to the actual situation. When facing complex scenarios, they still adopt fixed processes and cannot flexibly adjust strategies according to the specific characteristics of the environment, images, and operation modes. For example, in some scenarios, it may be necessary to identify detailed abnormalities to enhance the accuracy of identification; while in other scenarios, more attention may be paid to efficiently and quickly processing a large number of images and reducing unnecessary interventions. The existing technology fails to solve this problem, resulting in the inability of the image judgment method to flexibly adapt to different application scenarios, which affects the overall effect of security inspection work.

[0034] Based on this, the embodiments of the present disclosure provide a method for discriminating security inspection images, including: selecting a target working mode; if the target working mode is selected as the first mode, obtaining a second discrimination result of the security inspection image after obtaining the first discrimination result of the security inspection image, where the first discrimination result is the discrimination result obtained based on manual image judgment operations, and the second discrimination result is the discrimination result obtained based on image analysis and recognition algorithms; if the target working mode is selected as the second mode, obtaining the first discrimination result and the second discrimination result in real time; and generating a first discrimination conclusion based on the first discrimination result and the second discrimination result. According to the method for discriminating security inspection images of the present disclosure, different working modes can be switched during security inspection work to meet the requirements in different scenarios: in an environment that requires high-precision judgment, the on-site observation ability and intuitive judgment advantages of security inspection personnel can be fully utilized, and the AI system will not interfere with the judgment of security inspection personnel, thereby ensuring the accuracy of identification. Further, two different discrimination methods can be provided according to the actual requirements in this environment: one is fast and efficient discrimination, which can quickly screen a large number of images and improve work efficiency; the other is precise marking of suspicious areas, which helps security inspection personnel more accurately identify potential threats by visually marking suspicious areas; in an environment that requires efficient and quick processing of a large number of images, the security inspection speed can be maximally improved while ensuring the accuracy of security inspection, ensuring that the discrimination task is completed quickly and accurately in a busy workplace.

[0035] Figure 1A schematic application scenario diagram of a method, apparatus, device, medium, and program product for discriminating security inspection images according to an embodiment of the present disclosure is shown.

[0036] As Figure 1 shown, the application scenario 100 according to this embodiment may include at least one X-ray security inspection device (two are shown in the figure, X-ray security inspection device 101 and X-ray security inspection device 102), a scheduling server 103, a picture judgment station 104, and a picture judgment device 105.

[0037] The X-ray security inspection device 101 and the X-ray security inspection device 102 can scan the objects passing through their interiors to obtain security inspection images.

[0038] An image analysis and recognition algorithm is deployed in the picture judgment device 105, and the received security inspection image can be discriminated by using this image analysis and recognition algorithm to determine whether the items in the security inspection image contain prohibited items. Exemplarily, the image analysis and recognition algorithm can be an analysis algorithm generated by an artificial intelligence learning platform through learning X-ray scan images. For example, a target detection algorithm (convolutional neural network) based on deep learning extracts prohibited item data features from a large amount of X-ray scan image data, models the prohibited item features, and completes target detection and recognition.

[0039] The picture judgment station 104 may include a picture judgment terminal 114 used by on-site operators. In some embodiments, the picture judgment station 104 may further include a quality control server 124 and a quality control terminal 134 used by quality control personnel.

[0040] The scheduling server 103 can receive the security inspection images uploaded by the X-ray security inspection device 101 and the X-ray security inspection device 102, and send the security inspection images to the picture judgment station 104 and the picture judgment device 105 respectively, so that the operator and the artificial intelligence model can perform image picture judgment synchronously and independently.

[0041] According to an embodiment of the present disclosure, a target working mode can be selected. When selecting the first mode, after the system obtains the first discrimination result of the security inspection image (that is, the result of the operator performing image discrimination using the picture judgment terminal 114), it will then obtain the second discrimination result (that is, the automatic discrimination result of the image analysis and the artificial intelligence model).

[0042] When continuing to select the first sub-mode of the first mode, the first discrimination result and the second discrimination result include that the security inspection image is suspicious or the security inspection image is not suspicious; when continuing to select the second sub-mode of the first mode, the first discrimination result and the second discrimination result include visual marking of the suspicious area of the security inspection image. And it can be judged whether secondary discrimination is required according to the conclusions of the two.

[0043] When the second mode is selected, the system will obtain the first discrimination result and the second discrimination result in real time. The operator can make a judgment based on their own observations and the prompts provided by the image discrimination device, and generate a discrimination conclusion.

[0044] The following will be based on Figure 1 the described scenario to describe in detail the method for discriminating security inspection images of the embodiments of the present disclosure.

[0045] Figure 2 Schematically shows a flowchart of the method for discriminating security inspection images according to an embodiment of the present disclosure.

[0046] As Figure 2 shown, the method for discriminating security inspection images may include operation S210 to operation S260.

[0047] In operation S210, first select a target working mode, and the target working mode will determine the discrimination strategy for the security inspection images. Specifically, the first mode is applicable to environments that require higher-precision discrimination, aiming to ensure that the image discrimination device does not interfere with the operator's judgment. This mode focuses on providing detailed and accurate discrimination results, ensuring that each discrimination is double-verified by both humans and the image discrimination device; while the second mode focuses on efficiently processing a large number of images, can obtain discrimination results in real time, and quickly feedback to the operator, and is applicable to large-scale security inspection tasks or busy security inspection environments. In this mode, the judgment results of the image discrimination device and the operator will be processed synchronously to ensure that neither efficiency is sacrificed nor relative accuracy is maintained.

[0048] If the target working mode is selected as the first mode, then in operation S220, after obtaining the first discrimination result of the security inspection image, obtain the second discrimination result of the security inspection image. Among them, the first discrimination result is the result of the operator's judgment on the security inspection image based on their experience and observations, while the second discrimination result is the automatic discrimination result based on image analysis and recognition algorithms. Through this dual discrimination method combining manual and automation, high-precision image analysis and more comprehensive security inspections can be ensured. More advantageously, in this mode, the image discrimination device can not interfere with the independent image discrimination of the image discriminator, avoiding the image discriminator's dependence on AI technology.

[0049] Furthermore, in the embodiments of the present disclosure, the first mode can be subdivided into a first sub-mode and a second sub-mode to further optimize the discrimination strategy according to different requirements in environments that require higher-precision discrimination.

[0050] If the first mode is selected as the first sub - mode, then in operation S230, the discrimination result may include determining whether there is a suspicion in the image, and the specific results include "the security inspection image is suspicious" or "the security inspection image is not suspicious". The first sub - mode is applicable to the rapid screening of security inspection images on the premise of double verification by manual and image - judging equipment, and can effectively make a preliminary determination in an environment with high - precision requirements and quickly identify potential risks. In this mode, the operator can quickly process the non - suspicious images, while screening out the suspicious images for further review.

[0051] If the first mode is selected as the second sub - mode, then in operation S240, the first discrimination result and the second discrimination result include visual marking of the suspicious area of the security inspection image.

[0052] In the embodiments of the present disclosure, the second sub - mode can perform visual marking on the suspicious area of the security inspection image. The operator and the image - judging equipment will respectively highlight or frame the suspicious area in the image, which is convenient for quickly focusing on the possible risky parts. This way of visual marking helps to improve the accuracy of image discrimination. Especially in complex scenarios, the operator can quickly lock the problem area through clear visual cues, thereby reducing the risks of missed judgment and misjudgment.

[0053] If the target working mode is selected as the second mode, then in operation S250, the first discrimination result and the second discrimination result can be obtained in real - time. In this mode, the operator can make a judgment by combining his own observation and the prompts of the image - judging equipment in real - time, quickly process a large number of security inspection tasks, and provide real - time feedback. It can be seen that this mode is applicable to environments that need to process a large number of images and can ensure efficient security inspection operations.

[0054] Specifically, when the X - ray security inspection equipment scans an object and generates an X - ray image, the image can be immediately displayed on the operator's image - judging terminal. At the same time, the image - judging equipment can analyze the image in real - time, extract the characteristics of possible suspicious objects, such as the shape, size, density, etc. of the object, and combine with a pre - trained model to automatically mark the suspicious area. These analysis results, including the marking of suspicious objects, category prompts, etc., will be directly displayed on the image and combined with the operator's observation in real - time.

[0055] When the operator views the image, the image - judging equipment will provide immediate feedback according to the image content. For example, when the image - judging equipment detects that an object in the image may be a prohibited item, it will mark the area on the image and display the category information of the object or other relevant prompts, such as "suspicious object", "battery - type item", etc. The operator can quickly make a judgment based on the real - time prompts provided by the image - judging equipment and his own professional judgment during this process.

[0056] Furthermore, the above real-time image judgment process can allow for immediate determination based on image data columns (i.e., partial image data, rather than the complete row packet image). Since the image data columns do not need to wait for the complete analysis of the entire image, judgments can be made when the image has not been fully scanned, providing real-time or delayed display results. This can not only speed up the security inspection process but also quickly screen suspicious luggage within a short time, providing a direct basis for subsequent sorting operations.

[0057] In operation S260, based on the first discrimination result and the second discrimination result, a first discrimination conclusion is generated. The first discrimination conclusion can indicate whether to conduct a manual inspection of the target object in the security inspection image, whether to allow the target object to pass, or whether a second manual discrimination is required to obtain a second discrimination conclusion. If a second manual discrimination is needed, the second discrimination conclusion will guide whether to continue the manual inspection or finally release the target object.

[0058] In fact, the first sub-mode and the second sub-mode correspond to the full image judgment mode and the suspicious image judgment mode. In the full image judgment mode, the operator needs to give a clear manual discrimination conclusion for each security inspection image, that is, whether the image is suspicious; while in the suspicious image judgment mode, the operator mainly makes a manual mark on the prohibited item area in the image. If there are no prohibited items in the image, the operator does not need to perform any operation; when the image contains suspicious items, the operator will directly mark on the image.

[0059] Figure 3A Schematically shows a flowchart of a discrimination method for the full image judgment mode according to an embodiment of the present disclosure; Figure 3B Schematically shows a schematic diagram of a display interface for the full image judgment mode according to an embodiment of the present disclosure; Figure 4A Schematically shows a flowchart of a discrimination method for the suspicious image judgment mode according to an embodiment of the present disclosure; Figure 4B Schematically shows a schematic diagram of a display interface for the suspicious image judgment mode according to an embodiment of the present disclosure.

[0060] Combined with reference to Figure 3A and Figure 3B, in the full image judgment mode, different first judgment conclusions can be corresponding to different contents of the first judgment result and the second judgment result. If the first judgment result indicates that the image is suspected, regardless of the second judgment result, the first judgment conclusion is that the target object in the image needs to be manually inspected. This means that once the security inspection image is manually marked as "suspected", the target object corresponding to the security inspection image will enter the manual inspection process and be carefully checked by security personnel. This method ensures the priority of manual image judgment. Although the image judgment device can quickly identify abnormalities or potential threats in the image through deep learning and a large amount of data analysis, in some complex or non-standard scenarios, for example, poor image quality, object overlap, insufficient light, or objects with irregular shapes, it may lead to inaccurate judgments by the image judgment device. Therefore, manual image judgment can make more accurate identifications based on the experience and intuition of security personnel, and the first judgment result corresponding to manual image judgment has a higher priority.

[0061] If both the first judgment result and the second judgment result indicate that the image is not suspected, the first judgment conclusion is to release the target object. In this case, the target object corresponding to the security inspection image is considered safe and does not require further intervention or inspection. Security personnel can allow the target item to pass through without causing unnecessary delays.

[0062] If the first judgment result is "not suspected", but the second judgment result shows that the image is suspected, the manual secondary image judgment process will be initiated. In this case, the operator will re-review the image to obtain the second judgment conclusion. This process is to ensure that even when initially manually judged as "not suspected", if the image judgment device discovers details that may have been overlooked, further manual confirmation can be carried out to avoid misjudgment or missed judgment.

[0063] In the embodiments of the present disclosure, when the second judgment result is "not suspected", the result of the image judgment device may not be displayed, and the first judgment result shall be directly used as the standard, so as not to interfere with the on-site manual operation.

[0064] Combined with reference Figure 4A and Figure 4B , in the suspected image judgment mode, similarly, different first judgment conclusions can be corresponding to different contents of the first judgment result and the second judgment result. When the second judgment result does not have a visual mark for the suspected area of the security inspection image, that is, the image judgment device does not mark any suspected areas, the first judgment conclusion can be generated only based on the first judgment result of the second sub-mode. In this case, if the first judgment result indicates that the image is not suspected, the first judgment conclusion is to release the target object without further manual intervention; when the first judgment result has a visual mark, the first judgment conclusion is to manually inspect the target object.

[0065] In an embodiment of the present disclosure, the visualization marker may be a marker box.

[0066] When the first discrimination result does not have a visualization marker for the suspicious area of the security inspection image, that is, no suspicious area is marked manually, but the second discrimination result has a visualization marker for the suspicious area of the security inspection image, the first discrimination conclusion is to perform a manual secondary image review on the security inspection image to obtain the second discrimination conclusion. In this case, as Figure 4B shown, the system will prompt the operator to perform a manual secondary image review. Therefore, the operator can check again for details in the image that may have been overlooked, avoiding misjudgment or missed judgment.

[0067] When both the first discrimination result and the second discrimination result have a visualization marker for the suspicious area of the security inspection image, the first discrimination conclusion is to perform a manual inspection on the target object in the image. At this time, whether it is the automatic marker of the image review device or the careful observation of the manual image review, they ultimately point to the same conclusion, that is, the target object poses a potential threat and requires more detailed review.

[0068] Furthermore, if the areas marked by the visualization markers in the first discrimination result and the second discrimination result are the same, only the first discrimination result will be displayed, and the second discrimination result will not be displayed. At this time, the analysis results of the image review device and the manual image review consistently indicate that the potential threat area in the image has been accurately identified. Therefore, the second discrimination result of the second sub-mode will not be displayed additionally, avoiding information redundancy and not interfering with the on-site manual operation.

[0069] If the areas marked by the visualization markers in the first discrimination result and the second discrimination result are different, the second discrimination result needs to be displayed separately. At this time, there are differences in the analysis results of the image review device and the manual image review, showing different suspicious areas. To ensure that the operator can comprehensively understand all potential risk areas, the marked area of the second discrimination result can be displayed separately. For example, a marker box with a different color from the first discrimination result can be used to provide the operator for further judgment and operation. In this way, the operator can conduct a detailed review based on the differences between the two to ensure that no possible threats are missed.

[0070] Furthermore, in the case where the second discrimination result does not have a visualization marker for the suspicious area of the security inspection image, the marker of the image review device can be not displayed, and directly take the marker of the first discrimination result as the standard, so as not to interfere with the on-site manual operation.

[0071] The applicant has further found through research that the image recognition algorithms in AI-assisted image judgment devices usually adopt fixed sensitivity settings and cannot be adaptively adjusted according to actual situations. The sensitivity setting directly affects the discrimination accuracy and error rate of the system. However, due to the lack of adaptive adjustment to the actual security inspection environment and image features, this fixed setting may lead to high false alarm rates and missed alarm rates of the system in different security inspection scenarios, thus affecting the security inspection efficiency and accuracy. Especially when encountering high-density items, occlusions, or low-quality images, the fixed sensitivity cannot flexibly respond, resulting in missed judgments of important items or unnecessary false alarms, increasing the burden of manual intervention and reducing the overall efficiency of the system.

[0072] Based on this, the image analysis and recognition algorithm according to the embodiments of the present disclosure includes a sensitivity parameter that can be dynamically adjusted, and this sensitivity parameter is used to adjust the recognition sensitivity of the image analysis and recognition algorithm to abnormal objects in the security inspection image.

[0073] Exemplarily, the sensitivity parameter may include a threshold for image feature extraction and a confidence threshold for target detection.

[0074] The image feature extraction threshold determines the sensitivity of extracting features in the image. A lower image feature extraction threshold means being more sensitive to the detailed features in the image, capable of capturing more minute abnormalities or uncommon details, thus helping to discover potential threats, especially when processing security inspection images with poor image quality or complex backgrounds.

[0075] The confidence threshold for target detection controls the strictness of the judgment on whether a security inspection object is a suspicious target. A higher confidence threshold requires the image judgment device to have a higher certainty about whether the target is a prohibited item or a suspicious item when making a judgment. Therefore, a higher confidence threshold can effectively reduce the possibility of false alarms and prevent misjudging normal items as threats.

[0076] In the embodiments of the present disclosure, the sensitivity parameter can be dynamically adjusted based on the credit level. Specifically, the threshold of the sensitivity parameter can be adjusted based on the credit level of the target object.

[0077] Exemplarily, for objects with a high credit level, the confidence threshold for target detection can be appropriately increased, and the threshold for image feature extraction can be decreased. This means that the image judgment will be relatively lenient, reducing unnecessary inspection prompts and manual intervention. Therefore, it can help improve the passing efficiency, enabling objects with low risks to pass through the security inspection more quickly and avoiding over-inspection and unnecessary delays.

[0078] For objects with a medium credit level, the conventional sensitivity parameter settings can be adopted. At this time, the image judgment device works according to the standard image judgment rules and does not impose additional strict inspection requirements on all images.

[0079] For objects with a low credit rating, the confidence threshold for object detection can be lowered, the threshold for image feature extraction can be increased, and the sensitivity coefficient for the recognition of abnormal items can be increased. This setting enables the image judgment device to conduct more meticulous and stringent inspections on objects with a low credit rating, thereby enhancing the ability to identify potential threats.

[0080] For example, the confidence threshold for object detection of a high credit rating can be set at 0.9, indicating that the system has relatively loose requirements for the judgment of this item; while for an item with a low credit rating, the confidence threshold for object detection can be lowered to 0.5 to increase the inspection intensity of the system and the ability to identify abnormal items.

[0081] In an embodiment of the present disclosure, the associated information of the target object in the security inspection image can also be provided by obtaining multi-dimensional information. The multi-dimensional information may include, but is not limited to, the customs declaration information, logistics information, etc. of the goods, aiming to enhance the judgment ability of the target object in the image by associating with other data sources. Based on these multi-dimensional information, an auxiliary discrimination result can be generated, which is used to correct or verify the second discrimination result generated by the image judgment device to ensure the comprehensiveness and accuracy of the discrimination process.

[0082] Exemplarily, the scheduling server can obtain the goods information on the customs declaration form and perform an associated comparison between the security inspection images of the goods in the same batch and the goods information on the customs declaration form. For example, data such as the name, specification, quantity, and value of the goods can be compared. For example, if the customs declaration form declares a batch of fruits of the same specification, but the security inspection image shows some metal parts, it can be determined that there is an abnormality in the security inspection image. This method helps to more accurately classify the items in the security inspection image and prevent incorrect judgments caused by misidentification or data inconsistency.

[0083] Exemplarily, the scheduling server can obtain the logistics information of the object, such as the place of dispatch, place of receipt, transportation route, and transportation mode, etc., to assist in judging the type difference of the items in the same batch. Certain specific regions may be famous for producing or exporting specific types of goods. Therefore, through the logistics information, preliminary screening and judgment can be carried out, which can improve the accuracy and efficiency of recognition. For example, assume that a batch of goods is declared to come from a tea-producing area of a certain region, but other items unrelated to tea appear in the security inspection image. Combining the logistics information, such abnormal situations can be discovered, thereby triggering further inspections or correcting the discrimination results.

[0084] The discrimination method of security inspection images disclosed in this disclosure aims to improve the efficiency and accuracy of security inspection by combining artificial intelligence technology and the professional judgment of security inspection image interpreters, especially in the identification and processing of high-risk items. For example, AI algorithms can search and identify target categories with shape and material property characteristics, such as explosive devices, firearms and ammunition, pressure gas cylinders, etc. These items have unique shapes, material properties or abnormal characteristics, and the AI system can efficiently identify and mark them, providing strong auxiliary judgment for security inspection image interpreters.

[0085] It should be noted that the intelligent image interpretation mode and sensitivity differential adjustment technologies described in this disclosure are not limited to X-ray machines. This technology is also applicable to other imaging technologies such as CT and backscatter imaging. That is to say, the method of this disclosure can adapt to different types of security inspection equipment, improving the working efficiency and accuracy of different equipment in various scenarios.

[0086] Based on the above discrimination method of security inspection images, this disclosure also provides a discrimination device for security inspection images. The following will be combined with Figure 5 to describe this device in detail.

[0087] Figure 5 The structural block diagram of the discrimination device for security inspection images according to an embodiment of this disclosure is schematically shown.

[0088] As Figure 5 shown, the discrimination device 500 for security inspection images in this embodiment includes a mode selection module 510, a dual image interpretation module 520, an auxiliary image interpretation module 530, and a first discrimination conclusion generation module 540.

[0089] The mode selection module 510 can be used to select the target working mode. In one embodiment, the mode selection module 510 can be used to perform the operation S210 described above, which will not be elaborated here.

[0090] The dual image interpretation module 520 can be used to obtain the second discrimination result of the security inspection image after obtaining the first discrimination result of the security inspection image if the target working mode is selected as the first mode, where the first discrimination result is the image interpretation result obtained based on manual image interpretation operation, and the second discrimination result is the image interpretation result obtained based on image analysis and recognition algorithms. The first mode includes a first sub-mode and a second sub-mode, where the first discrimination result and the second discrimination result of the first sub-mode include that the security inspection image is suspected or the security inspection image is not suspected; the first discrimination result and the second discrimination result of the second sub-mode include visual marking of the suspected area of the security inspection image. In one embodiment, the dual image interpretation module 520 can be used to perform the operations S220 - S240 described above, which will not be elaborated here.

[0091] The auxiliary image discrimination module 530 can be used to obtain the first discrimination result and the second discrimination result in real time if the target working mode is selected as the second mode. In one embodiment, the auxiliary image discrimination module 530 can be used to perform the operation S250 described above, which will not be elaborated here.

[0092] The first discrimination conclusion generation module 540 can be used to generate a first discrimination conclusion based on the first discrimination result and the second discrimination result. In one embodiment, the first discrimination conclusion generation module 540 can be used to perform the operation S260 described above, which will not be elaborated here.

[0093] According to an embodiment of the present disclosure, the dual image discrimination module 520 can also be used to, when the first discrimination result of the first sub-mode is that the security inspection image is suspicious and the first discrimination conclusion is to manually inspect the target object in the security inspection image; when the first discrimination result of the first sub-mode is that the security inspection image is not suspicious and the second discrimination result of the first sub-mode is that the security inspection image is not suspicious, the first discrimination conclusion is to release the target object in the security inspection image; and when the first discrimination result of the first sub-mode is that the security inspection image is not suspicious and the second discrimination result of the first sub-mode is that the security inspection image is suspicious, the first discrimination conclusion is to perform a second manual image discrimination on the security inspection image to obtain a second discrimination conclusion.

[0094] According to an embodiment of the present disclosure, the dual image discrimination module 520 can also be used to, when the second discrimination result of the second sub-mode does not have a visual marker for the suspicious area of the security inspection image, generate a first discrimination conclusion based on the first discrimination result of the second sub-mode; when the first discrimination result of the second sub-mode does not have a visual marker for the suspicious area of the security inspection image and the second discrimination result of the second sub-mode has a visual marker for the suspicious area of the security inspection image, the first discrimination conclusion is to perform a second manual image discrimination on the security inspection image to obtain a second discrimination conclusion; and when both the first discrimination result and the second discrimination result of the second sub-mode have visual markers for the suspicious area of the security inspection image, the first discrimination conclusion is to manually inspect the target object in the security inspection image.

[0095] According to an embodiment of the present disclosure, the dual image discrimination module 520 can also be used to, if the areas marked visually in the first discrimination result and the second discrimination result of the second sub-mode are the same, not display the second discrimination result of the second sub-mode; and if the areas marked visually in the first discrimination result and the second discrimination result of the second sub-mode are different, display the second discrimination result of the second sub-mode.

[0096] According to an embodiment of the present disclosure, the dual image discrimination module 530 can also be used when the first discrimination result of the second sub-mode does not have a visual mark for the suspicious area of the security inspection image, and the first discrimination conclusion is to release the target object in the security inspection image; and when the first discrimination result of the second sub-mode has a visual mark for the suspicious area of the security inspection image, the first discrimination conclusion is to conduct manual inspection on the target object in the security inspection image.

[0097] According to an embodiment of the present disclosure, the discrimination device 500 for security inspection images can also be used to adjust the threshold of the sensitivity parameter based on the credit level of the target object in the security inspection image, and the sensitivity parameter is used to adjust the recognition sensitivity of the image analysis and recognition algorithm to abnormal objects in the security inspection image.

[0098] According to an embodiment of the present disclosure, the discrimination device 500 for security inspection images can also be used to obtain multi-dimensional information, where the multi-dimensional information is used to provide associated information of the target object in the security inspection image; and based on the multi-dimensional information, generate an auxiliary discrimination result, where the auxiliary discrimination result is used to correct or verify the second discrimination result.

[0099] According to an embodiment of the present disclosure, any multiple of the mode selection module 510, the dual image discrimination module 520, the auxiliary image discrimination module 530, and the first discrimination conclusion generation module 540 can be combined and implemented in one module, or any one of them can be split into multiple modules. Or, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the mode selection module 510, the dual image discrimination module 520, the auxiliary image discrimination module 530, and the first discrimination conclusion generation module 540 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on chip, a system on substrate, a system on package, an application specific integrated circuit (ASIC), or can be implemented by any other reasonable way of integrating or packaging circuits, etc., by hardware or firmware, or implemented in any one of the three implementation ways of software, hardware, and firmware, or in any appropriate combination of them. Or, at least one of the mode selection module 510, the dual image discrimination module 520, the auxiliary image discrimination module 530, and the first discrimination conclusion generation module 540 can be at least partially implemented as a computer program module, and when the computer program module is run, it can execute the corresponding functions.

[0100] Figure 6 A block diagram of an electronic device suitable for implementing the discrimination method of security inspection images according to an embodiment of the present disclosure is schematically shown.

[0101] As Figure 6As shown, the electronic device 600 according to an embodiment of the present disclosure includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage section 608 into a random access memory (RAM) 603. The processor 601 can include, for example, a general microprocessor (such as a CPU), an instruction set processor, and / or a related chipset, and / or a dedicated microprocessor (such as an application specific integrated circuit (ASIC)), etc. The processor 601 can also include on-board memory for caching purposes. The processor 601 can include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0102] In the RAM 603, various programs and data required for the operation of the electronic device 600 are stored. The processor 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. The processor 601 performs various operations of the method flow according to an embodiment of the present disclosure by executing the program in the ROM 602 and / or the RAM 603. It should be noted that the program can also be stored in one or more memories other than the ROM 602 and the RAM 603. The processor 601 can also perform various operations of the method flow according to an embodiment of the present disclosure by executing the program stored in the one or more memories.

[0103] According to an embodiment of the present disclosure, the electronic device 600 may further include an input / output (I / O) interface 605, and the input / output (I / O) interface 605 is also connected to the bus 604. The electronic device 600 may further include one or more of the following components connected to the input / output (I / O) interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including, for example, a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed so that a computer program read therefrom can be installed into the storage section 608 as needed.

[0104] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the foregoing embodiments; or may exist independently without being assembled into the device / apparatus / system. The foregoing computer-readable storage medium stores one or more programs, and when the one or more programs are executed, the methods according to the embodiments of the present disclosure are implemented.

[0105] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. In the present disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program may be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 602 and / or the RAM 603 described above and / or one or more memories other than the ROM 602 and the RAM 603.

[0106] An embodiment of the present disclosure further includes a computer program product, which includes a computer program, and the computer program includes program code for executing the method shown in the flowchart. When the computer program product runs in a computer system, the program code is used to cause the computer system to implement the method for discriminating security inspection images provided by the embodiments of the present disclosure.

[0107] When the computer program is executed by the processor 601, the above functions defined in the system / apparatus of the embodiments of the present disclosure are executed. According to an embodiment of the present disclosure, the foregoing systems, apparatuses, modules, units, etc. may be implemented by computer program modules.

[0108] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices and magnetic storage devices. In another embodiment, the computer program may also be transmitted and distributed in the form of a signal on a network medium, and is downloaded and installed through the communication part 609, and / or installed from the removable medium 611. The program code included in the computer program may be transmitted by any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the foregoing.

[0109] In such an embodiment, the computer program can be downloaded and installed from a network through the communication section 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, the above functions defined in the system of the embodiments of the present disclosure are executed. According to the embodiments of the present disclosure, the above-described systems, devices, apparatuses, modules, units, etc. can be implemented by computer program modules.

[0110] According to the embodiments of the present disclosure, the program code for executing the computer program provided by the embodiments of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages include, but are not limited to, such as Java, C++, python, the "C" language, or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, by using an Internet service provider to connect through the Internet).

[0111] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the above module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram or flowchart, and the combination of blocks in the block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0112] The above describes the embodiments of the present disclosure. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described separately above, this does not mean that the measures in each embodiment cannot be used advantageously in combination. Without departing from the scope of the present disclosure, those skilled in the art can make various substitutions and modifications, and these substitutions and modifications should fall within the scope of the present disclosure.

Claims

1. A method for distinguishing security inspection images, characterized in that: The method comprises: Select the target working mode; If the target working mode is selected as the first mode, a second discrimination result of the security inspection image is obtained after the first discrimination result of the security inspection image is obtained, wherein the first discrimination result is a discrimination result obtained based on a manual image discrimination operation, and the second discrimination result is a discrimination result obtained based on an image analysis and recognition algorithm; If the target working mode is selected as the second mode, obtaining the first discrimination result and the second discrimination result in real time; and Based on the first discrimination result and the second discrimination result, a first discrimination conclusion is generated, Among them, the first mode includes a first sub-mode and a second sub-mode, wherein the first judgment result and the second judgment result of the first sub-mode include that the security inspection image is suspicious or the security inspection image is not suspicious; the first judgment result and the second judgment result of the second sub-mode include visual marking of the suspicious area of ​​the security inspection image.

2. The method according to claim 1, characterized in that If the target working mode is selected as the first sub-mode, generating a first discrimination conclusion based on the first discrimination result and the second discrimination result specifically includes: When the first discrimination result of the first sub-mode is that the security inspection image is suspicious, the first discrimination conclusion is to manually check the target object in the security inspection image; When the first discrimination result of the first sub-mode is that the security inspection image is not suspicious and the second discrimination result of the first sub-mode is that the security inspection image is not suspicious, the first discrimination conclusion is to release the target object in the security inspection image; and When the first discrimination result of the first sub-mode is that the security inspection image is not suspicious and the second discrimination result of the first sub-mode is that the security inspection image is suspicious, the first discrimination conclusion is to manually perform a second discrimination on the security inspection image to obtain the second discrimination conclusion.

3. The method according to claim 1, characterized in that If the target working mode is selected as the second sub-mode, generating a first discrimination conclusion based on the first discrimination result and the second discrimination result specifically includes: When the second discrimination result of the second sub-mode does not have a visual mark for the suspected area of ​​the security inspection image, generating a first discrimination conclusion based on the first discrimination result of the second sub-mode; When the first discrimination result of the second sub-mode does not have a visual mark for the suspected area of ​​the security inspection image, and the second discrimination result of the second sub-mode has a visual mark for the suspected area of ​​the security inspection image, the first discrimination conclusion is to manually perform a second discrimination on the security inspection image to obtain the second discrimination conclusion; and When the first discrimination result of the second sub-mode and the second discrimination result of the second sub-mode both have a visual mark for the suspected area of ​​the security inspection image, the first discrimination conclusion is a manual inspection of the target object in the security inspection image.

4. The method according to claim 3, characterized in that When the first discrimination result of the second sub-mode and the second discrimination result of the second sub-mode both have a visual mark for the suspected area of ​​the security inspection image, the method further includes: If the first discrimination result of the second sub-mode and the area of ​​the visual mark in the second discrimination result of the second sub-mode are consistent, the second discrimination result of the second sub-mode is not displayed; and If the first discrimination result of the second sub-mode and the area of ​​the visual mark in the second discrimination result of the second sub-mode are inconsistent, the second discrimination result of the second sub-mode is displayed.

5. The method according to claim 3 or 4, characterized in that: When the second discrimination result of the second sub-mode does not have a visual mark for the suspected area of ​​the security inspection image, generating a first discrimination conclusion based on the first discrimination result of the second sub-mode specifically includes: When the first discrimination result of the second sub-mode does not have a visual mark for the suspicious area of ​​the security inspection image, the first discrimination conclusion is to release the target object in the security inspection image; and When the first discrimination result of the second sub-mode has a visual mark for the suspicious area of ​​the security inspection image, the first discrimination conclusion is a manual inspection of the target object in the security inspection image.

6. The method according to any one of claims 1 to 4, characterized in that: The image analysis and recognition algorithm includes a dynamically adjustable sensitivity parameter, and the sensitivity parameter is used to adjust the recognition sensitivity of the image analysis and recognition algorithm to abnormal objects in the security inspection image.

7. The method according to claim 6, characterized in that The threshold of the sensitivity parameter is adjusted based on the credit rating of the target object in the security inspection image.

8. The method according to any one of claims 1 to 4 and 7, characterized in that The method further comprises: Acquiring multi-dimensional information, where the multi-dimensional information is used to provide association information of a target object in the security inspection image; and Based on the multi-dimensional information, an auxiliary discrimination result is generated, and the auxiliary discrimination result is used to correct or verify the second discrimination result.

9. A security inspection image identification device, characterized in that: The device comprises: A mode selection module is used to: select a target working mode; A dual image discrimination module, for: if the target working mode is selected as the first mode, obtaining a second discrimination result of the security inspection image after obtaining the first discrimination result of the security inspection image, wherein the first discrimination result is a discrimination result obtained based on a manual image discrimination operation, and the second discrimination result is a discrimination result obtained based on an image analysis and recognition algorithm, and the first mode includes a first sub-mode and a second sub-mode, wherein the first discrimination result and the second discrimination result of the first sub-mode include that the security inspection image is suspicious or that the security inspection image is not suspicious; the first discrimination result and the second discrimination result of the second sub-mode include visual marking of a suspicious area of ​​the security inspection image; an auxiliary image discrimination module, configured to: if the target working mode is selected as the second mode, obtain the first discrimination result and the second discrimination result in real time; and The first discrimination conclusion generating module is used to generate a first discrimination conclusion based on the first discrimination result and the second discrimination result.

10. A security inspection system, comprising: An image acquisition device, wherein the image acquisition device is used to scan a target object to obtain a security inspection image; An image judging device, wherein the image judging device is provided with an image analysis and recognition algorithm, and is used to generate a second judging result based on the security inspection image; Image judging station; as well as a scheduling server, wherein the scheduling server is in communication with the image acquisition device, the image judgment device and the image judgment station; wherein the scheduling server is used to obtain the security inspection image uploaded by the image acquisition device, and send the security inspection image to the image judgment station and the image judgment device respectively, Wherein, the image judging station is used for: Select the target working mode; If the target working mode is selected as the first mode, the second discrimination result is obtained after the first discrimination result of the security inspection image is obtained, wherein the first discrimination result is a discrimination result obtained based on a manual image discrimination operation; If the target working mode is selected as the second mode, obtaining the first discrimination result and the second discrimination result in real time; and Based on the first discrimination result and the second discrimination result, a first discrimination conclusion is generated, Among them, the first mode includes a first sub-mode and a second sub-mode, wherein the first judgment result and the second judgment result of the first sub-mode include that the security inspection image is suspicious or the security inspection image is not suspicious; the first judgment result and the second judgment result of the second sub-mode include visual marking of the suspicious area of ​​the security inspection image.

11. An electronic device, comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more The processor executes the method according to any one of claims 1 to 8.

12. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method according to any one of claims 1 to 8.

13. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images