Non-business time personnel access vault detection method and related device

Abstract

The invention discloses a method for detecting personnel entering and exiting a vault in non-business time and a related device, which can be applied to the field of artificial intelligence or finance, and can be used for detecting personnel entering and exiting the vault in non-business time by identifying the brightness of a monitoring image and determining the brightness level corresponding to the brightness after the monitoring image is acquired in the non-business time of the vault. Under the condition that the brightness of the monitoring image is lower than the preset normal brightness threshold value, the brightness enhancement model corresponding to the brightness level of the monitoring image is called to enhance the brightness of the monitoring image to obtain the to-be-detected image, compared with the condition that unified brightness enhancement is adopted for all dark light images, the brightness enhancement effect on the monitoring image is improved, and the detection efficiency is improved. On the basis, the to-be-detected image is input into the vault personnel target detection model for target detection, the success rate of personnel identification under different brightness conditions is improved, manual parameters are not needed in the whole process, and the detection efficiency and accuracy of personnel entering and exiting from the vault in non-business time are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method and related device for detecting people entering and leaving a safe during non-business hours. Background Art

[0002] The safe is a very important security area. In order to protect property safety and reduce operational risks, bank staff must follow strict operating standards and approval requirements for entry and exit of the safe. People are strictly prohibited from entering and exiting the safe during non-business hours.

[0003] At present, the detection method of entering and leaving the vault during non-business hours mainly relies on manual spot checks of vault entry and exit records or monitoring. This method has disadvantages such as high labor costs, untimely detection, and easy omissions. In addition, the non-business hours of the vault span different time periods such as day and night, and the light brightness in the vault varies in different time periods. Some general computer vision recognition technologies are difficult to accurately identify vault scene images in different time periods. Summary of the invention

[0004] In view of the above problems, the present application provides a method and related device for detecting people entering and leaving the safe during non-business hours, so as to improve the efficiency and accuracy of detecting people entering and leaving the safe during non-business hours. The specific scheme is as follows:

[0005] The first aspect of the present application provides a method for detecting people entering and exiting a safe during non-business hours, comprising:

[0006] Obtain surveillance images during the vault's non-business hours;

[0007] Identifying the brightness of the surveillance image and determining a brightness level corresponding to the brightness;

[0008] If the brightness is lower than a preset normal brightness threshold, a brightness enhancement model corresponding to the brightness level of the monitoring image is called to perform brightness enhancement on the monitoring image to obtain an image to be detected;

[0009] Inputting the image to be detected into the vault personnel target detection model to obtain the target detection confidence output by the vault personnel target detection model;

[0010] If the target detection confidence is greater than the confidence threshold, it is determined that the surveillance image indicates that a person has entered or exited the safe.

[0011] In a possible implementation, the calling of a brightness enhancement model corresponding to the brightness level of the surveillance image to perform brightness enhancement on the surveillance image to obtain the image to be detected includes:

[0012] If the brightness level of the monitoring image is a first brightness, calling a first brightness enhancement model to perform brightness enhancement on the monitoring image to obtain an image to be detected, wherein the first brightness enhancement model is trained using training samples annotated with the first brightness and normal brightness;

[0013] If the brightness level of the surveillance image is the second brightness, call the second brightness enhancement model to perform brightness enhancement on the surveillance image to obtain the image to be detected, the second brightness enhancement model is trained using training samples labeled with the second brightness and the normal brightness, the first brightness is greater than the second brightness and the first brightness is less than the normal brightness.

[0014] In a possible implementation, the confidence threshold is positively correlated with the brightness of the monitoring image.

[0015] In a possible implementation, the method for detecting personnel entering and exiting the vault during non-business hours further includes:

[0016] If the brightness level of the monitoring image is normal brightness, the monitoring image is determined as the image to be detected, and the step of inputting the image to be detected into the safe personnel target detection model is performed.

[0017] In a possible implementation, after determining that the surveillance image indicates that a person has entered or exited the safe, the method for detecting a person entering or exiting the safe during non-business hours further includes:

[0018] Sending warning information to the cloud platform so that the cloud platform forwards the warning information to a terminal held by at least one target manager, the warning information including the monitoring image, the acquisition time of the monitoring image and the vault identification.

[0019] A second aspect of the present application provides a non-business hours personnel entry and exit detection device, comprising:

[0020] An image acquisition unit, used to acquire surveillance images during non-business hours of the vault;

[0021] A brightness recognition unit, used to recognize the brightness of the monitoring image and determine the brightness level corresponding to the brightness;

[0022] A brightness enhancement unit, configured to, if the brightness is lower than a preset normal brightness threshold, call a brightness enhancement model corresponding to the brightness level of the monitoring image to perform brightness enhancement on the monitoring image to obtain an image to be detected;

[0023] A target detection unit, used for inputting the image to be detected into a vault personnel target detection model to obtain the target detection confidence output by the vault personnel target detection model;

[0024] The result determination unit is used to determine that the monitoring image indicates that a person has entered or exited the safe if the target detection confidence is greater than a confidence threshold.

[0025] A third aspect of the present application provides an edge computing device, comprising at least one processor and a memory connected to the processor, wherein:

[0026] The memory is used to store computer programs;

[0027] The processor is used to execute the computer program so that the edge computing device can implement the method for detecting people entering and exiting the vault during non-business hours according to the first aspect or any implementation method of the first aspect.

[0028] A fourth aspect of the present application provides a non-business hours personnel entry and exit vault detection system, comprising: a cloud platform and at least one edge computing device;

[0029] The cloud platform remotely communicates with at least one edge computing device respectively;

[0030] The edge computing device is communicatively connected to at least one surveillance camera in the safe, and is used to implement the method for detecting people entering and exiting the safe during non-business hours according to the first aspect or any implementation method of the first aspect.

[0031] In a fifth aspect, the present application provides a computer program product, including computer-readable instructions. When the computer-readable instructions are executed on an edge computing device, the edge computing device implements the method for detecting people entering and exiting a safe during non-business hours according to the first aspect or any implementation method of the first aspect.

[0032] The sixth aspect of the present application provides a computer storage medium, characterized in that the storage medium carries one or more computer programs. When the one or more computer programs are executed by an edge computing device, the edge computing device can implement the method for detecting people entering and exiting the vault during non-business hours according to the first aspect or any implementation method of the first aspect.

[0033] Through the above technical scheme, the present application provides a method and related device for detecting people entering and leaving the safe during non-business hours. After obtaining a surveillance image during the safe's non-business hours, the brightness of the surveillance image is identified and the brightness level corresponding to the brightness is determined. When the brightness of the surveillance image is lower than a preset normal brightness threshold, a brightness enhancement model corresponding to the brightness level of the surveillance image is called to perform brightness enhancement on the surveillance image to obtain an image to be detected. Compared with using a uniform brightness enhancement for all dark light images, the brightness enhancement effect of the surveillance image is improved. On this basis, the image to be detected is input into the safe's personnel target detection model for target detection, thereby improving the success rate of personnel identification under different brightness conditions. The entire process does not require manual parameters, thereby improving the efficiency and accuracy of detecting people entering and leaving the safe during non-business hours. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0035] Figure 1 A system architecture diagram provided for an embodiment of the present application;

[0036] Figure 2 A schematic diagram of the structure of an edge computing device provided in an embodiment of the present application;

[0037] Figure 3 A schematic diagram of the structure of a cloud platform provided in an embodiment of the present application;

[0038] Figure 4 A flow chart of a method for detecting personnel entering and exiting a vault during non-business hours provided in an embodiment of the present application;

[0039] Figure 5 A flow chart of another method for detecting personnel entering and exiting a safe during non-business hours provided in an embodiment of the present application;

[0040] Figure 6 A schematic diagram of the structure of a non-business hours personnel entry and exit detection device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] In the following description, in order to ensure the accuracy of the citation and the fluency of reading, the key technical terms, abbreviations or abbreviations involved in the text are summarized and explained as follows:

[0042] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0043] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0044] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0045] See also Figure 1 , Figure 1 A schematic diagram of a system architecture is shown. The system may include at least one edge computing device 100, a cloud platform 200, and a surveillance camera 300. The cloud platform 200 may include one or more servers ( Figure 1 A server is included as an example for explanation).

[0046] The cloud platform 200 remotely communicates with at least one edge computing device 100, and the edge computing device 100 is communicatively connected with at least one surveillance camera 300 in the vault.

[0047] Among them, an application can be installed on the edge computing device 100. The edge computing device 100 is started during the non-business hours of the safe, and sends warning information to the cloud platform 200 when it detects that someone is entering or leaving the safe. The cloud platform 200 can issue a configuration file or software upgrade instruction to the edge computing device 100. The edge computing device 100 can update the non-business hours, detection frequency, etc. of the safe according to the configuration file, and can also respond to the software upgrade instruction and perform software upgrades based on the upgrade package issued by the cloud platform 200.

[0048] It should be understood that in some optional implementations, the edge computing device 100 can also complete the action of obtaining the processing result based on the received parameters (configuration information, upgrade instructions, etc.) by itself, or directly send warning information to the target management personnel without the need for the cooperation of the cloud platform 200, and the embodiments of the present application are not limited.

[0049] Next describe Figure 1 The product form of the edge computing device 100;

[0050] The edge computing device 100 in the embodiment of the present application can be a network device with edge computing function (such as a smart router, a smart switch, a smart gateway, etc.), a smart camera, an edge server, a PLC (programmable logic controller), a microcontroller, a mobile device (such as a smart phone, a tablet computer, etc.), etc., and the embodiment of the present application does not impose any restrictions on this.

[0051] like Figure 2 As shown, the edge computing device 100 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 101, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 102 or a program loaded from a storage device 108 to a random access memory (RAM) 103. When the edge computing device 100 is powered on, the RAM 103 also stores various programs and data required for the operation of the edge computing device 100. The processing device 101, the ROM 102, and the RAM 103 are connected to each other via a bus 104. An input / output (I / O) interface 105 is also connected to the bus 104.

[0052] Typically, the following devices may be connected to the I / O interface 105: an input device 106, an output device 107, a storage device 108 including, for example, a memory card, a hard disk, etc., and a communication device 109. The communication device 109 may allow the edge computing device 100 to communicate with other devices wirelessly or by wire to exchange data. Figure 2 The edge computing device 100 having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0053] Next describe Figure 1 The product form of Zhongyun Platform 200;

[0054] Figure 3 A schematic diagram of the structure of a cloud platform 200 is provided, such as Figure 3 As shown, the cloud platform 200 includes a bus 201, a processor 202, a communication interface 203, and a memory 204. The processor 202, the memory 204, and the communication interface 203 communicate with each other via the bus 201.

[0055] The bus 201 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0056] The processor 202 may be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0057] The memory 204 may include a volatile memory, such as a random access memory (RAM). The memory 204 may also include a nonvolatile memory, such as a read only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0058] Among them, the memory 204 can be used to store software codes related to the method for detecting people entering and exiting the safe during non-business hours, and the processor 202 can execute the steps of the chip's method for detecting people entering and exiting the safe during non-business hours, and can also schedule other units to implement corresponding functions.

[0059] It should be understood that the above-mentioned edge computing device 100 and cloud platform 200 can be centralized or distributed devices, and the processors in the above-mentioned edge computing device 100 and cloud platform 200 can be hardware circuits (such as application specific integrated circuits (ASIC), field programmable gate arrays (FPGA), general-purpose processors, digital signal processors (digital signal processing, DSP), microprocessors or microcontrollers, etc.), or a combination of these hardware circuits. For example, the processor can be a hardware system with an instruction execution function, such as a CPU, DSP, etc., or a hardware system without an instruction execution function, such as an ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without an instruction execution function and hardware systems with an instruction execution function.

[0060] The embodiment of the present application provides a method for detecting people entering and exiting a safe during non-business hours, which is applied to an edge computing device in a safe. The method for detecting people entering and exiting a safe during non-business hours according to the embodiment of the present application is described in detail below in conjunction with the accompanying drawings.

[0061] Reference Figure 4 , Figure 4 A flow chart of a method for detecting personnel entering and exiting a vault during non-business hours provided in an embodiment of the present application is shown as follows: Figure 4 As shown, a method for detecting people entering and exiting a safe during non-business hours provided in an embodiment of the present application may include steps 401 to 403, and these steps are described in detail below.

[0062] 401: Obtaining surveillance images during the vault's non-business hours;

[0063] The non-business hours for the vault are pre-set and can be updated.

[0064] The edge computing device is connected to at least one surveillance camera in the vault to obtain surveillance video data of the vault collected by the surveillance camera, and performs frame extraction preprocessing on the surveillance video data according to a preset frame rate to obtain a surveillance image.

[0065] 402: Identify the brightness of the monitoring image and determine the brightness level corresponding to the brightness;

[0066] Since the non-business hours of the safe cover different time periods, the illumination of the scene varies greatly. In order to facilitate subsequent brightness enhancement according to the brightness level of the surveillance image, it is necessary to identify the brightness of the surveillance image.

[0067] Exemplarily, the brightness of the monitoring image is identified according to the following formula:

[0068]

[0069] Among them, I(x i ,y i ) represents the coordinates of pixel i in the monitoring image, N is the total number of pixels in the monitoring image, G(I(x i ,y i )) represents grayscale processing of pixel i, Norm represents normalization calculation, maps the value range to 0~1, and calculates T as the normalized brightness value of the monitoring image.

[0070] The brightness level includes at least three brightness levels, that is, the brightness level includes normal brightness and at least two abnormal brightness levels. The finer the division of abnormal brightness, i.e. low brightness, the better the subsequent brightness enhancement effect, but the greater the workload of model training. In actual applications, the brightness levels can be divided according to application requirements.

[0071] Exemplarily, the brightness level includes: first brightness, second brightness and normal brightness, wherein the first brightness is greater than the second brightness and the first brightness is less than the normal brightness. The brightness level is divided according to the image brightness as follows:

[0072]

[0073] Among them, Θ1 and Θ2 are the first threshold and the second threshold respectively, normal is the normal brightness, low1 is the first brightness, low2 is the second brightness, and L is the brightness level division result of the monitoring image.

[0074] If the brightness level of the monitoring image is normal brightness, no brightness enhancement is required, the monitoring image is determined as an image to be detected, and step 404 is executed.

[0075] 403: If the brightness is lower than the preset normal brightness threshold, a brightness enhancement model corresponding to the brightness level of the monitoring image is called to enhance the brightness of the monitoring image to obtain an image to be detected;

[0076] Taking the brightness levels including: first brightness, second brightness and normal brightness as an example, the preset normal brightness threshold is the above-mentioned first threshold Θ1.

[0077] In addition to normal brightness, each brightness level corresponds to a brightness enhancement model to improve the brightness enhancement effect.

[0078] The input data of the brightness enhancement model is the surveillance image, and the output data is the image to be detected after brightness enhancement. Brightness enhancement is used to improve the success rate of the vault personnel target detection model in identifying personnel under dark light conditions.

[0079] 404: Input the image to be detected into the vault personnel target detection model, and obtain the target detection confidence output by the vault personnel target detection model;

[0080] The safe personnel target detection model is used to detect whether there is a person in the image to be detected, that is, to detect whether the surveillance image shows that there are people entering or leaving the safe.

[0081] The target detection confidence indicates the degree of credibility of the output of the vault personnel target detection model that there is a person in the image to be detected.

[0082] 405: If the target detection confidence is greater than the confidence threshold, it is determined that the surveillance image indicates that a person enters or leaves the safe.

[0083] The confidence threshold can be a fixed value or a dynamic value.

[0084] The present embodiment discloses a method for detecting people entering and leaving a safe during non-business hours. After acquiring a surveillance image during the safe's non-business hours, the method identifies the brightness of the surveillance image and determines the brightness level corresponding to the brightness. When the brightness of the surveillance image is lower than a preset normal brightness threshold, a brightness enhancement model corresponding to the brightness level of the surveillance image is called to perform brightness enhancement on the surveillance image to obtain an image to be detected. Compared with using a uniform brightness enhancement for all dark light images, the brightness enhancement effect of the surveillance image is improved. On this basis, the image to be detected is input into a safe's personnel target detection model for target detection, thereby improving the success rate of identifying people under different brightness conditions. The entire process does not require manual parameters, thereby improving the efficiency and accuracy of detecting people entering and leaving the safe during non-business hours.

[0085] The inventor has found through research that the confidence level of the personnel recognition result of the surveillance image with normal brightness is higher, while the confidence level of the personnel recognition result of the surveillance image with low brightness after brightness enhancement processing is lower, and the lower the brightness of the surveillance image, the lower the confidence level of the personnel recognition result after brightness enhancement processing. If a fixed confidence threshold is used, it may affect the judgment result of whether there is a person entering or leaving the vault.

[0086] In order to solve the above technical problem, this embodiment dynamically sets the confidence threshold according to the brightness of the monitoring image, and the confidence threshold is positively correlated with the brightness of the monitoring image.

[0087] Exemplarily, the calculation formula of the confidence threshold corresponding to the brightness of the monitoring image is as follows:

[0088] Θ person =αT

[0089] Among them, Θ person It represents the confidence threshold corresponding to the brightness of the monitoring image, T is the brightness of the monitoring image, and α represents the threshold coefficient, which is a constant.

[0090] This embodiment discloses dynamically setting a confidence threshold according to the brightness of a surveillance image, thereby improving the success rate of person recognition when the brightness of the surveillance image is low.

[0091] It should be noted that the brightness enhancement model and the vault personnel target detection model corresponding to different brightness levels in the above embodiment need to be trained.

[0092] For example, the brightness levels include: first brightness, second brightness and normal brightness. Figure 5 As shown, Figure 5 The lower brightness image in is the image of the first brightness, the extremely low brightness image is the image of the second brightness, the brightness enhancement model of the lower brightness image is the first brightness enhancement model, and the brightness enhancement model of the extremely low brightness image is the second brightness enhancement model. The training process of the brightness enhancement model is as follows A1-A4:

[0093] A1: Obtain vault scene image data as training samples, and calculate image brightness and brightness level division;

[0094] A2: Label the training samples according to their brightness levels. The labels of the training samples include normal brightness, lower brightness, and extremely low brightness.

[0095] A3: Use training samples labeled with lower brightness and normal brightness to train the model and obtain a brightness enhancement model for lower brightness images;

[0096] A4: Use training samples labeled with extremely low brightness and normal brightness to train the model and obtain an extremely low brightness image brightness enhancement model.

[0097] The above brightness enhancement model can use any image brightness enhancement model based on deep learning, such as the DeepUPE network model. The lower brightness image brightness enhancement model and the extremely low brightness image brightness enhancement model have the same training process except that the training samples are different. For example, both can use the Adam optimizer for optimization training, and set the learning rate to 0.0001, the batch size to 16, and train for 40 rounds to converge. In addition, the trained lower brightness image brightness enhancement model and the extremely low brightness image brightness enhancement model can also be tested and verified.

[0098] The training samples of the vault personnel target detection model include normal brightness surveillance images and surveillance images after the above brightness enhancement. The training samples need to be pre-labeled with people and no people. Then the model is trained using the labeled training samples with people and no people. For example, the YOLOv5 target detection model is used for training, and the Adam optimizer is used for optimization training. The learning rate is set to 0.001, the batch size is 64, and a total of 500 rounds of training convergence are set. The personnel data of the vault scene can also be used to fine-tune and optimize the YOLOv5 model that has been pre-trained using the COCO public data set. In addition, the vault personnel target detection model can also be tested and verified.

[0099] The above-trained low-brightness image brightness enhancement model, extremely low-brightness image brightness enhancement model, and vault personnel target detection model need to be converted into a format, for example, into a cross-platform universal ONNX format, and finally deployed on edge computing devices.

[0100] After the model is deployed, it can be put into use. The edge computing device needs to set the trigger time condition of the vault's non-business hours, that is, the edge computing device is triggered during the vault's non-business hours, and then calculates the image brightness and divides the brightness level of the input real-time vault monitoring image data. If the current image is of low brightness, the trained low-brightness image brightness enhancement model is used for brightness enhancement; if it is of extremely low brightness, the trained extremely low-brightness image brightness enhancement model is used for brightness enhancement; if it is of normal brightness, no brightness enhancement processing is required. Subsequently, the trained vault personnel target detection model is used to identify personnel on the enhanced image or normal brightness image.

[0101] If the target detection confidence output by the vault personnel target detection model is greater than the confidence threshold, it is determined that the surveillance image indicates that a person has entered or exited the vault, and an early warning message is sent to the cloud platform so that the cloud platform forwards the early warning message to at least one target manager, for example, by email, SMS, push message, etc. The early warning message is forwarded to the interruption device of at least one target manager. The early warning information includes the surveillance image, the acquisition time of the surveillance image, and the vault identification, so that the target manager can promptly understand which vault has a safety hazard and deal with it in a timely manner.

[0102] The embodiment of the present application provides a non-business hours personnel entry and exit vault detection system, including a cloud platform and at least one edge computing device, the cloud platform remotely communicates with at least one edge computing device respectively, the edge computing device communicates with at least one surveillance camera in the vault, and the system only needs to deploy a portable edge computing device in the vault environment and connect the vault surveillance camera to complete the environment deployment, without the installation and configuration of other sensors such as fingerprints, voiceprints, and irises. In addition, considering the different light brightness conditions of the vault during non-business hours, it is proposed to divide the surveillance images of the vault into different brightness levels according to the brightness, and call the corresponding brightness enhancement model based on the brightness level of the surveillance image to perform brightness enhancement processing, and then perform personnel target detection and recognition on these brightness enhanced images. In addition, considering that the confidence of the target detection result after low-brightness image enhancement is low, a method for dynamically calculating the target detection personnel recognition confidence threshold according to the current image brightness is proposed to improve the recognition rate of personnel in low-brightness environment. Finally, the non-business hours personnel entry and exit vault detection system disclosed in this embodiment can send the personnel recognition warning image, the location of the warning, and other information found to enter and exit the vault during non-business hours to the cloud platform in a timely manner, save and forward to the target management personnel, so that the target management personnel can handle it in time.

[0103] A method for detecting people entering and exiting a safe during non-business hours provided in an embodiment of the present application is introduced above. A device for executing the method for detecting people entering and exiting a safe during non-business hours will be introduced below.

[0104] See also Figure 6 , Figure 6 This is a schematic diagram of a non-business hours personnel entry and exit detection device provided in an embodiment of the present application. Figure 6 As shown, the non-business hours personnel entry and exit detection device includes:

[0105] An image acquisition unit 601 is used to acquire surveillance images during non-business hours of the vault;

[0106] A brightness recognition unit 602, used to recognize the brightness of the monitoring image and determine the brightness level corresponding to the brightness;

[0107] A brightness enhancement unit 603 is used to call a brightness enhancement model corresponding to the brightness level of the monitoring image to perform brightness enhancement on the monitoring image to obtain an image to be detected if the brightness is lower than a preset normal brightness threshold;

[0108] The target detection unit 604 is used to input the image to be detected into the safe personnel target detection model to obtain the target detection confidence output by the safe personnel target detection model;

[0109] The result determination unit 605 is configured to determine that the surveillance image indicates that a person has entered or exited the safe if the target detection confidence is greater than a confidence threshold.

[0110] In a possible implementation, the brightness enhancement unit 603 is specifically used to, if the brightness level of the monitoring image is a first brightness, call a first brightness enhancement model to perform brightness enhancement on the monitoring image to obtain an image to be detected, and the first brightness enhancement model is trained using training samples labeled with the first brightness and normal brightness; if the brightness level of the monitoring image is a second brightness, call a second brightness enhancement model to perform brightness enhancement on the monitoring image to obtain an image to be detected, and the second brightness enhancement model is trained using training samples labeled with the second brightness and the normal brightness, and the first brightness is greater than the second brightness and the first brightness is less than the normal brightness.

[0111] In a possible implementation, the confidence threshold is positively correlated with the brightness of the monitoring image.

[0112] In a possible implementation, the brightness recognition unit 602 is further configured to determine the surveillance image as the image to be detected and trigger the target detection unit 604 if the brightness level of the surveillance image is normal brightness.

[0113] In a possible implementation, the non-business hours personnel entry and exit detection device further includes:

[0114] The warning information sending unit is used to send warning information to the cloud platform so that the cloud platform forwards the warning information to a terminal held by at least one target manager. The warning information includes the monitoring image, the acquisition time of the monitoring image and the vault identification.

[0115] The present embodiment discloses a device for detecting people entering and leaving a safe during non-business hours. After acquiring a surveillance image during the safe's non-business hours, the device identifies the brightness of the surveillance image and determines the brightness level corresponding to the brightness. When the brightness of the surveillance image is lower than a preset normal brightness threshold, the device calls a brightness enhancement model corresponding to the brightness level of the surveillance image to perform brightness enhancement on the surveillance image to obtain an image to be detected. Compared with using a uniform brightness enhancement for all dark light images, the brightness enhancement effect of the surveillance image is improved. On this basis, the image to be detected is input into a safe's personnel target detection model for target detection, thereby improving the success rate of identifying people under different brightness conditions. The entire process does not require manual parameters, thereby improving the efficiency and accuracy of detecting people entering and leaving the safe during non-business hours.

[0116] Also provided in an embodiment of the present application is a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an edge computing device, the edge computing device implements any of the methods for detecting entry and exit of personnel into a safe during non-business hours provided in an embodiment of the present application.

[0117] A computer-readable storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an edge computing device, the edge computing device can implement any of the methods for detecting entry and exit of personnel into the safe during non-business hours provided in the embodiments of the present application.

[0118] The present application provides a method and a related device for detecting people entering and exiting a safe during non-business hours, which can be used in the field of artificial intelligence or finance. The above is only an example and does not limit the application field of the method and a related device for detecting people entering and exiting a safe during non-business hours provided by the present application.

[0119] It should also be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0120] Through the description of the above implementation mode, the technicians in the field can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present application, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0121] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0122] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, a computer, a training device, or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, training device, or data center. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

Claims

1. A method for detecting personnel entering and exiting a safe during non-business hours, characterized in that: include: Obtaining surveillance images during the vault's non-business hours; Identifying the brightness of the surveillance image and determining a brightness level corresponding to the brightness; If the brightness is lower than a preset normal brightness threshold, a brightness enhancement model corresponding to the brightness level of the monitoring image is called to perform brightness enhancement on the monitoring image to obtain an image to be detected; Inputting the image to be detected into the vault personnel target detection model to obtain the target detection confidence output by the vault personnel target detection model; If the target detection confidence is greater than the confidence threshold, it is determined that the surveillance image indicates that a person has entered or exited the safe.

2. The method for detecting personnel entering and exiting a safe during non-business hours according to claim 1, characterized in that: The step of calling a brightness enhancement model corresponding to the brightness level of the monitoring image to perform brightness enhancement on the monitoring image to obtain an image to be detected includes: If the brightness level of the monitoring image is a first brightness, calling a first brightness enhancement model to perform brightness enhancement on the monitoring image to obtain an image to be detected, wherein the first brightness enhancement model is trained using training samples annotated with the first brightness and normal brightness; If the brightness level of the surveillance image is the second brightness, call the second brightness enhancement model to perform brightness enhancement on the surveillance image to obtain the image to be detected, the second brightness enhancement model is trained using training samples labeled with the second brightness and the normal brightness, the first brightness is greater than the second brightness and the first brightness is less than the normal brightness.

3. The method for detecting personnel entering and exiting a safe during non-business hours according to claim 1, characterized in that: The confidence threshold is positively correlated with the brightness of the monitoring image.

4. The method for detecting personnel entering and exiting a safe during non-business hours according to claim 1, characterized in that: The method for detecting personnel entering and exiting the vault during non-business hours also includes: If the brightness level of the monitoring image is normal brightness, the monitoring image is determined as the image to be detected, and the step of inputting the image to be detected into the safe personnel target detection model is performed.

5. The method for detecting personnel entering and exiting a safe during non-business hours according to claim 1, characterized in that: After determining that the surveillance image indicates that a person has entered or exited the safe, the method for detecting a person entering or exiting the safe during non-business hours further includes: Sending warning information to the cloud platform so that the cloud platform forwards the warning information to a terminal held by at least one target manager, the warning information including the monitoring image, the acquisition time of the monitoring image and the vault identification.

6. A device for detecting people entering and leaving a safe during non-business hours, characterized in that: include: An image acquisition unit, used to acquire surveillance images during non-business hours of the vault; A brightness recognition unit, used to recognize the brightness of the monitoring image and determine the brightness level corresponding to the brightness; A brightness enhancement unit, configured to, if the brightness is lower than a preset normal brightness threshold, call a brightness enhancement model corresponding to the brightness level of the monitoring image to perform brightness enhancement on the monitoring image to obtain an image to be detected; A target detection unit, used to input the image to be detected into a vault personnel target detection model to obtain the target detection confidence output by the vault personnel target detection model; The result determination unit is used to determine that the monitoring image indicates that a person has entered or exited the safe if the target detection confidence is greater than a confidence threshold.

7. An edge computing device, characterized in that: The method comprises at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the edge computing device can implement the method for detecting people entering and exiting the safe during non-business hours as described in any one of claims 1 to 5.

8. A system for detecting personnel entering and leaving a safe during non-business hours, characterized in that: include: A cloud platform and at least one edge computing device; The cloud platform remotely communicates with at least one of the edge computing devices respectively; The edge computing device is communicatively connected to at least one surveillance camera in the safe, and is used to implement the method for detecting people entering and exiting the safe during non-business hours as described in any one of claims 1 to 5.

9. A computer program product, characterized in that It includes computer-readable instructions. When the computer-readable instructions are executed on an edge computing device, the edge computing device implements the method for detecting the entry and exit of personnel into a safe during non-business hours as described in any one of claims 1 to 5.

10. A computer storage medium, characterized in that: The storage medium carries one or more computer programs. When the one or more computer programs are executed by the edge computing device, the edge computing device can implement the method for detecting the entry and exit of personnel into the safe during non-business hours as described in any one of claims 1 to 5.

Images