Risk identification method and risk identification device based on image detection

Abstract

The invention discloses a risk identification method and risk identification device based on image detection, and the method comprises the steps: obtaining a to-be-processed image of target equipment, and calling a standard detection model to carry out the multi-class target detection of the to-be-processed image, outputting category labels of a plurality of functional parts of the target equipment and bounding boxes corresponding to the functional parts; according to the category label of each functional part and based on the spatial parameter of the bounding box, identifying the spatial configuration feature of each functional part in the structural layout of the target equipment so as to judge the type of the target equipment; and if the type of the target equipment is a first preset type, determining a display area of the target equipment from the to-be-processed image according to the spatial parameters, and judging whether a detected object has a security risk or not according to the display content of the display area. According to the technical scheme provided by the invention, the electrification risk of the electrical equipment can be quickly identified.

Description

Technical Field

[0001] This application relates to the field of industrial safety monitoring technology, and in particular to a risk identification method and device based on image detection. Background Technology

[0002] In high-risk work environments such as power, petrochemical, and rail transportation, workers must use multimeters, insulation meters, and other equipment to test parameters such as voltage and insulation resistance of electrical equipment to ensure the safety of the work environment. For example, during the maintenance of wind turbine frequency converters, operators must first disconnect the power supply to the box-type transformer and verify that the equipment is in a de-energized state using instruments such as multimeters before disassembling and installing components. This process requires operators to wear protective equipment and take photos of the multimeter and other instrument displays, then upload the images to the system backend for risk identification to ensure that the tested object complies with safety regulations.

[0003] However, existing risk identification methods rely on manual visual reading of display screens, which has several drawbacks. For example, reviewers are susceptible to visual fatigue and unclear images, leading to misreading of values ​​and incorrectly classifying a live state as a de-energized state, thus causing electric shock accidents. Furthermore, manual image review involves large amounts of data and is inefficient, making it difficult for humans to comprehensively cover all detection scenarios. These problems not only endanger the personal safety of operators but also reduce the reliability of risk identification work.

[0004] Therefore, it is necessary to provide a new risk identification method and device to address the above-mentioned shortcomings. Summary of the Invention

[0005] The purpose of this application is to provide a risk identification method and device based on image detection, which can quickly identify the electrical risks of electrical equipment.

[0006] To achieve the above objectives, this application provides a risk identification method based on image detection. The method includes: acquiring a target device image to be processed, and calling a standard detection model to perform multi-category target detection on the target device image to output category labels of multiple functional components of the target device and bounding boxes corresponding to each functional component; identifying the spatial configuration features of each functional component in the structural layout of the target device based on the category labels of each functional component and the spatial parameters of the bounding boxes to determine the type of the target device; if the type of the target device is a first preset type, determining the display area of ​​the target device from the target device image according to the spatial parameters, and determining whether the object under test has a security risk based on the display content of the display area.

[0007] To achieve the above objectives, this application also provides a risk identification device, comprising: an image processing module, configured to acquire an image of a target device to be processed, and call a standard detection model to perform multi-category target detection on the image to be processed, so as to output category labels of multiple functional components of the target device and bounding boxes corresponding to each functional component; a device type determination module, configured to identify the spatial configuration features of each functional component in the structural layout of the target device based on the category labels of each functional component and the spatial parameters of the bounding boxes, so as to determine the type of the target device; and a risk identification module, configured to determine the display area of ​​the target device from the image to be processed based on the spatial parameters if the type of the target device is a first preset type, and determine whether the tested object has a safety risk based on the display content of the display area.

[0008] To achieve the above objectives, this application also provides a risk identification device, comprising: a memory for storing a computer program; and a processor for executing the computer program stored in the memory, so that the device performs the image detection-based risk identification method as described above.

[0009] To achieve the above objectives, this application also provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, cause the processor to implement the image detection-based risk identification method as described above.

[0010] To achieve the above objectives, this application also provides a computer program product comprising computer program code, which, when run on a computer, enables the computer to implement the image detection-based risk identification method as described above.

[0011] To achieve the above objectives, this application also provides a chip that includes circuitry for performing the image detection-based risk identification method described above.

[0012] Therefore, the technical solution provided in this application, through directed bounding box target detection technology, can not only accurately distinguish different types of testing instruments, but also analyze the spatial layout characteristics of each functional component on the device body. Combined with image rotation correction and adaptive cropping technology, the instrument display screen area can be accurately located and extracted. Furthermore, by performing optical character recognition on the displayed content (such as voltage values, unit identifiers, etc.) and combining it with preset safety thresholds and compliance judgment rules, the system can automatically determine whether the tested device is in a live state. Compared with the traditional manual visual reading method, this solution significantly improves the accuracy and reliability of identifying the risk of device energization, effectively avoiding the risk of missed or incorrect judgments caused by visual fatigue, environmental interference, or human error. At the same time, the entire detection, analysis, and judgment process requires no manual intervention, greatly improving the efficiency of image review. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0014] Figure 1 This is a flowchart of a risk identification method based on image detection in one embodiment of this application; Figure 2 This is a flowchart illustrating the process of an image detection-based risk identification method according to one embodiment of this application. Figure 3 This is a schematic diagram of the functional modules of the risk identification device in the embodiments of this application; Figure 4 This is a schematic diagram of the risk identification device in the embodiments of this application. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be further described in detail below with reference to the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of this application, the terms "first," "second," "third," etc., are only used to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances. Furthermore, in the description of this application, unless otherwise stated, when an element is described as "connected" to another element, it can be directly connected to the other element, or there can be one or more intermediate elements between them. "Multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0016] In high-risk industrial operation scenarios such as power, petrochemical, and rail transportation, operators must use specialized instruments such as multimeters and insulation resistance testers to measure key electrical parameters of electrical equipment, such as voltage, insulation resistance, and grounding continuity. Only after confirming that the equipment is in a safe state with no power and no residual charge can subsequent maintenance or disassembly operations be carried out.

[0017] Taking the maintenance of wind turbine frequency converters in a wind farm as an example, operators must first disconnect the power input of the upstream box-type transformer and then use testing instruments such as multimeters or insulation meters to perform voltage tests on key nodes such as the DC bus and AC input terminals of the frequency converter. Only after confirming that the voltage reading is zero or within the millivolt safe range can operators proceed with the disassembly and replacement of components. This process not only requires operators to wear standard personal protective equipment throughout, but also requires them to photograph the display screen of the testing instrument and upload the image to the safety monitoring system. The back-end system then performs risk identification to ensure that the tested equipment complies with the "no-power" safety regulations.

[0018] However, existing risk identification solutions heavily rely on manual review mechanisms. Reviewers must visually read instrument readings from a large number of uploaded images and manually determine whether they meet safety thresholds. This model has several drawbacks. For example, reviewers are susceptible to visual fatigue, distraction, blurry images, screen glare, and tilted shooting angles, leading to misjudgments of numbers, units (such as "V" and "mV"), or decimal points. This can result in incorrectly classifying a live state as a de-energized state, causing electric shock accidents. Furthermore, as the number of uploaded detection images increases, manual review reaches a bottleneck, making full coverage difficult. To control costs, only a portion of images are often sampled, resulting in a lack of effective supervision for many high-risk operations. Additionally, manual image review is slow and cannot provide real-time alarms.

[0019] Therefore, how to intelligently analyze images collected at the work site and quickly and accurately identify the electrical risks of equipment has become an urgent issue to be addressed in this field.

[0020] The present application will now be described in more detail with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are for illustrative and explanatory purposes only and do not constitute a limitation on the embodiments of the present application. The embodiments described herein are only a part of the embodiments of the present application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this application. It should be noted that the implementing entity of the embodiments of this application is a risk identification device, which may specifically take the form of an electronic device with corresponding hardware and software environment, such as a server, mobile phone, laptop computer, tablet computer, vehicle-mounted device, etc.

[0021] Please refer to the following: Figure 1 and Figure 2 , Figure 1 This is a flowchart of a risk identification method based on image detection in one embodiment of this application. Figure 2 This is a flowchart illustrating the process of a risk identification method based on image detection in one embodiment of this application.

[0022] S101: Obtain the image to be processed of the target device, and call the standard detection model to perform multi-class target detection on the image to be processed, so as to output the category labels of multiple functional components of the target device and the bounding boxes corresponding to each functional component.

[0023] In this embodiment, the risk identification device can retrieve the image of the target device from the server. For example, the system can retrieve a digital image containing a multimeter or insulation meter from the backend server. This digital image can be acquired in real time using a vision sensor mounted on an inspection robot, or it can be manually captured and uploaded by an operator via a mobile terminal. It should be noted that the target device is used by the operator, and the type of target device includes, but is not limited to, testing devices such as multimeters or insulation meters, which are used to perform voltage testing on electrical equipment.

[0024] Once the risk identification device acquires the image to be processed, it can invoke pre-trained standard detection models, such as the OBB (Oriented Bounding Box) object detection model and the Mask R-CNN (Mask Region-based Convolutional Neural Network) detection model, to perform multi-class object detection on the image to identify its key features. For example, the risk identification device can use the OBB object detection model to extract multi-level feature maps from the image. These feature maps contain feature information at different levels of abstraction, such as the edges, textures, and functional components of the target device. Then, based on these feature maps, analysis and regression are performed, ultimately outputting a series of detected targets with oriented bounding boxes and corresponding category names. The bounding boxes at least include spatial parameters such as spatial location, orientation angle, and scale dimension. Specifically, the risk identification device can use the OBB object detection model to output bounding boxes for functional components such as the device body, display screen, high-voltage button of the insulation meter, test socket, and function knob, and output the category label for each bounding box.

[0025] It is particularly important to note that during the construction and training phase of the standard detection model, architectures such as YOLO11-obb and RetinaNet can be selected as the base network. Based on this, a directed bounding box regression head is introduced. By outputting the center coordinates, width and height dimensions, and rotation angle parameters of the rotated rectangle, accurate localization of each functional component under tilted and deflected states can be achieved. The labeled sample library of training data should cover different brands, models, and usage scenarios of target devices such as multimeters and insulation meters. The functional components in each image (including the device body, display screen, operation buttons, test jacks, adjustment knobs, etc.) should be labeled with directed bounding boxes according to a unified standard, and each functional component should be assigned a unique category label. During model training, data augmentation strategies such as random rotation, scaling, color perturbation, and mosaic enhancement can be used to improve the model's generalization performance in complex industrial scenarios.

[0026] S102: Based on the category labels of each functional component and the spatial parameters of the bounding box, identify the spatial configuration features of each functional component in the target device structural layout to determine the type of the target device.

[0027] In this embodiment, after outputting the category labels of functional components and the corresponding bounding boxes of each functional component, the risk identification device can calculate the center point coordinates, rotation angle, width, height, and other information of each functional component in the image coordinate system based on the category labels of each functional component and the spatial parameters of its corresponding bounding box in the target device. This allows the construction of the spatial topological relationships between the functional components, thereby identifying the spatial configuration features of each functional component in the structural layout of the target device. These spatial configuration features characterize the relative geometric relationships between the functional components. Then, by matching the relative geometric relationships between the functional components with a preset device layout pattern, the type of the target device can be determined.

[0028] For example, a risk identification device can construct a structural feature vector representing the relative geometric relationships between functional components by considering the distance between their center points, relative azimuth angles, and spatial arrangement order. This structural feature vector is then matched against preset structural layout templates for various types of equipment, where each template corresponds to a type of equipment (e.g., template one corresponds to a multimeter, template two to an insulation meter). If the matching degree exceeds a set threshold, the target equipment can be determined to belong to the corresponding equipment type.

[0029] It should be noted that, for different types of target equipment, their structural layout features can be standardized and modeled in advance. For example, by collecting sample images of various types of equipment under standard operating conditions, the core features such as the category combination of equipment-specific functional components, bounding box spatial parameters, and relative geometric relationships between components can be extracted and solidified to construct standardized structural layout templates corresponding to each type of equipment.

[0030] In practical applications, the testing instruments used by operators are usually uniformly distributed by their organizations, with relatively fixed models and significant and stable differences in structural characteristics between different types of instruments. Under this premise, identifying the spatial configuration characteristics of each functional component within the target equipment's structural layout to determine the target equipment type can be further simplified. For example, if it is known that a certain type of insulation meter structurally includes a specific "insulation meter button" functional component, while a multimeter used in the same scenario does not contain this component, then the equipment type can be directly determined based on the functional component category label output by the risk identification device. Specifically, if the category label output by the risk identification device includes the "insulation meter button" category label, the target equipment type can be directly determined to be an insulation meter; if the category label output by the risk identification device does not include the "insulation meter button" category label, and the candidate equipment type only includes multimeters and insulation meters, then the target equipment type can be directly determined to be a multimeter.

[0031] In practical applications, the background pattern of the image to be processed may contain icons that are visually similar to the "insulation meter button". Such similar icons may cause the model to mistakenly identify other types of testing instruments as insulation meters.

[0032] To reduce the risk of misclassification of device type caused by the above reasons, in one embodiment, when the category label output by the risk identification device includes the category label "insulation meter button," it can further determine whether the bounding box corresponding to the insulation meter button and the bounding box of the device body satisfy a preset attachment relationship in spatial position. If the preset attachment relationship is satisfied, the target device type can be determined to be an insulation meter. In this embodiment, the preset attachment relationship refers to the spatial correlation between the insulation meter button and the device body, which can be defined by a relative distance threshold between bounding boxes, an overlap area ratio, or a topological adjacency relationship. For example, when the boundary distance between the bounding box corresponding to the "insulation meter button" label and the bounding box corresponding to the device body is within a reasonable range, it can be considered that the preset attachment relationship is satisfied. Using the preset attachment relationship can verify the rationality of the physical connection between the functional component and the device body, avoiding misclassification of type due to image interference or false detection.

[0033] S103: If the target device is of the first preset type, the display area of ​​the target device is determined from the image to be processed according to the spatial parameters, and the safety risk of the object under test is determined according to the display content of the display area.

[0034] In this embodiment, when the risk identification device determines that the target device belongs to a first preset type, it can invoke the safety specification check logic corresponding to that device type. This logic is then used to analyze the display content of the screen area in the image to be processed, thereby determining whether the current measurement state of the object under test complies with relevant safety specifications. When determining the display area of ​​the target device, the risk identification device can determine candidate positions of the display area based on the spatial parameters of each functional component in the target device, combined with the relative positional relationship between the display area and adjacent functional components in the preset device structure layout. Then, through visual feature analysis and contextual verification, the device determines the accurate position of the display area within the target device structure layout and extracts the display area.

[0035] In another feasible implementation, the risk identification device can extract the orientation angle of the main bounding box of the target device based on the OBB detection results, and use the aforementioned orientation angle to perform rotation correction on the image to be processed. After obtaining the corrected image, the risk identification device can accurately locate and crop the display area based on the spatial position of the display bounding box corresponding to the display component.

[0036] After the risk identification device captures the aforementioned display area, it can use optical character recognition technology to extract text information from the display area and use preset security standard check logic to interpret the text information, such as analyzing whether the numerical value or character combination meets the security threshold requirements, and then determine whether the current measurement status of the object being measured meets the relevant security standards.

[0037] In one implementation, determining whether the object under test poses a security risk based on the content displayed in the display area can be achieved in the following way: First, the risk identification device determines whether a first specified character exists in the displayed content. If the first specified character (e.g., mV or mv) exists, it can be directly determined that the current measurement state of the object under test does not pose a safety risk. If the first specified character does not exist, it is necessary to further determine whether a second specified character (e.g., V or v) exists. If the second specified character exists, it is further determined whether the value in the displayed content is greater than a specified threshold (e.g., 12 or 24). If the value in the displayed content is greater than the specified threshold, it can be determined that the voltage of the object under test exceeds the safe voltage for the human body, and the object under test poses a safety risk; the operator cannot disassemble or replace components. If the value in the displayed content is less than or equal to the specified threshold, it can be considered that the voltage of the object under test is within the safe voltage range for the human body, and the object under test does not pose a safety risk; the operator can disassemble or replace components.

[0038] Furthermore, if neither the first nor the second specified character is present in the displayed content, it indicates that the target device may be malfunctioning, or that the operator has incorrectly used the target device. In this case, the displayed content cannot reflect the true electrical condition of the object under test. To reduce risk in this scenario, the risk identification device can determine that the object under test poses a safety risk, and the operator must not disassemble or replace any components.

[0039] In practical applications, some testing devices, despite their different structures, can be used functionally equivalently. For example, an insulation meter can be used as a multimeter for voltage testing in certain scenarios. When the target device is of the second preset type (such as an insulation meter), its structural layout differs significantly from the first preset type (such as a multimeter). This will prevent the risk identification device from classifying it as the first preset type, thus failing to trigger the corresponding safety regulation check logic. However, if its current operating function is substantially equivalent to the first preset type (such as performing the same type of voltage measurement), it still needs to be included in the corresponding safety supervision.

[0040] To address the aforementioned issues, in one embodiment, if the risk identification device determines that the target device is of a second preset type, it can further determine whether the target device is functionally equivalent to the first preset type device. If the target device is functionally equivalent to the first preset type device, the display area of ​​the target device can be determined from the image to be processed based on the spatial parameters of each functional component in the target device, and the safety risk of the tested object can be determined based on the display content of the display area. The process of determining whether the tested object has a safety risk based on the display content of the display area can be referred to the content in the foregoing embodiments, and will not be repeated here.

[0041] In one implementation, determining whether the target device is functionally equivalent to a first preset type of device can be achieved in the following way: First, the risk identification device can extract the main bounding box corresponding to the target device body and the interface bounding box corresponding to the signal interface component based on the OBB detection results of the target device. Then, it establishes a device body coordinate system using the relevant spatial parameters of the main bounding box. Next, based on the relevant spatial parameters of the interface bounding box, it determines whether the interface bounding box is located within a preset area of ​​the device body coordinate system. This preset area can be set according to the structural layout of a second preset type. If the interface bounding box is located within the preset area of ​​the device body coordinate system, the target device can be considered functionally equivalent to a device of the first preset type.

[0042] For example, if the target device is an insulation meter and its insulation test jack is located near the lower left corner of the device body, then when the risk identification device detects that the insulation test jack is in the lower left corner of the device body based on the OBB test results of the target device, it means that the insulation test jack is not blocked (i.e., the insulation test jack is not used). This indicates that the test lead is plugged into another jack (usually the right-side jack), and the target device is being used as a multimeter, not as an insulation meter. Therefore, the insulation meter is functionally equivalent to a multimeter at this time. Thus, the risk identification device can use the safety specification check logic corresponding to the multimeter (i.e., the first preset type of device) to interpret the display content of the target device, and then determine whether the tested object has a safety risk.

[0043] It should be noted that if the risk identification device identifies the target device as an insulation meter and does not detect the insulation test socket in the lower left corner of the device body, it means that the test leads are inserted into the insulation test socket, i.e., the target device is being used as an insulation meter. Accordingly, the reading displayed by the target device cannot reflect the energization status of the object under test. Therefore, the risk identification device can determine that the object under test poses a safety risk, and the operator must not disassemble or replace components on the object under test.

[0044] Please see Figure 3 This application also provides a risk identification device, the device comprising: The image processing module is used to acquire the image to be processed of the target device, and call the standard detection model to perform multi-class target detection on the image to be processed, so as to output the category labels of multiple functional components of the target device and the bounding boxes corresponding to each functional component; The device type determination module is used to identify the spatial configuration features of each functional component in the target device structure layout based on the category label of each functional component and the spatial parameters of the bounding box, so as to determine the type of the target device; The risk identification module is used to determine the display area of ​​the target device from the image to be processed based on the spatial parameters if the type of the target device is a first preset type, and to determine whether the object under test has a safety risk based on the display content of the display area.

[0045] In one feasible implementation, the risk identification module is further configured to: If the target device is of the second preset type, it is determined whether the target device is functionally equivalent to the first preset type device. If so, the display area of ​​the target device is determined from the image to be processed according to the spatial parameters, and the test object is determined to have a security risk based on the display content of the display area.

[0046] Please see Figure 4 This application also provides a risk identification device, which includes a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, it can implement the image detection-based risk identification method described above. Specifically, at the hardware level, the risk identification device may include a processor, an internal bus, and a memory. The memory may include main memory and non-volatile memory. The processor reads the corresponding computer program from the non-volatile memory into main memory and then runs it. Those skilled in the art will understand that... Figure 4 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned risk identification device. For example, the aforementioned risk identification device may also include a... Figure 4 The components shown may include more or fewer components, such as other processing hardware like a GPU (Graphics Processing Unit) or external communication ports. Of course, this application does not exclude other implementation methods besides software implementations, such as logic devices or a combination of hardware and software.

[0047] In this embodiment, the processor may include a central processing unit (CPU) or a graphics processing unit (GPU), and may also include other microcontrollers, logic gates, integrated circuits, or appropriate combinations thereof with logic processing capabilities. The memory described in this embodiment can be a storage device for storing information. In digital systems, a device capable of storing binary data can be a memory; in integrated circuits, a circuit without physical form but with storage function can also be a memory, such as RAM or FIFO; in a system, a storage device with physical form can also be called a memory. In implementation, this memory can also be implemented using a cloud storage method; the specific implementation method is not limited in this specification.

[0048] It should be noted that the specific implementation of the risk identification device in this specification can be found in the description of the method implementation method, and will not be elaborated here.

[0049] This application also provides a computer-readable medium storing instructions that, when executed by a processor, enable the processor to implement the image detection-based risk identification method described in the above embodiments.

[0050] This application also provides a computer program product, which includes computer program code. When the computer program code is run on a computer, the computer can implement the risk identification method based on image detection in the above embodiments.

[0051] This application also provides a chip, the chip including circuitry, the circuitry being used to perform the image detection-based risk identification method in the above embodiments.

[0052] Therefore, the technical solution provided in this application, through directed bounding box target detection technology, can not only accurately distinguish different types of testing instruments, but also analyze the spatial layout characteristics of each functional component on the device body. Combined with image rotation correction and adaptive cropping technology, the instrument display screen area can be accurately located and extracted. Furthermore, by performing optical character recognition on the displayed content (such as voltage values, unit identifiers, etc.) and combining it with preset safety thresholds and compliance judgment rules, the system can automatically determine whether the tested device is in a live state. Compared with the traditional manual visual reading method, this solution significantly improves the accuracy and reliability of identifying the risk of device energization, effectively avoiding the risk of missed or incorrect judgments caused by visual fatigue, environmental interference, or human error. At the same time, the entire detection, analysis, and judgment process requires no manual intervention, greatly improving the efficiency of image review.

[0053] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A risk identification method based on image detection, characterized in that, The method comprises: obtaining a to-be-processed image of a target device, and calling a standard detection model to perform multi-class target detection on the to-be-processed image to output class labels of multiple functional components of the target device and a bounding box corresponding to each functional component; according to the class labels of each functional component and based on spatial parameters of the bounding box, identifying spatial configuration features of each functional component in a structural layout of the target device to determine a type of the target device; if the type of the target device is a first preset type, determining a display area of the target device from the to-be-processed image according to the spatial parameters, and determining whether a measured object has a safety risk according to display content of the display area.

2. The method of claim 1, wherein, if the type of the target device is a second preset type, determining whether the target device is functionally equivalent to the first preset type device, if yes, determining a display area of the target device from the to-be-processed image according to the spatial parameters, and determining whether a measured object has a safety risk according to display content of the display area.

3. The method of claim 2, wherein, determining whether the target device is functionally equivalent to the first preset type device comprises: establishing a device body coordinate system based on a body bounding box corresponding to a main body of the target device; determining whether an interface bounding box corresponding to a signal interface component is located in a preset area of the device body coordinate system, if yes, the target device is functionally equivalent to the first preset type device.

4. The method of claim 3, wherein, the first preset type is a multimeter, and the second preset type is an insulation meter.

5. The method of claim 4, wherein, identifying spatial configuration features of each functional component in a structural layout of the target device to determine a type of the target device comprises: if there is an insulation meter button class label, and a bounding box corresponding to the insulation meter button satisfies a preset attachment relationship with the body bounding box in terms of spatial position, the target device is an insulation meter; if there is no insulation meter button class label, the target device is a multimeter.

6. The method of claim 5, wherein, determining a display area of the target device from the to-be-processed image according to the spatial parameters comprises: performing rotation correction on the to-be-processed image according to an orientation angle of the body bounding box; in the corrected image, cutting out the display area according to spatial position and scale parameters of a display bounding box corresponding to a display component.

7. The method of claim 6, wherein, determining whether a measured object has a safety risk according to display content of the display area comprises: determining whether there is a first specified character in the display content, if there is the first specified character, the measured object does not have a safety risk; if there is no first specified character, determining whether there is a second specified character in the display content, if there is the second specified character, determining whether a numerical value in the display content is greater than a specified threshold, if greater than the specified threshold, the measured object has a safety risk, and if less than or equal to the specified threshold, the measured object does not have a safety risk.

8. The method of claim 7, wherein, determining whether a measured object has a safety risk according to display content of the display area further comprises: if there is no first specified character and no second specified character in the display content, the measured object has a safety risk.

9. A risk identification apparatus, characterized by, the device comprises: An image processing module is configured to acquire a to-be-processed image of a target device, and call a standard detection model to perform multi-class target detection on the to-be-processed image to output class labels of multiple functional components of the target device and a bounding box corresponding to each functional component. A device type judgment module is configured to identify spatial configuration features of each functional component in a structural layout of the target device according to the class labels of the functional components and based on spatial parameters of the bounding box, to judge a type of the target device. A risk identification module is configured to, if the type of the target device is a first preset type, determine a display area of the target device from the to-be-processed image according to the spatial parameters, and judge whether a measured object has a security risk according to display content of the display area.

10. The risk identification apparatus according to claim 9, characterized by The risk identification module is further configured to: If the type of the target device is a second preset type, judge whether the target device is functionally equivalent to the first preset type of device, if yes, determine a display area of the target device from the to-be-processed image according to the spatial parameters, and judge whether a measured object has a security risk according to display content of the display area.

11. A risk identification apparatus, characterized by, The apparatus comprises: a memory configured to store a computer program; a processor configured to execute the computer program stored in the memory, so that the apparatus performs the method of any one of claims 1 to 8.

12. A computer-readable storage medium, characterized in that, An instruction is stored thereon, and the instruction is executed by a processor to enable the processor to implement the method of any one of claims 1 to 8.

13. A computer program product, characterised in that, The computer program product comprises computer program code, which, when executed on a computer, enables the computer to implement the method of any one of claims 1 to 8.

14. A chip, characterized by The chip comprises a circuit configured to execute the method of any one of claims 1 to 8.

Images