A medicine distribution verification method and system based on image recognition

Abstract

The application provides a medicine distribution verification method and system based on image recognition. The delivery process including a continuous action sequence is forcibly guided and verified, which fundamentally eliminates cheating behaviors such as pre-stored photos and staged photos. The system verification must rely on specific actions completed in real time, which improves the cost of fraud. At the same time, the verification process is visualized, standardized, automated, evidence-based and traceable. Through the collection and fusion analysis of action information such as action video, environmental continuous parameters and medicine packaging information, the compliance, effectiveness, authenticity and continuity of the medical and pharmaceutical delivery process are ensured, and an unalterable verification record is generated, solving the credibility and traceability problem of the end verification of medicine distribution.

Description

Technical Field

[0001] This application relates to the field of image processing technology, specifically to a method and system for verifying pharmaceutical delivery based on image recognition. Background Technology

[0002] Currently, most methods for ensuring accurate and safe delivery of medicines rely on manual verification. This involves delivery personnel contacting the recipient by phone to verbally verify delivery information or the last four digits of the phone number before completing the delivery. This method heavily depends on the delivery personnel's sense of responsibility and focus. In bulk delivery scenarios, verification errors are highly likely to occur, leading to misdelivery of medicines. While some companies require delivery personnel to take delivery photos as proof, this process is usually manually triggered and independent of the delivery workflow. This results in inconsistent photo quality, non-standard content, and even instances of photos being taken afterward or falsified to pass inspections. Back-end administrators cannot verify the actual time and location of the photos in real time, indicating a significant flaw in the verification process.

[0003] Existing technologies include some systems that use scanning or simple photography for recording, but these systems typically only collect and store information in isolation, failing to deeply bind and link the delivery person's real-time location, order expectation information, on-site visual evidence, and automated system decisions for verification. Therefore, there is an urgent need for a mandatory, automated, and multi-dimensional cross-validation method and system for verifying pharmaceutical deliveries to address the technical problems of low reliability, inefficiency, and difficulty in traceability associated with manual verification. Summary of the Invention

[0004] This application aims to achieve mandatory and automated verification of pharmaceutical delivery with high credibility, high efficiency, and traceability, and provides a pharmaceutical delivery verification method and system based on image recognition.

[0005] To achieve this objective, the following technical solution is adopted in this application: A method for verifying pharmaceutical delivery based on image recognition is provided, comprising the following steps: S1, when the delivery terminal simultaneously meets the spatial location condition and the drug identity condition, and determines that the trigger condition is met, the delivery personnel are prompted to execute a delivery action sequence that is preset for the current delivery target order and contains at least two consecutive actions. At the same time, the image acquisition device of the delivery terminal is forcibly activated to record a process video covering the entire delivery action sequence, and the action information generated during this period is collected synchronously. S2, after successfully verifying the compliance of the recorded process video, extract key frame information from the process video and compare it with the target order information for consistency, and output the medical delivery verification result.

[0006] Preferably, the method of recording a process video covering the entire sequence of delivery actions for the current delivery target order includes the steps of: S11, the back-end system generates a delivery action execution prompt and pushes it to the delivery terminal based on the delivery terminal meeting the triggering condition; S12, the delivery terminal, upon receiving the delivery action execution prompt, begins recording the process video, and during the recording process, verifies the validity of the delivery action and determines whether the validity verification of the current delivery action passes. If so, proceed to step S13; If not, the current delivery action is deemed invalid and an error alert is issued, prompting the delivery person to re-execute the previously executed delivery action that is sequential with the current delivery action; S13, determine whether the delivery action sequence has been completed. If so, the recorded process video is uploaded to the backend system; If not, continue recording the process video.

[0007] Preferably, in step S12, the method for validating the current delivery action includes the following steps: S121, determine whether the stable maintenance duration set for the current delivery action has reached the preset qualified duration threshold. If so, proceed to step S122; If not, the validity check of the current delivery action is deemed to have failed, and the delivery personnel are prompted to re-execute the previously executed delivery action that is currently being executed. S122, determine whether valid process data, including drug packaging information, environmental information, and delivery personnel identification information, has been collected during the execution of the current delivery action. If so, then the execution of the current delivery action is deemed valid; If not, the validity check of the current delivery action is deemed to have failed, and the delivery personnel are prompted to re-execute the previously executed delivery action that is currently being performed.

[0008] Preferably, the environmental information includes one or more of the following: ambient light intensity waveform, ambient sound waveform, and ambient image; and the identity information includes the delivery person's iris features or facial features.

[0009] Preferably, the preset delivery action sequence includes at least two of the following actions: a first action, a second action, and a third action, which are executed sequentially; wherein, the first action is: lifting the medicine packaging of the target order and holding it toward the camera for a first duration; the second action is: moving the medicine packaging to the recipient and displaying it stably to the recipient for a second duration; and the third action is: handing the medicine packaging to the recipient and maintaining contact for a third duration.

[0010] Preferably, in step S2, the method for compliance verification of the process video successfully recorded in step S1 includes the following steps: S21, the system acquires action information collected by the delivery terminal during the execution of each action constituting the delivery action sequence, including one or more of the following: drug packaging information, environmental dynamic parameters, delivery person identification results, and recipient information; S22, the acquired action information is associated with the corresponding action to be performed, and then the validity of the action information associated with each action is checked. If all checks pass, the compliance check of the process video successfully recorded in step S1 is deemed to have passed. If any check fails, the compliance verification of the process video is deemed to have failed, and the compliance verification of the process video is terminated. The method for validating the action information associated with each action includes: Determine whether each type of process information in the action information acquired for the current inspection action meets the corresponding preset accuracy requirements and / or the corresponding preset consistency requirements. If so, the validity check of the action information is deemed to have passed; If not, the validity check of the action information is deemed to have failed.

[0011] This application also provides a pharmaceutical delivery verification system based on image recognition, including: The delivery action execution prompt module is used to generate a trigger instruction to prompt the delivery personnel to execute a delivery action sequence that includes at least two consecutive actions for the current delivery target order when the delivery terminal simultaneously meets the spatial location condition and the drug identity condition. The video recording module is connected to the delivery action execution prompt module and is used to forcibly start the image acquisition device of the delivery terminal according to the trigger command, record a process video covering the entire delivery action sequence, and simultaneously acquire the action information generated during the time period. The video compliance verification module, connected to the video recording module, is used to verify the compliance of successfully recorded process videos. The keyframe information extraction module, connected to the video compliance verification module, is used to extract keyframe information from the action information and compare it with the target order information based on the instruction that the compliance verification of the process video has passed. Then, it outputs the medical delivery verification result based on the comparison result.

[0012] This application fundamentally eliminates cheating behaviors such as pre-stored photos and staged shots by forcibly guiding and verifying a delivery process containing a coherent sequence of actions. The system verification must rely on specific actions completed in real time, which increases the cost of forgery. At the same time, it realizes the visualization, standardization, automation, evidentiary and traceability of the verification process. By collecting action information such as action videos, continuous environmental parameters and drug packaging information and performing fusion analysis, it ensures the compliance, effectiveness, authenticity and continuity of the medical device delivery process, generates tamper-proof verification records, and solves the problems of credibility and traceability in the verification of the last mile of pharmaceutical distribution. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments of this application will be briefly described below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a diagram illustrating the implementation steps of a pharmaceutical delivery verification method based on image recognition, as provided in an embodiment of this application. Detailed Implementation

[0015] The technical solution of this application will be further described below with reference to the accompanying drawings and specific embodiments.

[0016] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual images. They should not be construed as limiting the scope of this application. To better illustrate the embodiments of this application, some components in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0017] In the accompanying drawings of the embodiments of this application, the same or similar reference numerals correspond to the same or similar components. In the description of this application, it should be understood that if terms such as "upper," "lower," "left," "right," "inner," and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting this application. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0018] In the description of this application, unless otherwise expressly specified and limited, the term "connection" or similar designation indicating a connection between components should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication between two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0019] This application provides an image recognition-based method for verifying pharmaceutical delivery, such as... Figure 1 As shown, it includes: S1. When the delivery terminal simultaneously meets the spatial location condition and the drug identity condition, it is determined that the trigger condition is met. Then, the delivery personnel are prompted to execute a delivery action sequence that is preset for the current delivery object and contains at least two consecutive actions. At the same time, the image acquisition device of the delivery terminal is forcibly activated to record a process video covering the entire delivery action sequence and simultaneously collect the environmental dynamic parameters during this period. In this embodiment, the method for determining whether the delivery terminal meets the spatial location conditions is as follows: The system continuously compares the real-time GPS or BeiDou positioning coordinates reported by the delivery terminal with the electronic geofence corresponding to the expected delivery address in the target order. If the terminal's delivery coordinates remain stable within the geofence for N seconds (e.g., 3 seconds) or within a radius of 50 meters, the terminal is determined to meet the spatial location requirements.

[0020] For example, when deliveryman Xiao Wang delivers medicine to Building 5 of Community A, once Xiao Wang's handheld terminal shows that his location has entered the electronic geofence pre-set for Building 5 of Community A and remains stable, the terminal or system records that the spatial location conditions for delivery verification of the target order have been met.

[0021] The method for determining whether a delivery terminal meets the criteria for drug identification is as follows: Delivery personnel scan the unique identification code on the packaging of the medicine to be delivered using a delivery terminal. This code may be an encrypted QR code or an RFID tag. The decoded information is then compared in real time with the medicine information in the current order to be delivered. The comparison includes information such as the batch number and order number. If all information matches perfectly, the medicine is deemed to meet the identification criteria.

[0022] For example, before arriving at the recipient's door, the delivery person uses a terminal to scan a special QR code on the medicine box. The terminal parses the encrypted information within the QR code, such as the order number: ORD1234, and compares it with the order number currently being executed in the terminal's delivery task list. If the comparison is successful, the terminal or system records that the delivery meets the requirements for identifying medicine.

[0023] Step S1, the method for recording a sequence of delivery actions for the current delivery target order, includes the following steps: S11, the back-end system generates a delivery action execution prompt based on the delivery terminal meeting the trigger conditions and pushes it to the delivery terminal. The terminal screen displays or broadcasts the preset delivery action sequence as guidance. S12, the delivery terminal, using the received delivery action execution prompt as an instruction, begins recording the process video, and during the recording process, verifies the validity of the delivery action and determines whether the validity check of the current delivery action has passed. If so, proceed to step S13; If not, the current delivery action is deemed invalid and an error alert is issued, prompting the delivery person to re-execute the previously executed delivery action that is sequential with the current delivery action; In this embodiment, the preset delivery action sequence preferably includes at least two of the following actions: a first action, a second action, and a third action, which are executed sequentially. The first action is preferably: lifting the medicine packaging of the target order to be delivered and holding it towards the camera for a first duration, such as at least 3 seconds; the second action is preferably: moving the medicine packaging to the recipient and displaying it stably to the recipient for a second duration, such as 2 seconds; the third action is preferably: handing the medicine packaging to the recipient and maintaining contact for a third duration, such as 2 seconds.

[0024] The two actions are executed consecutively as follows: Assume the delivery sequence includes a first action and a second action. After meeting the recipient, the delivery person first takes out the medicine packaging, holds it up facing the terminal camera for, for example, 3 seconds, and then moves the medicine packaging horizontally to the recipient and displays it stably for, for, 2 seconds. If the horizontal movement of the medicine packaging does not exceed 3 seconds, the system determines that the two actions are consecutive; otherwise, they are considered inconsequential. Assume the delivery sequence includes the first and third actions described above. If the system determines that the time taken for the delivery person to hand the medicine packaging to the recipient after completing the first action does not exceed a preset time threshold, then the first and third actions are considered consecutive.

[0025] The following example illustrates the method for recording process video via steps S11-S13, using a delivery person's use of a delivery terminal and wearing AI smart glasses that communicate with the delivery terminal as a pharmaceutical delivery scenario: When the delivery person's location reaches the delivery terminal they are carrying and both the spatial location and the drug identification conditions are met, and upon seeing the recipient, the delivery person presses the video recording button on the delivery terminal as prompted. The terminal then begins recording the process video, and the delivery person simultaneously executes the pre-set delivery sequence for the current order. The video frames captured in real-time by the delivery terminal are synchronously transmitted to the AI ​​smart glasses worn by the delivery person, allowing them to view the live video feed on their glasses.

[0026] Assuming the preset delivery sequence for the current target order includes the preferred first and third actions mentioned above, the delivery person, following the prompts from the delivery terminal, first lifts the medicine packaging of the target order towards the camera of the delivery terminal (whether it's hanging around the neck or handheld) and holds it for a first duration, such as 3 seconds. The side of the medicine packaging facing the camera is preferably the side with the order information affixed. During these 3 seconds, the camera not only collects information about the medicine packaging, including the shipping label information and packaging shape, but also collects environmental information, including ambient light intensity waveforms, ambient sound waveforms, and environmental images. Furthermore, it instructs the AI ​​smart glasses to capture the delivery person's iris or facial features. The AI ​​smart glasses determine the delivery person's identity based on the captured features and send the identity verification result to the delivery terminal.

[0027] In the second preferred action described above, the information collected by the camera preferably includes drug packaging information (including the shipping label information on the drug packaging, a stable image of the packaging displayed to the recipient, packaging shape information, etc.), recipient information such as facial information, and environmental information. In the third preferred action described above, the information collected by the camera preferably includes images of the contact between the drug packaging and the recipient.

[0028] The method for verifying the validity of the first action during the collaborative data collection process between the delivery terminal and the AI ​​smart glasses includes the following steps: S121, determine whether the stable maintenance duration set for the current delivery action has reached the preset qualified duration threshold. If so, proceed to step S122; If not, the validity check of the current delivery action is deemed to have failed, and the delivery personnel are prompted to re-execute the previously executed delivery action that is currently being executed. For example, in the above example, the current delivery action is the preferred first action. The stability maintenance duration set for this first action is a first duration, such as 3 seconds. Assuming the acceptable duration threshold set for this first duration is 2.5 seconds, then in step S121, it is determined to be acceptable and the process proceeds to step S122. Otherwise, the validity check of the current delivery action is deemed to have failed, and the delivery personnel are prompted to re-execute the previously executed delivery action that is consecutive to the current action. For example, assuming the delivery action sequence includes a first action and a third action, the first action is executed first, followed by the third action. When the validity check of the first action fails, since there is no prior executed delivery action, the delivery personnel are prompted to re-execute the first action. When the validity check of the first action succeeds, but the validity check of the third action fails, the delivery personnel are prompted to re-execute the previously executed delivery action that is consecutive to the current delivery action and the third action, i.e., the first action.

[0029] The method for determining whether the execution of the second or third action is effective is the same as the method for determining whether the execution of the first action is effective, and will not be repeated here.

[0030] S122, determine whether valid process data, including drug packaging information, environmental information, and delivery personnel identification information, has been collected during the execution of the current delivery action. If so, then the current delivery action is deemed valid; If not, the validity check of the current delivery action is deemed to have failed, and the delivery personnel are prompted to re-execute the previously executed delivery action that is currently being performed.

[0031] After successfully recording a process video covering the delivery action sequence through the above step S1, as follows: Figure 1 As shown, the steps of the image recognition-based medical delivery verification method provided in this application are as follows: S2, after successfully verifying the compliance of the recorded process video, extracts key frame information from the process video and compares it with the target order information for consistency, and outputs the medical delivery verification result.

[0032] Specifically, in step S2, the method for compliance verification of the process video successfully recorded in step S1 includes the following steps: S21, the system acquires action information collected by the delivery terminal during the execution of each action constituting the delivery action sequence, including one or more of the following: drug packaging information, environmental dynamic parameters, delivery person identification results, and consignee personnel information; S22: Associate the acquired action information with the corresponding actions to be performed, and then verify the validity of the action information associated with each action. If all checks pass, the compliance check of the process video successfully recorded in step S1 is deemed to have passed. If any check fails, the compliance verification of the process video is deemed to have failed, and the compliance verification of the process video is terminated.

[0033] The method for validating the action information associated with each action is as follows: Determine whether each type of process information in the action information acquired for the current inspection action meets the corresponding preset accuracy requirements and / or the corresponding preset consistency requirements. If so, the validity check of the action information is deemed to have passed; If not, the validity check of the action information is deemed to have failed.

[0034] The following is a brief explanation of the validity testing process, using the validity testing of drug packaging information collected in the first step as an example: In the first step of the example above, the required accuracy for collecting drug packaging information is to clearly detect the order number, recipient information, outer packaging shape, and color on the waybill. Therefore, if only the order number can be detected but the recipient information is not identified, the accuracy of the drug packaging information collection is deemed insufficient, meaning the compliance check for the drug packaging information fails. Only after all the drug packaging information collected in the first step passes the validity check is the validity check for the drug packaging information in the first step considered passed.

[0035] During the second action, drug packaging information is also collected, but the accuracy of this data collection is set to be the same as or different from that in the first action, depending on the scenario requirements. For example, in the second action, as long as the shipping label order number can be identified, the validity check of the drug packaging information in the second action is considered to have passed.

[0036] The following example illustrates the method for verifying whether the acquired action information meets the corresponding preset consistency requirements: For example, the analysis checks whether the environmental dynamic parameters collected in the first action and those collected in the subsequent second action exhibit a smooth and natural transition over time, rather than exhibiting abrupt changes, breakpoints, or obvious mismatches, such as jumps in light color or intensity at the action transition points. If the transition is smooth and natural, the consistency check of the environmental dynamic parameters for the first and second actions is considered passed. The principle for verifying the consistency of the delivery person's identity is the same as for the environmental dynamic parameters and will not be elaborated further.

[0037] In step S2, after the compliance verification of the successfully recorded process video is passed, key frame information is extracted from the process video. For example, the waybill information of the target order is extracted from the action information collected during the execution of the first action. Then, the order number identified from the waybill information is compared with the current delivery target order number. If the comparison is successful, the medical delivery verification result is output. For example, the result may state: The target order (order number a1001) has completed compliant delivery. Or, if the consistency comparison fails, the result may state: The compliance verification of the process video is passed, but the final consistency comparison fails. Please verify manually.

[0038] This application also provides a pharmaceutical delivery verification system based on image recognition, including: The delivery action execution prompt module is used to generate a trigger instruction to prompt the delivery personnel to execute a delivery action sequence that includes at least two consecutive actions for the current delivery target order when the delivery terminal simultaneously meets the spatial location condition and the drug identity condition. The video recording module is connected to the delivery action execution prompt module and is used to forcibly start the image acquisition device of the delivery terminal according to the trigger command, record a process video covering the entire delivery action sequence, and simultaneously acquire the action information generated during the time period. The video compliance verification module, connected to the video recording module, is used to verify the compliance of successfully recorded process videos. The keyframe information extraction module, connected to the video compliance verification module, is used to extract keyframe information from the action information and compare it with the target order information based on the instruction that the compliance verification of the process video has passed. Then, it outputs the medical delivery verification result based on the comparison result.

[0039] In summary, this application fundamentally eliminates cheating behaviors such as pre-stored photos and staged shots by forcibly guiding and verifying a delivery process containing a coherent sequence of actions. The system verification must rely on specific actions completed in real time, increasing the cost of forgery. At the same time, it realizes the visualization, standardization, automation, evidentiary value, and traceability of the verification process. By collecting and integrating action information such as action videos, continuous environmental parameters, and drug packaging information, it ensures the compliance, effectiveness, authenticity, and continuity of the medical device delivery process, generates tamper-proof verification records, and solves the problems of credibility and traceability in the verification of pharmaceutical distribution.

[0040] It should be stated that the above-described specific embodiments are merely preferred embodiments and technical principles applied in this application. Those skilled in the art should understand that various modifications, equivalent substitutions, and variations can be made to this application. However, such variations, as long as they do not depart from the spirit of this application, should be within the scope of protection of this application. Furthermore, some terminology used in this application's specification and claims is not limiting but merely for ease of description.

Claims

1. A method for verifying pharmaceutical delivery based on image recognition, characterized in that the steps include... include: S1, when the delivery terminal simultaneously meets the spatial location condition and the drug identity condition, and determines that the trigger condition is met, the delivery personnel are prompted to execute a delivery action sequence that is preset for the current delivery target order and contains at least two consecutive actions. At the same time, the image acquisition device of the delivery terminal is forcibly activated to record a process video covering the entire delivery action sequence, and the action information generated during this period is collected synchronously. S2, after successfully verifying the compliance of the recorded process video, extract key frame information from the process video and compare it with the target order information for consistency, and output the medical delivery verification result.

2. The method for verifying pharmaceutical delivery based on image recognition according to claim 1, characterized in that, A method for recording a process video covering the entire sequence of delivery actions for the current delivery target order includes the steps of: S11, the back-end system generates a delivery action execution prompt and pushes it to the delivery terminal based on the delivery terminal meeting the triggering condition; S12, the delivery terminal, upon receiving the delivery action execution prompt, begins recording the process video, and during the recording process, verifies the validity of the delivery action and determines whether the validity verification of the current delivery action passes. If so, proceed to step S13; If not, the current delivery action is deemed invalid and an error alert is issued, prompting the delivery person to re-execute the previously executed delivery action that is sequential with the current delivery action; S13, determine whether the delivery action sequence has been completed. If so, the recorded process video is uploaded to the backend system; If not, continue recording the process video.

3. The method for verifying pharmaceutical delivery based on image recognition according to claim 1, characterized in that, Step S12, the method for validating the current delivery action includes the following steps: S121, determine whether the stable maintenance duration set for the current delivery action has reached the preset qualified duration threshold. If so, proceed to step S122; If not, the validity check of the current delivery action is deemed to have failed, and the delivery personnel are prompted to re-execute the previously executed delivery action that is currently being executed. S122, determine whether valid process data, including drug packaging information, environmental information, and delivery personnel identification information, has been collected during the execution of the current delivery action. If so, then the execution of the current delivery action is deemed valid; If not, the validity check of the current delivery action is deemed to have failed, and the delivery personnel are prompted to re-execute the previously executed delivery action that is currently being performed.

4. The method for verifying pharmaceutical delivery based on image recognition according to claim 3, characterized in that, The environmental information includes one or more of the following: ambient light intensity waveform, ambient sound waveform, and ambient image; the identity information includes the delivery person's iris features or facial features.

5. The method for verifying pharmaceutical delivery based on image recognition according to claim 1, characterized in that, The preset delivery action sequence includes at least two of the following actions: a first action, a second action, and a third action, which are executed sequentially. The first action is: lifting the medicine packaging of the target order and holding it towards the camera for a first duration; the second action is: moving the medicine packaging to the recipient and displaying it stably to the recipient for a second duration; and the third action is: handing the medicine packaging to the recipient and maintaining contact for a third duration.

6. The method for verifying pharmaceutical delivery based on image recognition according to claim 1, characterized in that, In step S2, the method for compliance verification of the process video successfully recorded in step S1 includes the following steps: S21, the system acquires action information collected by the delivery terminal during the execution of each action constituting the delivery action sequence, including one or more of the following: drug packaging information, environmental dynamic parameters, delivery person identification results, and recipient information; S22, the acquired action information is associated with the corresponding action to be performed, and then the validity of the action information associated with each action is checked. If all checks pass, the compliance check of the process video successfully recorded in step S1 is deemed to have passed. If any check fails, the compliance verification of the process video is deemed to have failed, and the compliance verification of the process video is terminated. The method for validating the action information associated with each action includes: Determine whether each type of process information in the action information acquired for the current inspection action meets the corresponding preset accuracy requirements and / or the corresponding preset consistency requirements. If so, the validity check of the action information is deemed to have passed; If not, the validity check of the action information is deemed to have failed.

7. A pharmaceutical delivery verification system based on image recognition, characterized in that, include: deliver The action execution prompt module is used to generate a trigger command to prompt the delivery personnel to perform a delivery action sequence that includes at least two consecutive actions for the current delivery target order when the delivery terminal simultaneously meets the spatial location conditions and the drug identity conditions. The video recording module is connected to the delivery action execution prompt module and is used to forcibly start the image acquisition device of the delivery terminal according to the trigger command, record a process video covering the entire delivery action sequence, and simultaneously acquire the action information generated during the time period. The video compliance verification module, connected to the video recording module, is used to verify the compliance of successfully recorded process videos. The keyframe information extraction module, connected to the video compliance verification module, is used to extract keyframe information from the action information and compare it with the target order information based on the instruction that the compliance verification of the process video has passed. Then, it outputs the medical delivery verification result based on the comparison result.

Images