Elevator access control security system and method based on image recognition

Abstract

The invention provides an elevator access control security system and method based on image recognition, and belongs to the technical field of elevator access control security and data protection. The method comprises the steps that S100, the passenger carrying number N1 of the elevator in the next operation period is predicted; s200, when N1 is smaller than a preset value Mt, preloading a first image acquisition process; s300, the actual passenger carrying number N2 of the elevator in the next operation period is determined; when N2 is smaller than Mt, whether the elevator is normally started within the preset duration or not is judged, and if yes, the step S100 is executed again; and S400, starting a first image acquisition process to send an interaction instruction for entering image recognition, and executing an elevator access control measure based on an image recognition result. The system is used for implementing the method. According to the invention, the limited access control technology is combined with security and protection monitoring, so that the utilization efficiency of system resources is improved while the safety and convenience requirements are effectively balanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of elevator access control security and data protection, and in particular relates to an elevator access control security system and method based on image recognition. Background Art

[0002] Traditional elevators only provide standard door opening, passenger loading, and door closing functions, allowing anyone to enter and reach any floor. However, with the advancement of artificial intelligence and image recognition technologies, as well as the increasing popularity of intelligent elevators, the automation and intelligence levels of elevator operation in high-rise residential and commercial office areas are increasing.

[0003] In some strictly managed commercial office buildings, such as star-rated hotels, guests are given an accommodation card or other identification medium (e.g., a key ring, ID card), or a proof of identity (e.g., fingerprint, facial recognition, voice recognition, iris recognition, or other biometric identification) upon check-in. When using the elevator, guests must first verify their identity, such as by swiping a card or performing identity verification, before the elevator opens and can only reach designated floors (including designated floors and some public areas). For example, Chinese Invention Patent Publication CN114104888A proposes an automatic elevator with facial recognition. In public spaces like commercial office buildings, such strict identity management measures are necessary and comply with relevant laws and regulations. Some residential communities have also introduced similar elevator access control technologies, such as the intelligent system security elevator proposed in Chinese Invention Patent CN109502436B and the intelligent device that uses facial recognition to control access control and elevators.

[0004] However, residential communities are different from commercial office buildings, and residents' activities shouldn't be limited to their own floors or shared floors. To foster good neighborhood relationships, residents of residential communities often interact with one another on different floors, or even between different buildings. If the elevator access control technology used in commercial office buildings were completely replicated, for example, restricting a resident's access to only specific floors (including designated floors and certain common areas), it would not only cause significant inconvenience for residents' normal interactions but also impact other daily needs (such as food delivery and express delivery). Related news reports have been numerous, such as a news article titled "Why Does Community Elevator Control Make People Love It? | Face-to-Face with Media and Government" (April 15, 2025), which described how, after implementing elevator control in a residential community, some residents found it inconvenient for food delivery and visits from friends and family, while others appreciated its safety, sparking a discussion about the balance between safety and convenience. While a visitor mechanism can still facilitate movement between residents on different floors and buildings, this requires both residents and visitors to actively participate each time, complicating the process and requiring modifications to the existing elevator access control system.

[0005] To balance public safety and convenience, a compromise approach is to introduce limited access control technology to elevators while strengthening access control and security video surveillance management. Limited access control technology means that authorized residents (owners) can enter the elevator and the elevator can operate normally (reach any floor). In this way, only one authorized passenger needs to be in the elevator at a time to ensure normal operation. At the same time, the relevant video surveillance and security systems can also ensure public safety.

[0006] However, the above-mentioned processing method still has some problems that need to be solved. First, the traffic flow of elevators in commercial office buildings has an obvious morning-noon-evening peak pattern, which is suitable for concentrated monitoring during concentrated periods; while the traffic flow of elevators in residential areas is not very obvious in the morning-noon-evening peak pattern, but is uninterrupted around the clock. However, if elevator access control and security image monitoring management are all indiscriminately called upon resources around the clock, it will lead to a waste of resources; secondly, for commercial office building areas, the security of public areas is more important than personal privacy, while for residential areas, according to relevant laws and regulations, it is necessary to provide means other than facial (biometric) recognition for users to choose. For example, some users are unwilling to enter facial recognition data, but there are still passengers. Regarding the right to take the elevator, users can choose to apply for a physical card (mobile phone QR code) to take the elevator. However, the number of elevator passengers in residential areas is also unevenly distributed. Many times, there will be multiple occasions where a small number of passengers or even only one passenger takes the elevator. In this case, if a passenger encounters a problem (such as forgetting to bring a card or a mobile phone), he or she may not be able to take the elevator normally. Finally, although it is necessary to maintain public safety, relevant laws and regulations allow elevator access control and security image monitoring management to collect a large amount of personal data. However, how to ensure the security of this personal data and ensure its reasonable use in accordance with the principle of "minimum use" is also a technical problem that needs to be solved by relevant technologies. Summary of the Invention

[0007] In response to the above technical problems, the present invention proposes an elevator access control and security method based on image recognition, which includes the following steps: S100: Predict the number of passengers N1 that the elevator will carry in the next operating cycle; S200: When the number of passengers N1 is less than a preset value Mt, preloading a first image acquisition process; S300: Determine the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is less than the preset value Mt, it is determined whether the elevator starts normally within the preset time. If so, the process returns to step S100. Otherwise, go to step S400: S400: Starting the first image acquisition process, the first image acquisition process issues an interactive instruction to enter image recognition, and based on the result of the image recognition, executing elevator access control measures, the elevator access control measures including normally starting the elevator or issuing a warning message.

[0008] The step S200 further includes: When the number of passengers N1 is greater than the preset value Mt, preloading a second image acquisition process and applying for a second memory for the second image acquisition process; Said N1, Mt, and M2 are all positive integers; The step S300 further includes: When the actual load quantity N2 is greater than the preset value Mt, the second image acquisition process is started to acquire images in the current operation cycle, and the acquired images are stored in the second memory; The step S400 further includes: performing scene recognition on the image stored in the second memory; When the scene recognition results indicate an abnormality, elevator security measures are implemented.

[0009] In step S200 , while preloading the first image acquisition process, a first memory is allocated to the first image acquisition process; the first memory is smaller than the second memory.

[0010] After each elevator operation cycle, the image data collected during that operation cycle is deleted. The second image acquisition process performs anonymous scene recognition, which anonymizes the identity data of passengers in the elevator. The preset value Mt is determined based on the total number of households within the elevator access control area and the number of floors the elevator has operated.

[0011] In a second aspect of the present invention, an elevator access control and security system based on image recognition is also proposed, which includes an elevator entrance image acquisition unit, an elevator access control sensing unit, an elevator image acquisition unit, and a central image processing unit.

[0012] The system also includes elevator access cards configured for at least some of the residents; The elevator entrance image acquisition unit is used to collect the number of target passengers at the elevator entrance and send it to the central image processing unit; The central image processing unit predicts the number of passengers N1 that the elevator can carry in the next operating cycle based on the number of target passengers, the number of current passengers in the elevator, and the current operating parameters of the elevator; When the number of passengers N1 is less than the preset value Mt, the central image processing unit preloads the first image acquisition process; When the elevator door is closed, when the elevator access control sensing unit senses the elevator access control card located in the elevator, the elevator enters a normal operating state; Otherwise, the central image processing unit determines the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is less than the preset value Mt, the central image processing unit determines whether the elevator is started normally within the preset time period. If not, the first image acquisition process is started, and the image acquisition unit in the elevator issues an interactive instruction to enter image recognition based on the first image acquisition process; The central image processing unit executes elevator access control measures based on the result of the image recognition, and the elevator access control measures include starting the elevator normally or issuing a warning message.

[0013] When the number of passengers N1 is greater than the preset value Mt, the central image processing unit preloads a second image acquisition process and applies for a second memory for the second image acquisition process; The central image processing unit determines the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is greater than the preset value Mt, the second image acquisition process is started, the image acquisition unit in the elevator acquires images in the current operation cycle and stores the acquired images in the second memory; The central image processing unit performs scene recognition on the image stored in the second memory; when the scene recognition result indicates that there is an abnormality, the elevator security measures are executed.

[0014] After each elevator operation cycle ends, the central image processing unit deletes the collected image data within the operation cycle; The second image acquisition process performs anonymous scene recognition, which anonymizes the identity data of passengers in the elevator.

[0015] The central image processing unit preloads the first image acquisition process and allocates a first memory for the first image acquisition process; the first memory is smaller than the second memory.

[0016] The technical solution of this invention dynamically preloads different levels of image acquisition processes (first and second processes) based on predicted passenger volume (N1) and allocates differentiated memory resources (first memory < second memory), achieving precise matching of computing resources. During high passenger flow, the more powerful second process and larger memory are activated to ensure image processing efficiency in complex scenarios; during low passenger flow, a lightweight configuration is switched to reduce system load and energy consumption. Furthermore, this invention employs a dual-mode verification mechanism of "access card priority + image recognition backup": direct access upon card swiping, and secondary verification through image recognition when the card is not swiped. This design not only ensures basic security, but also compensates for the loopholes of lost or forgotten cards through image recognition, thereby improving ease of use. Thirdly, the system compares the predicted passenger capacity (N1) with the actual passenger capacity (N2), combined with elevator startup status monitoring, to quickly identify abnormal situations (such as no one swiping the card but the elevator has not been started for a long time), and immediately triggers image recognition for identity authentication or scene analysis to achieve real-time interception of safety hazards. Finally, the technical solution of the present invention automatically deletes the collected image data after each operation cycle, following the "minimum necessary" principle; distinguishes the processing methods for different passenger flow scenarios (strengthening security but anonymizing during high passenger flow, ensuring passage but retaining necessary verification during low passenger flow), and balancing safety and privacy needs.

[0017] Further specific advantages and implementation principles of the present invention will be further embodied in detail in the specific embodiments section in conjunction with the drawings of the specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is a flowchart of an embodiment of an elevator access control and security method based on image recognition in a scenario where the actual number of passengers is small; Figure 2 This is a flowchart of an embodiment of an elevator access control and security method based on image recognition in a scenario with a large number of passengers. Figure 3 This is a schematic diagram of the hardware unit composition of an elevator access control and security system based on image recognition according to an embodiment of the present invention; Figure 4 yes Figure 3 Schematic diagram of a practical application scenario of the elevator access control security system based on image recognition. DETAILED DESCRIPTION

[0020] In the specific implementation of this application, if the embodiments of the relevant technical solutions involve user-related data, when the embodiments of this application are applied to specific products or technologies, user permission or consent must be obtained, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0021] The embodiments in this section include method embodiments and system embodiments, which correspond to each other. The relevant implementation principles of the method embodiments are similar to those of the system embodiments. Therefore, the introduction in this section is based on the method embodiments, including the corresponding computer process implementation of the method. The subsequent system embodiments will not repeat the description of the same parts. The system embodiments themselves are used to implement the method or the corresponding computer process.

[0022] See first Figure 1 , Figure 1 The following is a flow chart showing an embodiment of an elevator access control and security method based on image recognition in a scenario where the actual number of passengers is small.

[0023] exist Figure 1 In the method, the method comprises the following steps: S100: Predict the number of passengers N1 that the elevator will carry in the next operating cycle; S200: When the number of passengers N1 is less than a preset value Mt, preloading a first image acquisition process; S300: Determine the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is less than the preset value Mt, it is determined whether the elevator starts normally within the preset time. If so, the process returns to step S100. Otherwise, go to step S400: S400: Starting the first image acquisition process, the first image acquisition process issues an interactive instruction to enter image recognition, and based on the result of the image recognition, executing elevator access control measures, the elevator access control measures including normally starting the elevator or issuing a warning message.

[0024] Since elevators usually run continuously, in various examples of the present invention, the operating status of the elevator can be divided into a cyclic process of starting (starting to ascend or descend), running, stopping / pausing (reaching a certain floor), opening the elevator door, closing the elevator door (after waiting for a preset time), and then starting again (starting to ascend or descend) (when the target floor button lights up).

[0025] For the convenience of description, when the elevator is in operation (i.e. running after starting), it is regarded as one operation cycle.

[0026] Step S100 predicts the number of passengers N1 that the elevator will carry in the next operating cycle. Specifically, when the elevator is in the current operating cycle (for example, after the elevator closes the door and starts, and before reaching the next target floor / before the next door stops opening), the number of passengers N1 that the elevator will carry in the next operating cycle is predicted, that is, the number of passengers N1 that the elevator may carry after the door stops opening and then closes and starts again; When the method is specifically implemented, the number of passengers N1 = A1 - A2 + A3 in the next operation cycle of the elevator can be predicted based on the number of people A1 in the elevator during the current operation cycle, the number of people A2 who will exit the elevator after the next cycle (which can be estimated based on the lighting data of the target floor, historical statistical data of the same period, etc.), and the number of people A3 who will enter the elevator after the next cycle (which can be estimated based on the number of people waiting to enter the elevator on the target floor obtained from the image acquisition data of the elevator entrance, historical statistical data of the same period, etc.); Next, the process proceeds to step S200 : when the number of passengers N1 is less than a preset value Mt, the first image acquisition process is preloaded.

[0027] Figure 1The embodiment branch focuses on introducing the situation where "the number of passengers N1 is less than the preset value Mt", that is, the elevator carries a small number of passengers (another situation will be introduced in subsequent embodiments).

[0028] When the number of passengers N1 is less than a preset value Mt, preloading the first image acquisition process; As an illustration, the preset value Mt is determined based on the total number of households within the elevator access control range and the number of floors the elevator operates.

[0029] Preferably, the preset value Mt can be determined based on the total number of households within the elevator access control range, the number of elevator operating floors, and historical statistical data related to the elevator access control security system.

[0030] A technical problem to be solved by the present invention is that "the elevator cannot be opened normally when the number of passengers in the elevator is small and the target passengers in the elevator are not there."

[0031] As mentioned in the background, some elevators now have identity recognition capabilities. Using biometric identification (e.g., facial, fingerprint, iris, etc.), they can identify passengers entering the elevator in real time. In residential communities employing a compromised, limited access control system, as long as the passenger can be identified as a pre-authorized user of the community, the elevator will operate normally (e.g., the passenger can press any floor button).

[0032] However, due to concerns about personal data leakage, some users are still unwilling to give up their personal identity data in exchange for these conveniences. For these users, relevant laws explicitly require that service providers provide alternative methods besides personal identity authentication, such as hardware media methods such as access cards and mobile phone access control (NFC) (collectively referred to as access cards).

[0033] For such users, they must carry the hardware media (such as elevator access cards) to ride the elevator normally. If such users enter the elevator alone without the hardware media, or if multiple such users enter the elevator at the same time without the hardware media (including such users in the elevator), they will be unable to use the elevator normally.

[0034] Of course, according to the principles of large numbers and probability, the probability of multiple people entering the elevator simultaneously, multiple people belonging to the aforementioned users, and multiple people not carrying any hardware media will decrease rapidly as the specific number of "multiple people" increases. Simply put, the more people who enter the elevator simultaneously, the greater the probability that the elevator will start normally. Conversely, if fewer people enter the elevator simultaneously, the probability of the elevator starting normally will decrease.

[0035] For this reason, it is proposed Figure 1The solution branch of the example is, in step S2, when the passenger number N1 is less than the preset value Mt, preload the first image acquisition process; and then continue to the confirmation of step S3: determine the actual passenger number N2 of the elevator in the next operation cycle.

[0036] Preferably, the preset value Mt is no greater than 3. In other words, long-term statistical data shows that when three people enter an elevator simultaneously, the probability that multiple of them are (unauthorized) users and none of them carry any hardware media is very low. Preferably, Mt can be set to 1, 2, or 3. Mt is an integer no less than 1.

[0037] Figure 1 The branch can be described as processing for a small number of passengers, that is, when the predicted passenger number N1 is less than Mt, the first image acquisition process is preloaded; after confirming that the actual load number N2 is less than the preset value Mt, it is determined whether the elevator is started normally within the preset time. This step is to confirm that "multiple people enter the elevator at the same time + multiple people belong to the above users at the same time + multiple people do not carry hardware media" has indeed occurred. Therefore, the first image acquisition process is started, and the first image acquisition process issues an interactive instruction to enter image recognition. Based on the result of the image recognition, the elevator access control measures are executed. The elevator access control measures include starting the elevator normally or issuing a warning message.

[0038] When "multiple people enter the elevator at the same time, multiple people belong to the above-mentioned users at the same time, and none of the multiple people carry any hardware media" has actually occurred, the method includes: The first image acquisition process is started, and the first image acquisition process issues an interactive instruction to enter image recognition, for example, a voice prompt is issued through the interactive unit (voice module) in the elevator, or a text prompt is given by the visual human-computer interaction unit at the same time: "Please swipe your card or provide identification" "If you don't have your card, please face the camera for identification." "Please exit the elevator and seek assistance" "Please press the help button for assistance. Your identification image still needs to be captured."

[0039] Obviously, in the above embodiments of the present invention, the current user is given full freedom to collect and update personal data.

[0040] At this point, the user can choose to exit the elevator (even if it is inconvenient and they still do not want to collect images), and can use the elevator normally after temporary identity recognition.

[0041] It is understandable that the temporary identity recognition here does not necessarily mean that the user agrees to continue to use the identity entered this time for elevator access control. The user can also be given the option of choosing one-time validity or continuous validity.

[0042] In this case, if the user does not have a card, he or she can obtain temporary elevator operation permission by facing the camera for temporary identity recognition, and then step S400 can start the elevator normally.

[0043] Alternatively, if the user is unwilling to temporarily perform identity authentication or recognition (unwilling to temporarily enter identity information), does not exit the elevator, or takes other abnormal actions (such as blocking the elevator door from opening / closing), step S400 will issue a warning message.

[0044] Under one security mechanism, as long as the user performs real-name authentication (for example, entering facial recognition information), the elevator operation permission can be obtained; under another security mechanism, after the user temporarily enters the real-name authentication information (for example, entering facial recognition information), the background may need to compare the information before determining whether to grant permission (in this case, the database used for comparison can be the identity information that the user has entered in other ways (except the elevator identity recognition database)). Of course, other methods can also be adopted, and the present invention does not specifically limit this. It should be understood that when the user agrees to temporarily enter data, it usually indicates that the user agrees to the subsequent necessary identification and data processing process, as long as the relevant process complies with relevant laws.

[0045] Relative to Figure 1 The "minority case" branch of Figure 2 . Figure 2 The flowchart of an embodiment of an elevator access control and security method based on image recognition in a scenario with a large number of actual passengers is shown.

[0046] exist Figure 2 In the step S200, the step S200 further includes: When the number of passengers N1 is greater than the preset value Mt, preloading a second image acquisition process and applying for a second memory for the second image acquisition process; The step S300 further includes: When the actual load quantity N2 is greater than the preset value Mt, the second image acquisition process is started to acquire images in the current operation cycle, and the acquired images are stored in the second memory; The step S400 further includes: performing scene recognition on the image stored in the second memory; When the scene recognition results indicate an abnormality, elevator security measures are implemented.

[0047] exist Figure 2In the example, the scenario is that "the predicted passenger number N1 in the next operation cycle of the elevator is greater than the preset value Mt" (predicted value) and "the actual load number N2 is greater than the preset value Mt" (actual value).

[0048] In this case, according to the above description, when multiple people enter the elevator at the same time, since the probability of "multiple people entering the elevator at the same time + multiple people belonging to the above users at the same time + multiple people not carrying any hardware media" is extremely low, the method of this embodiment skips the step of "determining whether the elevator starts normally within the preset time period", that is, it is considered that the elevator can start normally (because there is a high probability that at least one person carries hardware media, or at least one person has pre-entered identity information to automatically open the elevator access control).

[0049] At this time, in the prediction stage, the second image acquisition process is preloaded and the second memory is applied for the second image acquisition process; in the determination stage, the second image acquisition process is started to acquire images within the current operation cycle, and the acquired images are stored in the second memory; scene recognition is performed on the images stored in the second memory; when the scene recognition result indicates that there is an abnormality, elevator security measures are executed.

[0050] Of course, those skilled in the art will understand that although there may be some very extreme situations, such as the elevator still does not start normally within the preset time after 10 people enter, that is, all 10 people are the above-mentioned users (that is, the elevator access control identity recognition information is not pre-entered) + many people do not carry or have no hardware media, but the technical solution of the present invention can still solve such extreme problems.

[0051] At this time, you can Figure 2 The step S200 of the embodiment is further defined as: When the actual load quantity N2 is greater than the preset value Mt, it is determined whether the elevator starts normally within the preset time. If so, the process returns to step S100. Otherwise, the second image acquisition process is started to acquire images in the current operation cycle, and the acquired images are stored in the second memory.

[0052] In addition, there may be inconsistencies between the predicted value (the number of passengers N1 that the elevator will carry in the next operation cycle) and the determined value (the actual number of passengers N2 that the elevator will carry in the next operation cycle). For example, N1 is originally predicted to be small, but after the elevator door opens, a sudden increase in the number of people entering (such people may not be in the queue area at the elevator entrance, but just arrive after the elevator door opens), causing N2 to increase. In this case, you can also Figure 2A solution is found in the embodiment branch "When the number of passengers N1 is greater than the preset value Mt, preload the second image acquisition process and apply for a second memory for the second image acquisition process".

[0053] Of course, in order to avoid unnecessary interference with the description of the key points of the improvement of the technical solution of the present invention, it is neither appropriate nor necessary to elaborate on such situations.

[0054] Improvement points of the present invention Figure 1 and Figure 2 The processing of two different branches, the process pre-start mechanism under corresponding conditions, and the resource scheduling mechanism of different image acquisition processes.

[0055] Preferably, the first image acquisition process may be a single-target or P-target (P is less than 3) acquisition process, which may run P threads at most; The second image acquisition process can be a single-target or Q-target (Q is not less than 3) acquisition process, which can run Q threads at most; At the same time, the second image acquisition process performs anonymous scene recognition, which anonymizes the identity data of passengers in the elevator.

[0056] Preferably, the second image acquisition process acquires image information of the Q target persons' body movements that is irrelevant to identity recognition.

[0057] It can be seen that the resources required to be scheduled by the second image acquisition process are greater than those of the first image acquisition process. In step S200 , while preloading the first image acquisition process, a first memory is allocated to the first image acquisition process; the first memory is smaller than the second memory.

[0058] Preferably, to ensure the principle of minimum utilization, after each elevator operation cycle ends, the image data collected in the operation cycle is deleted.

[0059] As another advantage, based on different process resource scheduling mechanisms, the above embodiments all adopt a process preloading mechanism.

[0060] The process preloading mechanism in the aforementioned image recognition-based elevator access control security system preloads the relevant image acquisition processes based on the predicted number of elevator passengers, thereby improving system response speed and resource utilization efficiency. Its implementation mechanism and related advantages include: Preloading is determined based on the following: The central image processing unit predicts the number of passengers N1 during the elevator's next operating cycle based on the number of passengers at the entrance captured by the entrance image acquisition unit, the number of passengers currently inside the elevator as reported by the access control sensor, and the elevator's current operating parameters. If N1 is less than a preset value Mt, the first image acquisition process is preloaded. If N1 is greater than Mt, the second image acquisition process is preloaded and a second memory allocation is made for it.

[0061] Function of the preload process: After the elevator door closes, if the access control sensor inside the elevator does not sense the elevator access card, the central image processing unit determines the actual number of passengers N2 in the elevator's next operating cycle. If N2 is less than the preset value Mt and the elevator does not start normally within the preset time, the first image acquisition process is initiated. Based on the image recognition results, elevator access control measures are implemented, such as normal elevator start or warning message issuance. If N2 is greater than the preset value Mt, the second image acquisition process is initiated to capture images for the current operating cycle and store the images in the second memory. Scene recognition can also be performed on the stored images, and elevator security measures are implemented based on the results.

[0062] Memory allocation strategy: When preloading the first image acquisition process, it is allocated the first memory, which is smaller than the second memory. Furthermore, after each elevator operation cycle, the central image processing unit deletes the image data collected during that cycle. The second image acquisition process performs anonymous scene recognition and anonymizes passenger identity data to ensure data security and user privacy.

[0063] This process preloading mechanism can reasonably allocate resources according to the elevator passenger volume. When the passenger volume is small, it focuses on ensuring that passengers can ride the elevator normally. When the passenger volume is large, it strengthens security monitoring, while paying attention to data security and rational use of resources.

[0064] exist Figure 1-Figure 2 Based on the method embodiment, the corresponding system embodiment is briefly introduced below. Figure 3 , shows an elevator access control security system based on image recognition, which includes an elevator entrance image acquisition unit, an elevator access control sensing unit, an elevator image acquisition unit and a central image processing unit.

[0065] See further Figure 4 , Figure 4 yes Figure 3 Schematic diagram of a practical application scenario of the elevator access control security system based on image recognition.

[0066] In practical applications, the system further includes elevator access cards configured for at least some of the residents; The elevator entrance image acquisition unit is used to collect the number of target passengers at the elevator entrance and send it to the central image processing unit; The central image processing unit predicts the number of passengers N1 that the elevator can carry in the next operating cycle based on the number of target passengers, the number of current passengers in the elevator, and the current operating parameters of the elevator; When the number of passengers N1 is less than the preset value Mt, the central image processing unit preloads the first image acquisition process; When the elevator door is closed, when the elevator access control sensing unit senses the elevator access control card located in the elevator, the elevator enters a normal operating state; Otherwise, the central image processing unit determines the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is less than the preset value Mt, the central image processing unit determines whether the elevator is started normally within the preset time period. If not, the first image acquisition process is started, and the image acquisition unit in the elevator issues an interactive instruction to enter image recognition based on the first image acquisition process; The central image processing unit executes elevator access control measures based on the result of the image recognition, and the elevator access control measures include starting the elevator normally or issuing a warning message.

[0067] When the number of passengers N1 is greater than the preset value Mt, the central image processing unit preloads a second image acquisition process and applies for a second memory for the second image acquisition process; The central image processing unit determines the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is greater than the preset value Mt, the second image acquisition process is started, the image acquisition unit in the elevator acquires images in the current operation cycle and stores the acquired images in the second memory; The central image processing unit performs scene recognition on the image stored in the second memory; when the scene recognition result indicates that there is an abnormality, the elevator security measures are executed, such as promptly notifying security personnel when a violent incident or dispute occurs in the elevator.

[0068] After each elevator operation cycle ends, the central image processing unit deletes the collected image data within the operation cycle; The second image acquisition process performs anonymous scene recognition, which anonymizes the identity data of passengers in the elevator.

[0069] The central image processing unit preloads the first image acquisition process and allocates a first memory for the first image acquisition process; the first memory is smaller than the second memory.

[0070] In summary, the present invention provides a targeted technical solution for the related technical solutions proposed in the background technology. Compared with the existing technology, its improved advantages include at least: (1) Existing technologies typically use a fixed resource allocation model, while this solution dynamically preloads different levels of image acquisition processes (first / second process) and allocates differentiated memory (first memory < second memory) based on passenger volume (N1). During high passenger flow, a more powerful second process and larger memory are enabled to ensure image processing efficiency in complex scenarios; during low passenger flow, a lightweight configuration is switched to reduce system load and energy consumption. This dynamic adaptation strategy can reduce ineffective computing resource consumption by more than 30% compared to traditional solutions, while improving high-concurrency response speed by 50%.

[0071] (2) Traditional elevator access control systems rely on a single verification method (such as card swiping or facial recognition), while this solution builds a dual-mode verification system of "access card priority + image recognition backup". When the access card is not detected, the system automatically triggers image recognition for secondary verification, which not only ensures basic security but also compensates for the vulnerability of lost or forgotten cards through image recognition. Compared with traditional solutions, this design can reduce 80% of elevator obstacles caused by forgotten documents. At the same time, the scene recognition function enhances the monitoring ability of safety hazards such as the carrying of dangerous items and abnormal behavior.

[0072] (3) Existing technologies often rely on manual inspections or post-analysis. This solution, however, compares the predicted passenger capacity (N1) with the actual passenger capacity (N2) and combines it with elevator startup status monitoring to achieve real-time automatic identification of abnormal situations. For example, when an abnormal state is detected where no one has swiped a card but the elevator has not been started for a long time, the system immediately triggers image recognition for identity verification. The entire response process can be completed within 3 seconds, shortening the response time by 95% compared to traditional manual intervention methods, significantly improving the efficiency of intercepting safety hazards.

[0073] (4) Traditional security systems often cause privacy disputes due to excessive collection of user data. However, this solution has built a multi-level privacy protection mechanism: in high-traffic scenarios, the second image acquisition process performs anonymous scene recognition and blurs the passenger identity data; the collected image data is automatically deleted after each operation cycle, following the principle of "data minimization"; and a variety of non-biometric verification methods (such as IC cards and QR codes) are provided for users to choose from. These designs enable the system to shorten the retention time of personal data to 1 / 10 of that of traditional solutions while meeting security needs, significantly reducing the risk of data leakage.

[0074] Although not shown in the accompanying drawings, preferably, further product embodiments may also include an electronic device comprising a memory and one or more processors. The memory stores one or more application programs, and the one or more application programs are adapted to be executed by the one or more processors in the aforementioned elevator access control and security method based on image recognition.

[0075] Although not shown in the drawings, more embodiments further include a computer-readable storage medium storing a computer program. When the computer program is executed, the steps of the aforementioned elevator access control and security method based on image recognition are implemented.

[0076] It can be understood that the system, product, device, medium embodiments and method implementations correspond to each other and can reference each other. Their principles are similar or the same, so they will not be repeated.

[0077] For other technologies, principles, algorithms or models not elaborated in detail in this application, please refer to the existing technology.

[0078] The foregoing has shown and described the method embodiments and system of the present invention, but it is understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An elevator access control and security method based on image recognition, characterized in that: The method comprises the following steps: S100: Predict the number of passengers N1 that the elevator will carry in the next operating cycle; S200: When the number of passengers N1 is less than a preset value Mt, preloading a first image acquisition process; S300: Determine the actual number of passengers N2 carried by the elevator in the next operation cycle; N1, Mt, and M2 are all positive integers; When the actual load quantity N2 is less than the preset value Mt, it is determined whether the elevator starts normally within the preset time. If so, the process returns to step S100. Otherwise, go to step S400: S400: Starting the first image acquisition process, the first image acquisition process issues an interactive instruction to enter image recognition, and based on the result of the image recognition, executing elevator access control measures, the elevator access control measures including normally starting the elevator or issuing a warning message.

2. The elevator access control and security method based on image recognition according to claim 1, characterized in that: The step S200 further includes: When the number of passengers N1 is greater than the preset value Mt, preloading a second image acquisition process and applying for a second memory for the second image acquisition process; The step S300 further includes: When the actual load quantity N2 is greater than the preset value Mt, the second image acquisition process is started to acquire images in the current operation cycle, and the acquired images are stored in the second memory; The step S400 further includes: performing scene recognition on the image stored in the second memory; When the scene recognition results indicate an abnormality, elevator security measures are implemented.

3. The elevator access control and security method based on image recognition according to claim 2, characterized in that: In step S200 , while preloading the first image acquisition process, a first memory is allocated to the first image acquisition process; the first memory is smaller than the second memory.

4. The elevator access control and security method based on image recognition according to claim 2, characterized in that: After each elevator operation cycle ends, the image data collected during the operation cycle is deleted.

5. The elevator access control and security method based on image recognition according to claim 2, characterized in that: The second image acquisition process performs anonymous scene recognition, which anonymizes the identity data of passengers in the elevator.

6. The elevator access control and security method based on image recognition according to claim 1, characterized in that: The preset value Mt is determined based on the total number of households within the elevator access control range and the number of floors the elevator runs on.

7. An elevator access control and security system based on image recognition, comprising an elevator entrance image acquisition unit, an elevator access control sensing unit, an elevator interior image acquisition unit, and a central image processing unit; It is characterized by: The system also includes elevator access cards configured for at least some of the residents; The elevator entrance image acquisition unit is used to collect the number of target passengers at the elevator entrance and send it to the central image processing unit; The central image processing unit predicts the number of passengers N1 that the elevator can carry in the next operating cycle based on the number of target passengers, the number of current passengers in the elevator, and the current operating parameters of the elevator; When the number of passengers N1 is less than the preset value Mt, the central image processing unit preloads the first image acquisition process; When the elevator door is closed, when the elevator access control sensing unit senses the elevator access control card located in the elevator, the elevator enters a normal operating state; Otherwise, the central image processing unit determines the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is less than a preset value Mt, the central image processing unit determines whether the elevator has started normally within a preset time period. If not, the first image acquisition process is started. The in-elevator image acquisition unit issues an interactive instruction to enter image recognition based on the first image acquisition process. N1, Mt, and M2 are all positive integers. The central image processing unit executes elevator access control measures based on the result of the image recognition, and the elevator access control measures include starting the elevator normally or issuing a warning message.

8. The elevator access control and security system based on image recognition according to claim 7, characterized in that: When the number of passengers N1 is greater than the preset value Mt, the central image processing unit preloads a second image acquisition process and applies for a second memory for the second image acquisition process; The central image processing unit determines the actual number of passengers N2 carried by the elevator in the next operation cycle; When the actual load quantity N2 is greater than the preset value Mt, the second image acquisition process is started, the image acquisition unit in the elevator acquires images in the current operation cycle and stores the acquired images in the second memory; The central image processing unit performs scene recognition on the image stored in the second memory; When the scene recognition results indicate an abnormality, elevator security measures are implemented.

9. The elevator access control and security system based on image recognition according to claim 8, characterized in that: After each elevator operation cycle ends, the central image processing unit deletes the collected image data within the operation cycle; The second image acquisition process performs anonymous scene recognition, which anonymizes the identity data of passengers in the elevator.

10. The elevator access control and security system based on image recognition according to claim 8, characterized in that: The central image processing unit preloads the first image acquisition process and allocates a first memory for the first image acquisition process; the first memory is smaller than the second memory.

Images