Operation scene investigation auxiliary system, method and tool

Abstract

The invention discloses a work scene investigation auxiliary system, method and tool, and the system comprises an image collection unit which comprises a camera and a light supplement lamp and is used for collecting multi-angle images of a work scene; the mechanical support and motion module comprises an electric control lifting rod, an electric control holder and a driver, and is used for carrying and controlling the image acquisition unit to perform multi-degree-of-freedom motion in the working space; the control and processing unit comprises a main control board and an image processor in communication connection with the main control board; wherein the main control board is configured with an automatic path planning algorithm, controls the mechanical support and motion module to move according to a predetermined track, and synchronously triggers the image acquisition unit to perform image acquisition; the image processor is configured with an image processing algorithm and is used for processing the acquired multi-angle image to generate a panoramic image and / or state change information of the operation scene; according to the system, the storage medium, the method and the tool, the operation and maintenance work efficiency and accuracy can be remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of power equipment operation and maintenance technology, and in particular to auxiliary systems, methods and tools for surveying work scenarios. Background Technology

[0002] Currently, the status survey and recording of secondary switchgear in converter stations mainly relies on the traditional method of manual photography by maintenance personnel using handheld DSLR cameras or smartphones. This method has the following inherent drawbacks: (1) The survey is inefficient and relies on manual experience: Maintenance personnel need to adjust their body posture and shooting angle in front of each cabinet to frame, focus and take pictures within the limited time window of power outage maintenance. The whole process is time-consuming and labor-intensive, with low automation, and cannot quickly complete the general survey of large-scale cabinets.

[0003] (2) There are blind spots and the information is not complete: Due to the limited size of the equipment and the small space inside the cabinet, it is difficult to effectively cover key areas such as the back of the terminal block and the depth of the cable tray by manual shooting, resulting in blind spots in the collected image data, which poses a hidden danger for subsequent status assessment and fault tracing.

[0004] (3) Poor image quality and consistency: Due to the manual operation, it is difficult to keep the angle, focal length and lighting conditions consistent each time, resulting in inconsistent image quality of the acquired image sequence, which brings great difficulties to subsequent image stitching and historical comparison.

[0005] (4) Low reliability of status recognition: The existing status verification relies entirely on the operation and maintenance personnel to visually compare the current photo with the historical photo. This method is very prone to omission or misjudgment of minor but critical anomalies (such as loose wiring or component corrosion) due to visual fatigue, distraction or subjective judgment differences. It lacks objective and unified judgment standards and intelligent analysis capabilities.

[0006] Therefore, there is an urgent need in this field for a survey assistance tool that can achieve automated and standardized data collection and intelligent comparative analysis to improve the efficiency, quality and reliability of operation and maintenance work. Summary of the Invention

[0007] To address the shortcomings of the existing technologies, the present invention aims to provide a work scene survey assistance system, method, and tool, which aims to achieve blind-spot-free, high-definition, and standardized automatic image acquisition of the internal space of the cabinet, and possess intelligent image processing and status change recognition capabilities, thereby significantly improving the efficiency and accuracy of operation and maintenance work.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a work scene reconnaissance assistance system, which mainly includes: Mechanical support and motion module: Composed of an electrically controlled lifting rod, an electrically controlled pan-tilt unit, and a driver, forming a multi-degree-of-freedom precision motion system used to precisely control the movement trajectory of the image acquisition unit within the confined space inside the cabinet; Image acquisition unit: Includes a high-definition industrial-grade camera and a ring LED fill light to ensure that high-resolution, low-distortion, and detailed images can be acquired even under complex lighting conditions; Control and processing unit: includes a main control board and an image processor. The main control board serves as the edge computing and control hub, with a built-in automatic path planning algorithm to coordinate and control the automated acquisition process of the entire system; the image processor is specifically responsible for running a series of advanced image processing algorithms.

[0009] The image processing algorithm is the core of the intelligent implementation of this invention, and it includes at least: Panoramic Image Generation Algorithm: When performing feature matching on multi-angle images, this algorithm innovatively introduces geometric constraints (such as orientation consistency constraints and spatial distribution consistency constraints) based on prior knowledge of the physical layout of the device (such as the vertical / horizontal orientation of components). This can effectively filter out mismatched points generated in repetitive texture environments, thereby generating high-precision, seamlessly stitched panoramic images.

[0010] Real-time image quality assessment and control algorithm: This algorithm calculates the sharpness evaluation value (such as Laplacian variance) of each frame of image in real time at the acquisition end and compares it with a preset threshold. Once image blur is detected, a retake at the same location is immediately triggered, ensuring the quality of the input image from the source and providing a reliable data foundation for subsequent processing.

[0011] Intelligent Difference Detection Algorithm: This algorithm combines a Siamese neural network with a rule engine. The Siamese network is responsible for initially locating difference regions from massive image details and generating a difference heatmap; the rule engine performs logical reasoning and filtering on candidate differences based on a predefined device topology rule base (such as allowed offset ranges and legal port connection relationships), and finally outputs high-confidence anomaly alarms, greatly reducing the false alarm rate.

[0012] In a second aspect, the present invention provides a method for surveying a work scene, the method being applied to the aforementioned system and comprising: Through the coordinated operation of the mechanical support and motion module and the image acquisition unit, multi-angle image sequences of the work scene are automatically acquired; The multi-angle image sequence is processed to generate a panoramic image of the work scene; The currently generated panoramic image is compared with historical baseline panoramic images to detect changes in the state of the work scene.

[0013] In the preferred embodiment of the above method, the step of generating the panoramic image includes executing a panoramic image generation algorithm; the step of detecting state changes includes executing a difference detection algorithm.

[0014] In a third aspect, the present invention provides a work scene survey assistance tool, applied to the aforementioned system, comprising: Cameras are used to capture multi-angle images of the work scene; The supplementary light integrates a ring array LED light source and has intelligent adjustment functions for color temperature and brightness. The main control board, as the edge computing and control hub, runs embedded software to control image acquisition, motion unit coordination, and data communication. The electrically controlled lifting pole and tripod are used to support and control the camera's vertical movement; An electronically controlled pan-tilt head is used to control the horizontal tilt and rotation of a camera. The driver is used to receive signals from the main control board and drive the electric lifting rod and the electric pan-tilt unit to move. The main control board executes a path planning algorithm to achieve automated multi-angle image acquisition; the main control board stores the automatic path planning algorithm and the panoramic image generation algorithm.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. Comprehensive and efficient surveying: Through a multi-degree-of-freedom motion system and automatic path planning, automated image acquisition without blind spots inside the cabinet is achieved, completely solving the problem of visual blind spots and significantly shortening the time of a single survey, increasing efficiency several times over.

[0016] 2. High-quality and standardized data: Through real-time quality assessment and re-capture mechanisms, the high definition and consistency of all acquired images are ensured, laying the foundation for establishing a reliable equipment image archive.

[0017] 3. Intelligent and Highly Reliable Analysis: By integrating a series of algorithms such as improved feature matching, focus stack fusion, twin networks, and rule engines, the system achieves automatic and accurate identification and alarm of device status, reducing reliance on human experience and significantly improving the objectivity and accuracy of status identification.

[0018] 4. Flexible deployment and strong adaptability: The system adopts a modular and compact design, which does not require complicated fixed installation. It can be quickly adapted to different models and sizes of cabinets, and has good on-site applicability and promotion value. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall structural diagram of the work scene survey auxiliary tool of the present invention; Figure 2 This is a structural diagram of the system of the present invention. Detailed Implementation

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0021] It should be understood that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product, or device.

[0022] refer to Figure 1 and Figure 2 This embodiment provides a work scenario survey assistance system, which is specifically used for the internal survey of the secondary cabinets in a converter station. Its specific composition is as follows: The image acquisition unit 300 includes a camera 301 and a supplementary light 302, and is used to acquire multi-angle images of the work scene; The mechanical support and motion module 100 includes an electrically controlled lifting rod 101, an electrically controlled pan-tilt unit 103, and a driver 102, which is used to mount and control the image acquisition unit 300 to perform multi-degree-of-freedom motion in the work space. The control and processing unit 200 includes a main control board 202 and an image processor 201 that is communicatively connected to it. The main control board 202 is configured with an automatic path planning algorithm to control the mechanical support and motion module 100 to move along a predetermined trajectory and simultaneously trigger the image acquisition unit 300 to acquire images. The image processor 201 is configured with a panoramic image generation algorithm to process the acquired multi-angle images to generate panoramic images of the work scene and / or status change information.

[0023] Specifically, the mechanical support and motion module 100 is physically implemented as follows: Electric lifting rod 101: Made of high-strength lightweight alloy material, it can precisely control its extension stroke under the drive of driver 102, thereby driving the image acquisition unit 300 installed at its top to move smoothly in the vertical direction, covering all height levels of the cabinet from bottom to top; Electronically controlled gimbal 103: As a key component for achieving horizontal scanning without blind spots, this gimbal can perform pitch and rotation movements under the drive of the driver 102. It works in conjunction with the electronically controlled lifting rod 101 to form a multi-degree-of-freedom precision motion system; Driver 102: As a power execution unit, it can specifically adopt a stepper motor or a servo motor and its matching drive circuit; it receives low-voltage control signals from the main control board 202 and converts them into sufficient current and voltage to drive the extension and retraction of the electric lifting rod 101 and the rotation of the electric gimbal 103, ensuring that the motion mechanism can respond to control commands quickly, smoothly and accurately.

[0024] The image acquisition unit 300 is physically implemented as follows: Camera 301 is a high-definition industrial-grade camera: as the imaging core of the system, it uses a high-resolution, high-sensitivity miniature image sensor and is equipped with a professional-grade macro optical lens to ensure clear capture of details such as sub-millimeter level terminals and equipment labels. This camera 301 integrates global shutter technology to eliminate motion blur, providing high-quality, distortion-free original images for subsequent panoramic stitching and intelligent analysis. Fill light 302: An integrated ring-array LED light source is arranged around the lens of camera 301. This fill light 302 has intelligent color temperature and brightness adjustment functions and is equipped with a polarizing filter, which can effectively suppress strong reflections on the metal surface inside the cabinet, ensuring that images captured at different angles and under complex lighting conditions have good consistency, uniformity and detail.

[0025] The control and processing unit 200 is implemented physically and logically as follows: Main control board 202: As the edge computing and control hub of the device, it can adopt an embedded ARM or x86 architecture computing module. It is responsible for running embedded software, and its core functions include: receiving instructions from the host computer, executing automatic path planning algorithms (which are pre-set with scanning paths for different cabinet models), synchronously controlling all motion units and imaging units, implementing local image caching and preprocessing (such as quality verification and exposure equalization), and realizing data communication and interaction with remote servers through its network interface (such as Ethernet or 5G module). Image Processor 201: It can be a standalone GPU computing card or a powerful processing core (such as an FPGA or a SoC with high computing power) integrated on the main control board 202. It is specifically configured to execute image processing algorithms, including panoramic image generation algorithms.

[0026] The image processing algorithm includes a panoramic image generation algorithm, and the panoramic image generation algorithm executed by the image processor 201 includes the following steps: Feature point extraction and matching are performed on the input multi-angle image set to obtain an initial set of matching point pairs; Based on prior knowledge of the physical layout of the equipment in the aforementioned work scenario, geometric constraint filtering is applied to the initial set of matching point pairs to eliminate incorrect matching point pairs and obtain a set of matching points. The geometric constraints include at least directional consistency constraints, which are used to filter out matching point pairs whose principal direction difference is approximately 0° or an integer multiple of 90°. The spatial transformation relationship between images is calculated using the matching point set, and image stitching and fusion are performed to generate the panoramic image.

[0027] Specifically, when the image processor 201 executes the panoramic image generation algorithm, the specific process is as follows: Feature point extraction and matching: The SIFT (Scale Invariant Feature Transform) algorithm is used to extract key points and their descriptors from each input image. First, a Gaussian scale space L(x,y,σ)=G(x,y,σ)*I(x,y) is constructed, where G is the Gaussian kernel function and σ is the scale factor. Then, extreme points are detected in the Gaussian difference pyramid D(x,y,σ)=L(x,y,kσ)-L(x,y,σ) as candidate key points. Subsequently, by calculating the Euclidean distance between descriptors, preliminary matching of feature points from different images is performed to obtain an initial set of matching point pairs. ,in =( , ), and These are the corresponding feature points in the two images.

[0028] Geometric Constraint-Based Mismatch Filtering: Traditional methods (such as RANSAC) perform poorly in environments with repetitive textures and similar structures within the cabinet. This invention introduces prior knowledge of the physical layout of the cabinet equipment as a strong constraint. The geometric constraints also include spatial distribution consistency constraints and directional consistency constraints. The steps for spatial distribution consistency constraints include: Calculate the candidate space transformation model based on the candidate matching point set; Calculate the projection error of each matching point under the candidate model; Based on the statistical distribution of the projection error, matching points whose errors significantly deviate from the mainstream distribution are eliminated.

[0029] 1. Directional Consistency Constraint: Components within the cabinet (such as terminal blocks and wires) are typically arranged vertically or horizontally. Therefore, the correct matching point corresponds to the principal direction difference of m_i. It should be approximately an integer multiple of 0° or 90°, i.e., |Δθ_i-k*90°|<τ_θ, where k is an integer. This is a preset tolerance threshold (e.g., 5°). Matching points that do not meet this condition will be discarded.

[0030] 2. Spatial Distribution Consistency Constraint: Correct matching points should support a reasonable global transformation model (such as the homography matrix H) in terms of spatial distribution. We calculate a candidate transformation based on the current set of candidate matching points. Then, calculate the projection error of each matching point under this model: ε_i=||q_i-H_candidate*p_i||².

[0031] Then, the median absolute deviation (MAD) of all projection errors was calculated, and matching points with errors greater than twice the MAD (i.e., points whose errors significantly deviate from the mainstream distribution) were removed.

[0032] Panoramic image generation: using the matching point set filtered by the above dual constraints. Robust estimation algorithms, such as the least squares method, are used to estimate the optimal homography matrix H. All images I are then projected onto the same global coordinate system using their corresponding transformation matrices, and weighted fusion algorithms, such as multi-band fusion, are employed to eliminate seams, generating the final seamless, high-fidelity panoramic image of the screen cabinet. .

[0033] In a preferred embodiment, the control and processing unit 200 is further configured to execute a real-time image quality assessment and control algorithm, the algorithm including: During image acquisition, the sharpness evaluation value of the currently acquired image is calculated in real time; wherein, the sharpness evaluation value is the variance of the image after processing by the Laplacian operator; The clarity evaluation value is compared with a preset threshold. If the clarity evaluation value is lower than the preset threshold, the main control board 202 immediately controls the image acquisition unit 300 to re-acquire the image at the same location.

[0034] Specifically, the real-time image quality assessment and control algorithm is implemented as follows: The real-time image quality assessment and control algorithm executed by the control and processing unit 200 is specifically implemented as follows: Sharpness assessment: During the image acquisition process, each captured image frame is assessed. The sharpness evaluation value is calculated in real time at the edge (i.e., on the main control board 202 or the image processor 201). This value is obtained by calculating the variance of the image after processing with the Laplacian operator, i.e., V_i = Var(∇²F_i). The larger the value, the better the image. The richer the high-frequency information, the clearer the image.

[0035] Judgment and control: The calculated... Compared with the preset sharpness threshold Compare. If Below the threshold If the image frame is blurry, it is determined that the blurriness may be due to jitter or defocus. In this case, the system will not send this low-quality image to subsequent processing. Instead, the main control board 202 immediately sends instructions to the mechanical support and motion module 100 and the image acquisition unit 300, controlling them to re-acquire the image at the same location. This innovative mechanism ensures the quality of the original image input to subsequent algorithms from the data source, preventing post-processing failure due to poor input quality.

[0036] Specifically, for the image sequence F that has passed quality control, its Laplacian pyramid {L1,L2,…,Lk} and Gaussian pyramid {G1,G2,…,Gk} are constructed. For each layer l of the pyramid, pixel-level fusion is performed based on the sharpness of each image in that layer (the weight map W_i derived from V_i): L_fused_l=Σ_i[G{W_i}_l•L{F_i}_l], Where G{W_i}_l is the Gaussian pyramid representation of the weight map W_i at layer l, and • denotes dot product. Finally, the fused Laplacian pyramid L_fused is reconstructed to obtain the globally clear image F_fused.

[0037] In a preferred embodiment, the difference detection algorithm executed by the image processor 201 includes: The feature maps of the baseline panoramic image and the current panoramic image are extracted using a Siamese neural network, and a difference heatmap is generated. Based on a preset device topology rule base, logical reasoning and filtering are performed on the candidate difference regions identified from the difference heatmap; Output a filtered list of high-confidence disparity regions and their category labels.

[0038] The specific implementation of the intelligent difference detection algorithm is as follows: The difference detection algorithm executed by the image processor 201 includes the following specific steps: Siamese neural network feature matching: A pre-trained, weighted convolutional neural network (CNN_shared) is used to extract baseline panoramic images. Compared with the current panoramic image of the inspection The feature maps I_base and I_current are used. A difference heatmap is generated by calculating the cosine similarity of corresponding positions between the feature maps. The formula is: D_map(x,y)=1-(F_base(x,y)•F_current(x,y)) / (||F_base(x,y)||*||F_current(x,y)||). in, The higher the median value, the more likely the area is to have undergone changes.

[0039] Post-processing filtering in the rules engine: First, Medium above the threshold The region was initially defined as a candidate difference area. .

[0040] Subsequently, these regions are logically compared with the predefined device topology rule base R, specifically including: Apply unstructured change filtering rules: The rule base R contains clauses such as "device label position is allowed to have an offset of ±N pixels". If the candidate region... If the positional offset is determined to be within the allowable range, it will be filtered out and not considered an anomaly.

[0041] In a preferred embodiment, the step of performing logical reasoning and filtering based on the device topology rule base includes: Apply the non-structural change filtering rules; if a candidate difference region is determined to be a positional offset within the allowable range, then filter it out. And / or, Apply topology consistency check rules to determine whether the structural changes indicated by candidate difference regions conform to predefined legal connection relationships.

[0042] Topology consistency check rules: For structural changes such as adding new wiring, the rule base R defines legal port connection relationships. The system will determine whether the new wiring is connected to an invalid or disabled port. If the rules are violated, it will be marked as a high-risk anomaly and the alarm level will be escalated.

[0043] Output results: Finally, the output is a list of high-confidence difference regions B and their corresponding category labels after being strictly filtered by the rule engine, which greatly improves the accuracy and operability of alarm information.

[0044] This embodiment provides a method for surveying work scenarios. This method is applied to the system described above, and its core steps include: Automated image acquisition: Through the coordinated operation of the mechanical support and motion module 100 and the image acquisition unit 300, the main control board 202 automatically controls the camera 301 to complete the acquisition of multi-angle and multi-position image sequences inside the cabinet according to a predetermined path planning algorithm.

[0045] Scene image generation: The image processor 201 executes the above-mentioned panoramic image generation algorithm on the acquired multi-angle image sequence to generate a high-fidelity panoramic image of the inside of the cabinet.

[0046] Intelligent status change detection: The current panoramic image generated during the inspection is compared with historically uploaded and stored baseline panoramic images. This comparison process is executed by the image processor 201 using the **difference detection algorithm described in Example 4**, which automatically identifies and outputs high-confidence equipment status change information.

[0047] In a preferred embodiment, the step of generating the panoramic image includes executing a panoramic image generation algorithm; the step of detecting state changes includes executing a difference detection algorithm.

[0048] This embodiment provides a tool to assist in the surveying of work scenarios. This tool is a physical and integrated product manifestation of the aforementioned system. Its specific structure is as follows: Camera 301: the aforementioned high-definition industrial-grade camera 301; Fill light 302: namely the aforementioned ring array LED fill light 302; Main control board 202: As the edge computing and control center of this tool, it not only runs embedded software to control all hardware units, but also stores the automatic path planning algorithm and the panoramic image generation algorithm. In other words, the main control board 202 contains the program code required to implement all the aforementioned system functions.

[0049] Electrically controlled lifting pole 101 and tripod: forming the support and vertical movement framework of the tool; Electronically controlled gimbal 103: Enables precise horizontal movement; Driver 102: Integrated inside the tool, responsible for driving moving parts.

[0050] The tool runs an automatic path planning algorithm and a panoramic image generation algorithm through the program stored in its main control board 202, and is a complete and independent product invention.

[0051] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A work scene reconnaissance auxiliary system, characterized in that, include: The image acquisition unit, including a camera and a supplementary light, is used to acquire multi-angle images of the work scene; The mechanical support and motion module, including the electrically controlled lifting rod, the electrically controlled pan-tilt unit, and the driver, is used to mount and control the image acquisition unit to perform multi-degree-of-freedom motion within the workspace; The control and processing unit includes a main control board and an image processor that is connected in communication with it; The main control board is configured with an automatic path planning algorithm to control the mechanical support and motion module to move along a predetermined trajectory, and simultaneously triggers the image acquisition unit to acquire images. The image processor is configured with an image processing algorithm to process the acquired multi-angle images to generate panoramic images of the work scene and / or status change information.

2. The system according to claim 1, characterized in that, The image processing algorithm includes a panoramic image generation algorithm, and the panoramic image generation algorithm executed by the image processor includes the following steps: Feature point extraction and matching are performed on the input multi-angle image set to obtain an initial set of matching point pairs; Based on prior knowledge of the physical layout of the equipment in the aforementioned work scenario, geometric constraint filtering is applied to the initial set of matching point pairs to eliminate mismatched point pairs and obtain a set of matching points that meets the preset purity requirements. The geometric constraints include directional consistency constraints, which are used to filter out matching point pairs whose principal direction difference is approximately 0° or an integer multiple of 90°. The spatial transformation relationship between images is calculated using the matching point set, and image stitching and fusion are performed to generate the panoramic image.

3. The system according to claim 2, characterized in that, The geometric constraints also include spatial distribution consistency constraints, and the steps include: Calculate the candidate space transformation model based on the candidate matching point set; Calculate the projection error of each matching point under the candidate model; Based on the statistical distribution of the projection error, matching points whose errors significantly deviate from the mainstream distribution are eliminated.

4. The system according to claim 1, characterized in that, The control and processing unit is further configured to execute a real-time image quality assessment and control algorithm, the algorithm including: During image acquisition, the sharpness evaluation value of the currently acquired image is calculated in real time; wherein, the sharpness evaluation value is the variance of the image after processing by the Laplacian operator; The clarity evaluation value is compared with a preset threshold. If the clarity evaluation value is lower than the preset threshold, the main control board immediately controls the image acquisition unit to re-acquire the image at the same location.

5. The system according to claim 1, characterized in that, The difference detection algorithm executed by the image processor includes: The feature maps of the baseline panoramic image and the current panoramic image are extracted using a Siamese neural network, and a difference heatmap is generated. Based on a preset device topology rule base, logical reasoning and filtering are performed on the candidate difference regions identified from the difference heatmap; Output a filtered list of high-confidence disparity regions and their category labels.

6. The system according to claim 5, characterized in that, The steps for logical reasoning and filtering based on the device topology rule base include: Apply non-structural change filtering rules; if a candidate difference region is determined to be a positional offset within the allowable range, then filter it out.

7. The system according to claim 5, characterized in that, The steps for logical reasoning and filtering based on the device topology rule base include: Apply topology consistency check rules to determine whether the structural changes indicated by candidate difference regions conform to predefined legal connection relationships.

8. A method for surveying work scenarios, characterized in that, The method is applied to the system as described in any one of claims 1 to 6, and includes: Through the coordinated operation of the mechanical support and motion module and the image acquisition unit, multi-angle image sequences of the work scene are automatically acquired; The multi-angle image sequence is processed to generate a panoramic image of the work scene; The currently generated panoramic image is compared with historical baseline panoramic images to detect changes in the state of the work scene.

9. The method according to claim 8, characterized in that, The step of generating the panoramic image includes executing a panoramic image generation algorithm; the step of detecting state changes includes executing a difference detection algorithm.

10. A tool for assisting in the reconnaissance of work scenarios, characterized in that, include: Cameras are used to capture multi-angle images of the work scene; The supplementary light integrates a ring array LED light source and has intelligent adjustment functions for color temperature and brightness. The main control board, as the edge computing and control hub, runs embedded software to control image acquisition, motion unit coordination, and data communication. The electrically controlled lifting pole and tripod are used to support and control the camera's vertical movement; An electronically controlled pan-tilt head is used to control the horizontal tilt and rotation of a camera. The driver is used to receive signals from the main control board and drive the electric lifting rod and the electric pan-tilt unit to move. The main control board executes a path planning algorithm to achieve automated multi-angle image acquisition; the main control board stores the automatic path planning algorithm and the panoramic image generation algorithm.

Images