CN115389532A

CN115389532A - Article safety inspection method and device, computer equipment and storage medium

Info

Publication number: CN115389532A
Application number: CN202110495224.4A
Authority: CN
Inventors: 赖敬文; 李小帆; 招振荣; 刘庆龙; 尹东富; 余心浩
Original assignee: SF Technology Co Ltd
Current assignee: SF Technology Co Ltd
Priority date: 2021-05-07
Filing date: 2021-05-07
Publication date: 2022-11-25

Abstract

The application relates to an article security inspection method, an article security inspection device, a computer device and a storage medium. The method comprises the following steps: acquiring an image to be safely checked corresponding to an article to be safely checked, wherein the image to be safely checked carries binding identity information and a transportation sequence identifier; inputting the image to be safely checked into the trained target article detection model, and carrying out article type detection on the image to be safely checked through the trained target article detection model to obtain an article type detection result; obtaining a safety inspection result according to the article type in the article type detection result; when the security inspection result shows that the target contraband exists, generating an interception prompt according to the target contraband and the binding identity information, and determining a target prompt terminal according to the transportation sequence identifier; and pushing the interception prompt to a target prompt terminal. By adopting the method, the security inspection efficiency can be improved, and the security inspection effect can be improved.

Description

Article safety inspection method and device, computer equipment and storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for security inspection of an article, a computer device, and a storage medium.

Background

With the development of computer technology, a security inspection technology has appeared, which is to perform security inspection on articles to detect whether there are prohibited objects and process them, and can be applied to areas such as logistics transition, airports, etc.

In the conventional technology, a terminal is mainly used for scanning an article to be detected, and a security inspection image corresponding to the article to be detected is displayed, so that a worker beside the terminal can distinguish the article according to the security inspection image, and the security inspection is performed in a mode of obtaining a distinguishing result and feeding back the distinguishing result to process the distinguishing result.

However, in the conventional method, since the worker determines the security inspection image according to the displayed security inspection image and processes the security inspection image, the security inspection efficiency is low and the security inspection effect is poor.

Disclosure of Invention

In view of the above, it is necessary to provide an article security inspection method, apparatus, computer device and storage medium capable of improving security inspection efficiency and security inspection effect.

A method of security inspection of an article, the method comprising:

acquiring an image to be safely checked corresponding to an article to be safely checked, wherein the image to be safely checked carries binding identity information and a transportation sequence identifier;

inputting the image to be safely checked into the trained target object detection model, and performing object type detection on the image to be safely checked through the trained target object detection model to obtain a safety check result;

when the safety inspection result shows that the target contraband exists, generating an interception prompt according to the target contraband and the binding identity information, and determining a target prompt terminal according to the transportation sequence identifier;

and pushing the interception prompt to a target prompt terminal.

In one embodiment, acquiring an image to be security inspected corresponding to an article to be security inspected comprises:

acquiring a transportation sequence identifier corresponding to an article to be subjected to security inspection and first identity information before article security inspection scanning;

carrying out safety inspection scanning on an article to be subjected to safety inspection to obtain a scanning image corresponding to the safety inspection scanning;

acquiring second identity information after the article security inspection is scanned according to the corresponding transportation sequence identifier;

and carrying out identity verification on the second identity information according to the first identity information, carrying out image code binding on the scanned image according to the identity verification result, and associating a corresponding transportation sequence identifier to obtain an image to be subjected to security inspection.

In one embodiment, performing identity verification on the second identity information according to the first identity information, performing image code binding on the scanned image according to the identity verification result, and associating the corresponding transportation sequence identifier to obtain the image to be subjected to security inspection includes:

comparing the first identity information with the second identity information;

and when the second identity information corresponds to the first identity information, binding the second identity information and the scanned image, and associating the corresponding transportation sequence identifier to obtain an image to be safely checked.

In one embodiment, before inputting an image to be security-inspected into a trained target object detection model and performing object type detection on the image to be security-inspected through the trained target object detection model to obtain a security inspection result, the method further includes:

acquiring a training image set carrying article image annotation and an initial article detection model;

and training the initial article detection model according to the training image set to obtain a trained target article detection model.

In one embodiment, inputting an image to be security-inspected into a trained target object detection model, and performing object type detection on the image to be security-inspected through the trained target object detection model to obtain a security inspection result includes:

inputting the image to be safely checked into the trained target article detection model, and carrying out article type detection on the image to be safely checked through the trained target article detection model to obtain an article type detection result;

when the article type in the article type detection result comprises contraband, acquiring a type confidence coefficient corresponding to the contraband;

when the category confidence is greater than or equal to a preset confidence threshold, obtaining a safety inspection result as the existence of the target contraband;

and when the category confidence coefficient is smaller than the confidence coefficient threshold value, generating and displaying a security inspection image judgment page according to the image to be security inspected and the contraband, receiving an image judgment result corresponding to the security inspection image judgment page, and obtaining a security inspection result according to the image judgment result.

In one embodiment, determining the target prompting terminal according to the transportation sequence identifier comprises:

acquiring transportation sequence information of each transportation line;

and determining a target transport line and a target prompt terminal corresponding to the sorting operation on the target transport line according to the transport sequence information and the transport sequence identifier.

In one embodiment, after pushing the interception prompt to the target prompt terminal, the method further includes:

receiving an interception feedback result corresponding to the interception prompt;

and when the interception feedback result is an interception error, training a trained target article detection model according to the article detection result in the interception feedback result and the image to be subjected to security inspection.

An article security inspection apparatus, the apparatus comprising:

the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring an image to be safely checked corresponding to an article to be safely checked, and the image to be safely checked carries binding identity information and a transportation sequence identifier;

the detection module is used for inputting the image to be safely checked into the trained target article detection model and carrying out article type detection on the image to be safely checked through the trained target article detection model to obtain a safety check result;

the processing module is used for generating an interception prompt according to the target contraband and the binding identity information and determining a target prompt terminal according to the transportation sequence identifier when the safety inspection result shows that the target contraband exists;

and the pushing module is used for pushing the interception prompt to the target prompt terminal.

A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:

inputting the image to be safely checked into the trained target object detection model, and carrying out object type detection on the image to be safely checked through the trained target object detection model to obtain a safety check result;

when the security inspection result shows that the target contraband exists, generating an interception prompt according to the target contraband and the binding identity information, and determining a target prompt terminal according to the transportation sequence identifier;

and pushing the interception prompt to a target prompt terminal.

A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:

and pushing the interception prompt to a target prompt terminal.

According to the article security inspection method, the article security inspection device, the computer equipment and the storage medium, after the image to be security inspected corresponding to the article to be security inspected is obtained, the image to be security inspected is input into the trained target article detection model, article type detection is performed on the image to be security inspected through the trained target article detection model, accurate article type identification of the image to be security inspected can be achieved, security inspection effects and security inspection results are improved, when the security inspection results indicate that target contraband exists, an interception prompt is generated according to the target contraband and the binding identity information, the target prompt terminal is determined according to the transportation sequence identification, the interception prompt is pushed to the target prompt terminal, automatic security inspection is completed, and security inspection efficiency can be improved.

Drawings

FIG. 1 is a schematic flow chart diagram illustrating a security check method according to one embodiment;

FIG. 2 is a diagram of an exemplary security check method;

FIG. 3 is a diagram showing an application environment of a security check method in another embodiment;

FIG. 4 is a schematic flow chart of a security check method according to another embodiment;

FIG. 5 is a block diagram showing the structure of a security check apparatus according to an embodiment;

FIG. 6 is a diagram of the internal structure of a computer device in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

In an embodiment, as shown in fig. 1, an article security inspection method is provided, and this embodiment is illustrated by applying the method to a terminal, and it is to be understood that the method may also be applied to a server, and may also be applied to a system including the terminal and the server, and is implemented by interaction between the terminal and the server. In this embodiment, the method includes the steps of:

102, acquiring an image to be safely checked corresponding to an article to be safely checked, wherein the image to be safely checked carries binding identity information and a transportation sequence identifier.

The article to be subjected to security inspection refers to an article which needs to be subjected to security inspection. For example, the article to be safely inspected may specifically refer to a package in a logistics transit. As another example, the object to be security checked may specifically refer to luggage that needs to be security checked at an airport. The image to be safely checked refers to an image corresponding to the article to be safely checked, which is obtained through scanning. For example, the image to be security-inspected may specifically refer to an X-ray image obtained by scanning and corresponding to the article to be security-inspected. The binding identity information refers to identity information used for representing different articles to be safely inspected, for example, the binding identity information may specifically refer to a barcode corresponding to the article to be safely inspected, and the article to be safely inspected may specifically be a package. The transportation sequence identifier refers to an identifier for identifying the current transportation position of the items to be safely inspected, and the transportation sequence identifier of each item to be safely inspected is unique.

Specifically, when security inspection is required, the terminal can acquire an image to be securely inspected corresponding to an article to be securely inspected, and the image to be securely inspected carries binding identity information and a transportation sequence identifier.

And 104, inputting the image to be safely checked into the trained target object detection model, and performing object type detection on the image to be safely checked through the trained target object detection model to obtain a safety check result.

The trained target article detection model refers to a model which is trained in advance and used for detecting article types of images to be subjected to safety inspection. For example, the trained target object detection model may be a model constructed based on an R-CNN (R-Convolutional Neural Networks) network. For another example, the trained target object detection model may be a model constructed based on a YOLO (young Only Look Once) network. In this embodiment, the target article detection model is not specifically limited as long as article type detection can be achieved. The item type detection refers to detecting the item type existing in the image to be safely checked. For example, common categories of items include: lighters, mobile phones, milk powders, cosmetics, drinks, heat patches, dry batteries/lithium batteries, compressed gas tanks, cutters, water/unidentified liquids, powdered objects, tablet computers, guns, toy guns, iron case lighters, notebook computers, button batteries, scissors, charge pencils, explosives, folding umbrellas, and the like. The security inspection result refers to a result obtained by performing security inspection on an article to be security inspected, and includes the existence of target contraband and the absence of target contraband.

Specifically, after the image to be subjected to security inspection is acquired, the terminal inputs the image to be subjected to security inspection into the trained target article detection model, performs article type detection on the image to be subjected to security inspection through the trained target article detection model, obtains an article type in the image to be subjected to security inspection and a type confidence coefficient corresponding to the article type, namely an article type detection result, and further obtains a security inspection result according to the article type detection result. Further, when the trained target article detection model is used for detecting article types of images to be safely checked, the trained target article detection model firstly identifies the targets of the images to be safely checked, identifies areas where articles may exist in the images to be safely checked, then detects the types of the articles in the areas to obtain confidence coefficients of the articles belonging to the preset article types, orders the confidence coefficients, and determines the article types corresponding to the articles in the areas and the confidence coefficients corresponding to the article types. The preset item category may be set by itself as needed, for example, the preset item category may be specifically the above-mentioned common item category.

In order to improve the security inspection efficiency, in the present embodiment, the article type detection may be performed by parallel processing. For example, the parallel processing mode may specifically create multiple article type detection tasks at the same time, and perform article type detection on multiple images to be subjected to security inspection by using multiple trained target article detection models, respectively, to obtain an article type detection result.

Further, after the object type detection result is obtained, the terminal compares preset contraband according to the object type in the object type detection result to determine whether the contraband exists in the object type, when the contraband does not exist in the object type, the security inspection result is obtained that the target contraband does not exist, when the contraband exists in the object type, the terminal needs to further obtain a type confidence coefficient corresponding to the contraband, and the security inspection result is obtained according to the type confidence coefficient.

And 108, when the security inspection result shows that the target contraband exists, generating an interception prompt according to the target contraband and the binding identity information, and determining a target prompt terminal according to the transportation sequence identifier.

The interception prompt is a prompt for indicating a worker to intercept the article to be subjected to security inspection. The target prompt terminal is a terminal which can be used by a worker for intercepting the article to be safely checked. For example, the target prompt terminal may specifically refer to a branch line code scanning device, a trunk loading scanning device, and the like in a logistics transition.

Specifically, when the security inspection result shows that the target contraband exists, the object to be safely inspected needs to be intercepted, the terminal generates an interception prompt according to the target contraband and the binding identity information, positions the current transportation position of the object to be safely inspected according to the transportation sequence identifier, and determines the target prompt terminal which needs to push the interception prompt according to the current transportation position.

And step 110, pushing the interception prompt to a target prompt terminal.

Specifically, the terminal can propelling movement interception suggestion to target suggestion terminal, and the staff who uses target suggestion terminal can receive the interception suggestion to when utilizing target suggestion terminal scanning to treat the safety inspection article, can know through the interception suggestion and treat and probably contain the forbidden object of target in the safety inspection article, treat the safety inspection article and intercept. It should be noted that, in addition to pushing the interception prompt to the target prompt terminal, the terminal may also control the preset service system, and physically perform diversion interception on the object to be security-inspected including the target contraband, that is, by outputting a control instruction to the preset service system, the object to be security-inspected including the target contraband is directly diverted to the preset exception handling area. The preset business system is used for controlling sorting and shunting of articles to be safely inspected.

According to the article security inspection method, after the image to be safely inspected corresponding to the article to be safely inspected is obtained, the image to be safely inspected is input into the trained target article detection model, article type detection is carried out on the image to be safely inspected through the trained target article detection model, accurate article type identification of the image to be safely inspected can be achieved, the security inspection effect and the security inspection result are improved, when the security inspection result is that target contraband exists, an interception prompt is generated according to the target contraband and the binding identity information, the target prompt terminal is determined according to the transportation sequence identification, the interception prompt is pushed to the target prompt terminal, automatic security inspection is completed, and the security inspection efficiency can be improved.

The scanning image refers to a corresponding image obtained by scanning an article to be safely inspected. For example, the scanned image may specifically refer to an X-ray image obtained by scanning and corresponding to an article to be security-inspected. The first identity information refers to identity information obtained before security check scanning is performed on an article, for example, the first identity information may specifically refer to barcode information used for identifying an article to be security checked, and the barcode information may specifically refer to waybill information of a package, including one-dimensional barcode information and two-dimensional barcode information. The second identity information refers to identity information obtained after security inspection scanning is performed on an article, for example, the second identity information may specifically refer to barcode information used for identifying an article to be security inspected, and the barcode information may specifically refer to waybill information of a package, including one-dimensional barcode information and two-dimensional barcode information.

Specifically, when an image to be safely checked needs to be acquired, the terminal acquires a transportation sequence identifier corresponding to the article to be safely checked and first identity information before article safety check scanning, performs safety check scanning on the article to be safely checked to obtain a scanned image corresponding to the safety check scanning, establishes a corresponding relation between the scanned image and the first identity information, caches the scanned image and the first identity information into a preset scanned image queue according to the corresponding relation, acquires second identity information after article safety check scanning, performs identity check on the second identity information according to the first identity information to determine whether the second identity information is correct, binds the second identity information and the scanned image when the second identity information passes the check, associates the corresponding transportation sequence identifier to obtain the second identity information of the image to be safely checked, acquires other first identity information from the preset scanned image queue, performs identity check on the second identity information according to obtain an identity check result, binds a scanned image code according to the identity check result, and associates the corresponding transportation sequence identifier to obtain the image to be safely checked.

For example, the embodiment is described through an application scenario, where the terminal is a security inspection machine, in a security inspection process, an article to be securely inspected is transmitted on a transmission belt in real time, and may sequentially pass through the first scanning device, the terminal, and the second scanning device, when passing through the first scanning device, the first scanning device may acquire first identity information corresponding to the article to be securely inspected and output the first identity information to the security inspection machine, when passing through the security inspection machine, the security inspection machine may scan the article to be securely inspected to obtain a scanned image, and obtain a scanned image queue according to the scanned image, the first identity information corresponding to the scanned image, and a transportation sequence identifier, where the transportation sequence identifier is acquired by an acquisition device disposed on the transmission belt. When the second scanning device passes through the second scanning device, the second scanning device can acquire second identity information corresponding to an article to be safely checked and output the second identity information to the security check machine, when the security check machine receives the second identity information output by the second scanning device, the fact that the image code binding check needs to be carried out is indicated, the security check machine can carry out the identity check on the second identity information according to the first identity information, carry out the image code binding on the scanned image according to the identity check result, and associate the corresponding transportation sequence identification to obtain the image to be safely checked.

In this embodiment, by obtaining the transportation sequence identifier, the first identity information, the scanned image, and the second identity information, the identity of the second identity information can be verified according to the first identity information, the scanned image is bound with a graphic code after the identity is verified, and the corresponding transportation sequence identifier is associated to obtain an image to be subjected to security inspection.

comparing the first identity information with the second identity information;

Specifically, the terminal compares the first identity information with the second identity information, when the second identity information corresponds to the first identity information, the identity verification is passed, the terminal directly binds the second identity information with the scanned image, and associates the corresponding transportation sequence identifier to obtain the image to be subjected to security inspection. When the second identity information does not correspond to the first identity information, the identity verification is not passed, the terminal can acquire other first identity information from a preset scanning image queue, and identity verification is continuously performed on the second identity information according to the other first identity information.

When the first identity information does not correspond to the second identity information, the terminal needs to further acquire other first identity information except the first identity information from the scanned image queue, compare the other first identity information with the second identity information, determine matched first identity information corresponding to the second identity information, perform image code binding according to the corresponding relation between the matched first identity information and the scanned image, and associate a corresponding transportation sequence identifier to obtain an image to be subjected to security inspection. Further, when the first identity information is matched with the scanning image of the unbound identity information, the second identity information at the moment is available, the identity verification is passed, the terminal can bind the scanning image of the second identity information and the unbound identity information, the corresponding transportation sequence identifier is associated to obtain an image to be safely checked, when the first identity information is matched with the scanning image of the bound identity information, the second identity information at the moment is unavailable, the identity verification is not passed, the terminal can obtain new first identity information and second identity information again, and the identity verification is carried out on the new second identity information according to the new first identity information.

It should be noted that, when the first identity information is matched with the unbound identity information scan image, it indicates that the second identity information corresponds to the unbound identity information scan image, and in the scan image queue, the first identity information corresponding to the scan image before the scan image corresponding to the first identity information is entered into the queue is excessive data, and it needs to be removed, the terminal binds the second identity information and the unbound identity information scan image to obtain an image to be checked for security, removes the first identity information before the first identity information is matched and the scan image corresponding to the previous first identity information, and updates the scan image queue.

In this embodiment, by comparing the first identity information with the second identity information, when the second identity information corresponds to the first identity information, the second identity information and the scanned image are bound, and the corresponding transportation sequence identifier is associated to obtain the image to be subjected to security inspection, thereby obtaining the image to be subjected to security inspection.

In one embodiment, before inputting an image to be security-inspected into a trained target object detection model and performing object detection on the image to be security-inspected through the trained target object detection model to obtain an object detection result, the method further includes:

acquiring a training image set carrying article image labels and an initial article detection model;

The training image set comprises a plurality of training images, and the article image labels are used for labeling article images on the training images and comprise positions of the article images on the training images and article categories corresponding to the article images. The initial article detection model refers to a detection model for which parameter adjustment has not been performed.

Specifically, before article detection, the terminal obtains a training image set carrying article image labels and an initial article detection model, training images in the training image set are sequentially input into the initial article detection model to obtain a detection result corresponding to the training images, a model loss function is calculated by comparing the detection result with the article image labels carried by the training images, parameter adjustment is carried out on the initial article detection model according to the model loss function, the step of sequentially inputting the training images into the initial article detection model is returned to obtain the detection result corresponding to the training images, and a trained target article detection model is obtained until the model loss function meets the preset loss function requirement. The preset loss function requirement may be set as needed, for example, the preset loss function requirement may be that the loss function value is smaller than a preset loss function threshold or the loss function converges, and the embodiment is not limited herein.

It should be noted that, in this embodiment, when the terminal acquires the training image set carrying the object image annotation, because the scanning image, such as an X-ray image, is mainly detected in the present application, taking into consideration blocking imaging of the X-ray image, a color difference from a conventional image, and the like, when the training image set is acquired, the training image needs to be preprocessed by means of image enhancement and the like, so as to improve an accuracy of the trained target object detection model. Furthermore, after the trained target object detection model is obtained, the terminal can optimize the trained target object detection model through various modes such as Tensor RT acceleration, lightweight network design and the like, so that the security inspection efficiency is further improved.

In this embodiment, the trained target object detection model can be obtained by training the initial object detection model according to the training image set.

In one embodiment, inputting an image to be security-inspected into a trained target object detection model, and performing object type detection on the image to be security-inspected by using the trained target object detection model to obtain a security inspection result includes:

inputting the image to be safely checked into the trained target object detection model, and performing object type detection on the image to be safely checked through the trained target object detection model to obtain an object type detection result;

and when the category confidence coefficient is smaller than the confidence coefficient threshold, generating and displaying a security inspection image judgment page according to the image to be security inspected and the contraband, receiving an image judgment result corresponding to the security inspection image judgment page, and obtaining a security inspection result according to the image judgment result.

The category confidence is the confidence that the article image in the image to be safely checked output by the trained target article detection model belongs to a certain article category. And the preset confidence threshold is used for judging whether the class confidence is credible. For example, the confidence threshold may be implemented by presetting the product management information, and the product management information includes a preset land transportation confidence threshold, an air transportation confidence threshold, a type confidence threshold, and the like. The security inspection image judgment page is a page for a security inspector to judge whether a target contraband exists in an image to be subjected to security inspection, and the image judgment result is a judgment result fed back by the security inspector. For example, the image determination result may be one of pass and non-pass.

Specifically, when the object type includes contraband, the terminal may obtain a type confidence corresponding to the contraband, determine whether a target contraband really exists in the image to be security-inspected through the type confidence, when the type confidence is greater than or equal to a preset confidence threshold, the type confidence is trusted, the terminal may trust the object type detection result, obtain a security inspection result that the target contraband exists, when the type confidence is less than the confidence threshold, the type confidence is not trusted, the terminal may further generate and display a security inspection image determination page according to the image to be security-inspected and the contraband, so that a security inspector may determine whether the target contraband really exists in the image to be security-inspected according to the displayed security inspection image determination page, feed back an image determination result corresponding to the security inspection image determination page, and after receiving the image determination result, the terminal may obtain the security inspection result according to the image determination result. When the image judgment result is qualified, the target contraband does not exist in the image to be subjected to security inspection, the terminal can obtain the security inspection result, the target contraband does not exist, when the image judgment result is forbidden, the target contraband exists in the image to be subjected to security inspection, and the terminal can obtain the security inspection result, the target contraband exists.

Further, the confidence threshold in this embodiment may correspond to the transportation mode and the type of the article, that is, different confidence thresholds may be set for different transportation modes (such as land transportation, air transportation, water transportation, and the like) and different article types, respectively, so after the category confidence corresponding to the prohibited object is obtained, a corresponding preset confidence threshold needs to be determined according to the transportation mode of the article to be safely inspected and the type of the prohibited object, so as to improve the efficiency of the security inspection. In addition, when the category confidence is smaller than the confidence threshold, the terminal also refers to preset flow screening information when generating and displaying a security inspection image judgment page according to the image to be security inspected and the contraband. The flow direction filtering information may include a focus flow direction, a specified flow direction, and the like. The important attention flow direction refers to a flow direction needing important attention, and for the articles to be subjected to safety inspection in the flow direction, important attention prompts are generated in a safety inspection image judgment page at the same time. The designated flow direction refers to a preset flow direction which needs to be processed, namely, only the image to be safely checked corresponding to the article to be safely checked in the flow direction needs to be further judged by a safety inspector, and the image to be safely checked corresponding to the article to be safely checked out in the flow direction is directly judged by a confidence threshold value to obtain a safety check result. For example, when the flow direction a is a specific flow direction, the article category of the image to be security-inspected corresponding to the article to be security-inspected corresponding to the flow direction B includes contraband, and the confidence of the category corresponding to the contraband is smaller than the confidence threshold, the security inspection result can be directly obtained as that the target contraband does not exist.

In this embodiment, by determining whether the target contraband exists on the image to be security-inspected according to the category confidence, the confidence threshold, and the image determination result, accurate determination can be achieved, and thus an accurate security inspection result is obtained.

acquiring transportation sequence information of each transportation line;

The transportation sequence information refers to transportation sequence identification information of the articles to be subjected to security inspection which are transported on a transportation line, and the target transportation line refers to the transportation line which is transporting the articles to be subjected to security inspection and comprises target contraband.

Specifically, the terminal acquires the transportation sequence information of each transportation line, compares the transportation sequence identifier with the transportation sequence information, determines a target transportation line for transporting an object to be safely inspected, which corresponds to the transportation sequence identifier, and determines a target prompt terminal corresponding to the sorting operation on the target transportation line according to a preset relationship between the target transportation line and the terminal. It should be noted that the transportation lines are all provided with the acquisition devices, and the acquisition devices can be used for acquiring transportation sequence identifiers. For example, the collecting device may be a photoelectric sensor.

In this embodiment, by acquiring the transportation sequence information of each transportation line, the target transportation line and the target prompt terminal corresponding to the sorting operation on the target transportation line can be determined according to the transportation sequence information and the transportation sequence identifier.

and when the interception feedback result is an interception error, training a trained target object detection model according to the object detection result in the interception feedback result and the image to be checked for safety.

Specifically, after the interception is completed, the terminal receives an interception feedback result corresponding to the interception prompt, when the interception feedback result is an interception error, the accuracy of the trained target article detection model is low, the terminal marks an image to be safely checked according to an article detection result in the interception feedback result, and trains the trained target article detection model according to the marked image to be safely checked so as to improve the accuracy of the trained target article detection model. Preferably, the terminal trains the trained target object detection model after receiving the preset number of interception feedback results of the interception errors, wherein the preset number can be set according to needs.

In this embodiment, after the interception feedback result corresponding to the interception prompt is received, when the interception feedback result is an interception error, the trained target object detection model is trained according to the object detection result in the interception feedback result and the image to be subjected to security inspection, so that the accurate trained target object detection model can be obtained.

In an embodiment, as shown in fig. 2, an application scenario diagram is provided to illustrate the article security inspection method of the present application, where the terminal includes a security inspection machine, a security inspection machine information interaction module, a core module, an image recognition module, an AI (Artificial Intelligence) recognition service, and an interception module, and the application scenario of the article security inspection method is as follows:

the security inspection machine obtains image code data (namely an image to be safely inspected), uploads the image code data to a security inspection machine information interaction module according to a protocol, the security inspection machine information interaction module outputs the image code data and creates a graph judging task to a core module, the core module takes out a certain number of graph judging tasks from an image recognition module to judge a graph after receiving the graph judging task, the graph judging task uses the image code data as a parameter to call a view interface in AI recognition service, the AI recognition service returns a graph identifying result (namely, the image to be safely inspected is input into a trained target object detection model, the image to be safely inspected is subjected to object type detection through the trained target object detection model to obtain a safety inspection result), after receiving the graph identifying result, the image recognition module feeds back the graph identifying result to the core module, the core module stores the graph identifying result into a graph judging queue PendingTask and performs next step processing according to an interception model.

And when the interception mode is direct interception and the security check result indicates that target contraband exists, generating an interception prompt according to the target contraband and the binding identity information, determining a target prompt terminal according to the transportation sequence identifier, and pushing the interception prompt to the target prompt terminal.

When the interception mode is a mixed interception mode (namely an AI + artificial mixed interception mode), feeding back the image code data to a security inspector for judgment, receiving an image judgment result of the security inspector, storing the image judgment result in a FinishTask, updating a state task in a data storage module, obtaining a security inspection result according to the image judgment result, generating an interception prompt according to target contraband and binding identity information when the security inspection result shows that the target contraband exists, determining a target prompt terminal according to a transportation sequence identifier, and pushing the interception prompt to the target prompt terminal. It should be noted that, when the interception mode is the hybrid interception mode, the security inspector may pre-select the security inspection machine that needs to be determined, the core module may filter the graph judging queue PendingTask according to the selection of the security inspector, and only push the graph code data on the security inspection machine that needs to be determined and is selected by the security inspector to the display interface.

In one embodiment, as shown in fig. 3, an application scenario diagram of contraband packages processed in a logistics transition is provided to illustrate the article security inspection method of the present application, where the terminal includes a security inspection machine, a contraband identification system, and an AWSM system, and the application scenario of the article security inspection method is as follows:

packages are sorted on line on a security inspection machine, the security inspection machine performs single number acquisition and X-ray photographing on fast forward to obtain security inspection data (namely, images to be subjected to security inspection) corresponding to the packages, the security inspection data are output to a contraband identification system, the contraband identification system performs contraband identification according to the security inspection data, contraband and suspected contraband are determined according to confidence threshold values, further pushing manual judgment is performed on the suspected contraband, contraband in the suspected contraband is determined (namely, images to be subjected to security inspection are input into a trained target object detection model, object type detection is performed on the images to be subjected to security inspection through the trained target object detection model to obtain security inspection results), contraband data are generated according to the contraband, the general data are pushed (and) into an AWSM system, the AWSM system issues contraband to a front-end APP (namely, when the security inspection results are targeted contraband, blocking identification prompts are generated according to the target contraband and identity information sequence, the forward-end inspection prompt is generated, the contraband identification prompts are output to the front-end APP, and abnormal transportation prompt records are generated according to the forward-end APP.

It should be understood that, although the steps in the flowcharts related to the above embodiments are shown in sequence as indicated by the arrows, the steps are not necessarily executed in sequence as indicated by the arrows. The steps are not limited to being performed in the exact order illustrated and, unless explicitly stated herein, may be performed in other orders. Moreover, at least a part of the steps in each flowchart related to the above embodiments may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least a part of the steps or stages in other steps.

In one embodiment, as shown in fig. 4, the article security inspection method of the present application is described by a schematic flowchart, and specifically includes the following steps:

step 402, acquiring a transportation sequence identifier corresponding to an article to be subjected to security inspection and first identity information before article security inspection scanning;

404, performing security inspection scanning on an article to be security inspected to obtain a scanning image corresponding to the security inspection scanning;

step 406, acquiring second identity information after the article security inspection is scanned according to the corresponding transportation sequence identifier;

step 408, comparing the first identity information with the second identity information;

step 410, when the second identity information corresponds to the first identity information, binding the second identity information and the scanned image, and associating the corresponding transportation sequence identification to obtain an image to be safely checked;

step 412, acquiring a training image set carrying item image labels and an initial item detection model;

step 414, training the initial article detection model according to the training image set to obtain a trained target article detection model;

step 416, inputting the image to be safely checked into the trained target article detection model, and performing article type detection on the image to be safely checked through the trained target article detection model to obtain an article type detection result;

step 418, when the article type in the article type detection result includes contraband, obtaining a type confidence corresponding to the contraband;

step 420, when the category confidence is greater than or equal to a preset confidence threshold, obtaining a security check result that the target contraband exists;

step 422, when the category confidence is smaller than the confidence threshold, generating and displaying a security inspection image judgment page according to the image to be security inspected and the contraband, receiving an image judgment result corresponding to the security inspection image judgment page, and obtaining a security inspection result according to the image judgment result;

step 424, when the security inspection result shows that the target contraband exists, generating an interception prompt according to the target contraband and the binding identity information;

step 426, acquiring the transportation sequence information of each transportation line;

step 428, determining a target transportation line and a target prompt terminal corresponding to the sorting operation on the target transportation line according to the transportation sequence information and the transportation sequence identifier;

step 430, pushing the interception prompt to a target prompt terminal;

step 432, receiving an interception feedback result corresponding to the interception prompt;

and 434, training the trained target object detection model according to the object detection result in the interception feedback result and the image to be checked for safety when the interception feedback result is an interception error.

In one embodiment, as shown in fig. 5, there is provided an article security inspection apparatus including: an obtaining module 502, a detecting module 504, a processing module 506 and a pushing module 508, wherein:

an obtaining module 502, configured to obtain an image to be security-inspected corresponding to an article to be security-inspected, where the image to be security-inspected carries binding identity information and a transportation sequence identifier;

the detection module 504 is configured to input the image to be security-inspected into the trained target article detection model, and perform article type detection on the image to be security-inspected through the trained target article detection model to obtain a security inspection result;

a processing module 506, configured to generate an interception prompt according to the target contraband and the binding identity information when the security check result indicates that the target contraband exists, and determine a target prompt terminal according to the transportation sequence identifier;

and the pushing module 508 is configured to push the interception prompt to the target prompt terminal.

According to the article safety inspection device, after the image to be safely inspected corresponding to the article to be safely inspected is obtained, the image to be safely inspected is input into the trained target article detection model, article type detection is carried out on the image to be safely inspected through the trained target article detection model, accurate article type identification of the image to be safely inspected can be achieved, the safety inspection effect and the safety inspection result are improved, when the safety inspection result is that target contraband exists, an interception prompt is generated according to the target contraband and the binding identity information, the target prompt terminal is determined according to the transportation sequence identification, the interception prompt is pushed to the target prompt terminal, automatic safety inspection is completed, and safety inspection efficiency can be improved.

In an embodiment, the obtaining module is further configured to obtain a transportation sequence identifier corresponding to an object to be security-inspected and first identity information before the object security inspection is scanned, perform security inspection scanning on the object to be security-inspected to obtain a scanned image corresponding to the security inspection scanning, obtain second identity information after the object security inspection scanning according to the corresponding transportation sequence identifier, perform identity verification on the second identity information according to the first identity information, perform image-code binding on the scanned image according to an identity verification result, and associate the corresponding transportation sequence identifier to obtain the image to be security-inspected.

In one embodiment, the obtaining module is further configured to compare the first identity information with the second identity information, bind the second identity information and the scanned image when the second identity information corresponds to the first identity information, and associate the corresponding transportation sequence identifier to obtain the image to be subjected to security inspection.

In one embodiment, the detection module is further configured to obtain a training image set carrying the article image labels and an initial article detection model, and train the initial article detection model according to the training image set to obtain a trained target article detection model.

In one embodiment, the detection module is further configured to input the image to be security-inspected into a trained target object detection model, perform object type detection on the image to be security-inspected through the trained target object detection model to obtain an object type detection result, obtain a type confidence level corresponding to the contraband when the object type in the object type detection result includes the contraband, obtain that the security inspection result indicates that the target contraband exists when the type confidence level is greater than or equal to a preset confidence level threshold, generate and display a security inspection image determination page according to the image to be security-inspected and the contraband when the type confidence level is less than the confidence level threshold, receive an image determination result corresponding to the security inspection image determination page, and obtain the security inspection result according to the image determination result.

In one embodiment, the processing module is further configured to obtain transportation sequence information of each transportation line, and determine a target transportation line and a target prompt terminal corresponding to the sorting operation on the target transportation line according to the transportation sequence information and the transportation sequence identifier.

In one embodiment, the detection module is further configured to receive an interception feedback result corresponding to the interception prompt, and train the trained target item detection model according to an item detection result in the interception feedback result and the image to be subjected to security inspection when the interception feedback result is an interception error.

For the specific embodiment of the article security inspection apparatus, reference may be made to the above embodiments of the article security inspection method, and details are not described herein again. The modules in the above-mentioned article security inspection apparatus may be implemented in whole or in part by software, hardware, and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 6. The computer device comprises a processor, a memory, a communication interface, a display screen and an input device which are connected through a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile 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 an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a method of security inspection of an article. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

Those skilled in the art will appreciate that the architecture shown in fig. 6 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:

and pushing the interception prompt to a target prompt terminal.

In one embodiment, the processor, when executing the computer program, further performs the steps of: the method comprises the steps of obtaining a transportation sequence identification corresponding to an object to be safely checked and first identity information before object safety check scanning, conducting safety check scanning on the object to be safely checked to obtain a scanned image corresponding to the safety check scanning, obtaining second identity information after object safety check scanning according to the corresponding transportation sequence identification, conducting identity verification on the second identity information according to the first identity information, conducting image code binding on the scanned image according to an identity verification result, and associating the corresponding transportation sequence identification to obtain an image to be safely checked.

In one embodiment, the processor, when executing the computer program, further performs the steps of: and comparing the first identity information with the second identity information, binding the second identity information with the scanned image when the second identity information corresponds to the first identity information, and associating the corresponding transportation sequence identifier to obtain the image to be subjected to security inspection.

In one embodiment, the processor, when executing the computer program, further performs the steps of: and acquiring a training image set carrying the object image label and an initial object detection model, and training the initial object detection model according to the training image set to obtain a trained target object detection model.

In one embodiment, the processor, when executing the computer program, further performs the steps of: inputting an image to be safely checked into a trained target object detection model, carrying out object type detection on the image to be safely checked through the trained target object detection model to obtain an object type detection result, when the object type in the object type detection result comprises contraband, obtaining a type confidence coefficient corresponding to the contraband, when the type confidence coefficient is larger than or equal to a preset confidence coefficient threshold value, obtaining that the safety check result is that the target contraband exists, when the type confidence coefficient is smaller than the confidence coefficient threshold value, generating and displaying a safety check image judgment page according to the image to be safely checked and the contraband, receiving an image judgment result corresponding to the safety check image judgment page, and obtaining the safety check result according to the image judgment result.

In one embodiment, the processor when executing the computer program further performs the steps of: and acquiring the transportation sequence information of each transportation line, and determining a target transportation line and a target prompt terminal corresponding to the sorting operation on the target transportation line according to the transportation sequence information and the transportation sequence identifier.

In one embodiment, the processor when executing the computer program further performs the steps of: and receiving an interception feedback result corresponding to the interception prompt, and training a trained target article detection model according to an article detection result in the interception feedback result and the image to be subjected to security inspection when the interception feedback result is an interception error.

In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:

and pushing the interception prompt to a target prompt terminal.

In one embodiment, the computer program when executed by the processor further performs the steps of: the method comprises the steps of obtaining a transportation sequence identification corresponding to an object to be safely checked and first identity information before object safety check scanning, conducting safety check scanning on the object to be safely checked to obtain a scanned image corresponding to the safety check scanning, obtaining second identity information after object safety check scanning according to the corresponding transportation sequence identification, conducting identity verification on the second identity information according to the first identity information, conducting image code binding on the scanned image according to an identity verification result, and associating the corresponding transportation sequence identification to obtain an image to be safely checked.

In one embodiment, the computer program when executed by the processor further performs the steps of: and comparing the first identity information with the second identity information, binding the second identity information with the scanned image when the second identity information corresponds to the first identity information, and associating the corresponding transportation sequence identifier to obtain the image to be subjected to security inspection.

In one embodiment, the computer program when executed by the processor further performs the steps of: and acquiring a training image set carrying the article image label and an initial article detection model, and training the initial article detection model according to the training image set to obtain a trained target article detection model.

In one embodiment, the computer program when executed by the processor further performs the steps of: inputting an image to be safely checked into a trained target object detection model, carrying out object type detection on the image to be safely checked through the trained target object detection model to obtain an object type detection result, when the object type in the object type detection result comprises contraband, obtaining a type confidence coefficient corresponding to the contraband, when the type confidence coefficient is larger than or equal to a preset confidence coefficient threshold value, obtaining that the safety check result is that the target contraband exists, when the type confidence coefficient is smaller than the confidence coefficient threshold value, generating and displaying a safety check image judgment page according to the image to be safely checked and the contraband, receiving an image judgment result corresponding to the safety check image judgment page, and obtaining the safety check result according to the image judgment result.

In one embodiment, the computer program when executed by the processor further performs the steps of: and acquiring the transportation sequence information of each transportation line, and determining a target transportation line and a target prompt terminal corresponding to the sorting operation on the target transportation line according to the transportation sequence information and the transportation sequence identifier.

In one embodiment, the computer program when executed by the processor further performs the steps of: and receiving an interception feedback result corresponding to the interception prompt, and training a trained target object detection model according to an object detection result in the interception feedback result and the image to be subjected to security inspection when the interception feedback result is an interception error.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method for security inspection of an article, the method comprising:

inputting the image to be safely checked into a trained target object detection model, and carrying out object type detection on the image to be safely checked through the trained target object detection model to obtain a safety check result;

when the security inspection result shows that target contraband exists, generating an interception prompt according to the target contraband and the binding identity information, and determining a target prompt terminal according to the transportation sequence identifier;

and pushing the interception prompt to the target prompt terminal.

2. The method of claim 1, wherein said acquiring an image to be security inspected corresponding to an item to be security inspected comprises:

carrying out safety inspection scanning on the article to be subjected to safety inspection to obtain a scanning image corresponding to the safety inspection scanning;

and carrying out identity verification on the second identity information according to the first identity information, carrying out image code binding on the scanned image according to the identity verification result, and associating the corresponding transportation sequence identification to obtain an image to be subjected to security inspection.

3. The method of claim 2, wherein the performing identity verification on the second identity information according to the first identity information, performing image-code binding on the scanned image according to the identity verification result, and associating the corresponding transportation sequence identifier to obtain an image to be subjected to security inspection comprises:

comparing the first identity information with the second identity information;

and when the second identity information corresponds to the first identity information, binding the second identity information and the scanned image, and associating the corresponding transportation sequence identifier to obtain an image to be subjected to security inspection.

4. The method according to claim 1, wherein before inputting the image to be security-inspected into a trained target object detection model and performing object type detection on the image to be security-inspected by using the trained target object detection model to obtain a security inspection result, the method further comprises:

5. The method according to claim 1, wherein the inputting the image to be security-checked into a trained target object detection model, and performing object type detection on the image to be security-checked through the trained target object detection model to obtain a security check result comprises:

inputting the image to be safely checked into a trained target article detection model, and carrying out article type detection on the image to be safely checked through the trained target article detection model to obtain an article type detection result;

and when the category confidence is smaller than a confidence threshold, generating and displaying a security inspection image judgment page according to the image to be security inspected and the contraband, receiving an image judgment result corresponding to the security inspection image judgment page, and obtaining a security inspection result according to the image judgment result.

6. The method of claim 1, wherein the determining a target prompting terminal according to the transportation sequence identifier comprises:

acquiring transportation sequence information of each transportation line;

7. The method according to claim 1, wherein after the pushing the interception prompt to the target prompt terminal, the method further comprises:

and when the interception feedback result is an interception error, training the trained target object detection model according to the object detection result in the interception feedback result and the image to be subjected to security inspection.

8. A security inspection apparatus, characterized in that the apparatus comprises:

the detection module is used for inputting the image to be safely checked into a trained target article detection model, and performing article type detection on the image to be safely checked through the trained target article detection model to obtain a safety check result;

the processing module is used for generating an interception prompt according to the target contraband and the binding identity information and determining a target prompt terminal according to the transportation sequence identifier when the security check result shows that the target contraband exists;

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.