Communication machine room inspection device and method based on computer vision

Abstract

The invention belongs to the technical field of machine room inspection, and particularly discloses a communication machine room inspection device and method based on computer vision, and the device comprises an inspection vehicle, a buffer seat, a loading seat, a camera, an execution mechanism and an image processing system, the buffering seat is provided with mechanical anti-shake structures such as a sliding groove, a sliding block, a movable frame and a spring damper, and the influence of vehicle body vibration on image acquisition stability can be effectively reduced. The device further comprises functional components such as a laser distance sensor and a mechanical arm. The method comprises the steps of image acquisition, image comparison, abnormity identification, operation execution, pressure feedback control and the like, and can realize automatic abnormity identification, target positioning, intervention execution and alarm and report generation. The intelligent, automatic and anti-interference capabilities of communication machine room inspection are improved, and the system is suitable for high-frequency, large-scale and multi-type equipment inspection scenes.

Description

Technical Field

[0001] This invention belongs to the field of computer room inspection technology, specifically relating to a communication computer room inspection device and method based on computer vision. Background Technology

[0002] As the core operating environment of data communication systems, communication equipment rooms typically house a large number of servers, switches, power supply equipment, and environmental monitoring modules. To ensure the long-term stable operation of various communication devices, high-frequency and high-precision inspections are required. Traditional inspections mainly rely on manual patrols, using manual observation and recording to check the appearance, operating status, and environmental parameters of equipment. This method not only suffers from low inspection efficiency and a high rate of false positives and false negatives, but is also limited by the professional level and subjective judgment of the inspectors, making it difficult to achieve standardized and real-time inspection management.

[0003] In recent years, with the application of technologies such as image recognition, deep learning, and edge computing, computer vision-based automated inspection equipment has gradually developed. In existing technologies, a typical approach involves mounting cameras or image sensors on mobile platforms (such as tracked robots or vehicle chassis) to collect images of equipment in communication rooms. These images are then processed using image processing algorithms to identify and judge the equipment's operating status, thereby achieving automatic detection of equipment malfunctions. However, existing technologies still generally suffer from the following shortcomings: First, during the movement of the inspection device, especially when encountering complex terrain such as uneven ground, cable obstacles, protrusions, or small steps, the image acquisition equipment is prone to severe shaking or vibration, resulting in blurred images, information loss, or even misjudgment. Existing inspection platforms mostly lack effective anti-shake structures or rely solely on electronic image stabilization algorithms, making it difficult to buffer physical displacement under strong vibration conditions, thus limiting their stability. Summary of the Invention

[0004] This invention addresses the problems in the prior art by providing a communication equipment room inspection device and method based on computer vision, which solves the problem of decreased inspection accuracy caused by blurry images captured by mobile cameras in the prior art.

[0005] The technical solution adopted in this invention is as follows: In a first aspect, this application provides a communication equipment room inspection device, including an inspection vehicle. A buffer seat is fixedly connected to the top of the inspection vehicle. Two slide grooves are opened on the top of the buffer seat. Two sliders are slidably arranged in each slide groove. A spring damper is provided on the opposite side of the two sliders and the spring damper is arranged in the slide groove. A movable frame is rotatably connected to the adjacent side of the two sliders. A connector is rotatably connected between the two movable frames. A loading seat is fixedly connected to the top of the connector. A camera is arranged on the loading seat.

[0006] Preferably, an actuator is fixedly connected to the top of the loading seat, a placement slot is opened on the top of the loading seat, a first motor is fixedly connected in the placement slot, a turntable is fixedly connected to the output end of the first motor, an adjustment component is provided on the top of the turntable, a mounting base is connected to the top of the adjustment component, and a camera is provided on the top of the mounting base.

[0007] Preferably, the adjustment assembly includes a fixed base and a second motor. Both the fixed base and the second motor are fixedly connected to the turntable. A rotating rod is rotatably connected through the fixed base. Two connecting plates are fixedly connected to the outer ring of the rotating rod. The connecting plates are located on both sides of the fixed base. The top of both connecting plates is fixedly connected to the mounting base. A worm gear is fixedly connected to the right end of the rotating rod. A worm is fixedly connected to the output end of the second motor. The worm is located below the worm gear and is meshed with the worm gear.

[0008] Preferably, the top of the mounting base has a mounting groove, and two clamping blocks are slidably arranged in the mounting groove. A third motor is fixedly connected to the outside of the mounting base, and the output end of the third motor is fixedly connected to a bidirectional threaded rod. The outer ring of the bidirectional threaded rod is threaded to the two clamping blocks on both sides respectively. Limiting grooves are opened on the adjacent side of the two clamping blocks, and the camera is set in the limiting groove between the two clamping blocks.

[0009] Preferably, a buffer pad is provided in the limiting groove, and anti-slip ridges are provided on the contact surface between the buffer pad and the camera.

[0010] Preferably, a mounting frame is fixedly connected to the front end of the loading seat, and a lighting lamp is fixedly connected to the outside of the mounting frame.

[0011] Preferably, the camera is equipped with a laser distance sensor, and the optical axis of the laser distance sensor is set parallel to the optical axis of the camera.

[0012] Preferably, the actuator is a six-degree-of-freedom robotic arm, with an end effector at the end of the robotic arm. The end effector is an electric gripper, and a pressure sensor is installed between the electric gripper and the robotic arm. The two ends of the pressure sensor are respectively in contact with the electric gripper and the robotic arm.

[0013] Secondly, this application provides a method for inspecting a communication equipment room, employing the inspection device described in the first aspect, the method comprising: Control the inspection vehicle to move at a constant speed along a preset path in the computer room; Meanwhile, images are captured in real time using a camera mounted on the top of the inspection vehicle as it moves. During image acquisition, the camera's horizontal rotation is achieved by driving the turntable to rotate via the first motor according to the shooting requirements, and the camera's pitch angle is adjusted by driving the worm gear mechanism via the second motor. When vibrations occur during the movement of the inspection vehicle or when the inspection vehicle crosses an obstacle, the vibration is transmitted to the camera. During this process, the vibration is physically absorbed by the spring damper as it is transmitted from the inspection vehicle to the buffer seat, spring damper, slider, and movable frame.

[0014] Preferably, after the acquired image is preprocessed, it is compared with a preset comparison image; when an anomaly is found, the abnormal image is compared with the image in the preset abnormal operation table; when a comparison result is found, the corresponding preset control signal in the abnormal operation table is read and the corresponding preset control signal is sent to the actuator to handle the abnormal situation. If no comparison result is found, a corresponding inspection alarm signal is generated and sent to the remote monitoring center or the maintenance personnel terminal. At the same time, an inspection report is generated and stored in the historical database in the system. After the alarm is completed, the inspection continues until the inspection task of the preset path is completed. Upon detecting an anomaly, the inspection vehicle was brought to a stop, and images of the anomaly were captured. The abnormal image is compared with the image in the preset abnormal operation table. When a comparison result is found, the camera optical axis is aligned with the abnormal point, and the laser distance sensor collects the distance information between the camera and the object being measured in real time. By acquiring distance information and camera angle information of abnormal location, abnormal location information can be calculated, and the closest abnormal location information can be found in the abnormal operation table to perform the execution operation of the actuator. During operation, the pressure value of the pressure sensor is acquired in real time. When the pressure value exceeds the pressure threshold, the execution operation of the actuator is canceled, and a corresponding inspection alarm signal is generated and sent to the remote monitoring center or the terminal of the operation and maintenance personnel. At the same time, an inspection report is generated and stored in the historical database of the system.

[0015] As can be seen from the above technical solutions, the advantages of the present invention are: (1) This device, through the mechanical anti-shaking structure consisting of a chute, slider, movable frame, connector and spring damper, can effectively absorb the interference caused by external vibration or impact during the operation of the inspection vehicle, thereby achieving buffering and stable support for the loading seat and camera, avoiding image blurring and displacement, and improving image acquisition accuracy and system operation stability.

[0016] (2) By setting up a turntable and adjustment components with electric drive and multi-degree-of-freedom adjustment capabilities on the loading seat, the camera can be flexibly adjusted in the horizontal and vertical directions, enabling the inspection device to have a wider observation range and stronger target tracking capabilities, adapting to the complex layout in the communication equipment room, and improving the efficiency of target detection and image coverage.

[0017] The adjustment component adopts a structure in which a motor drives a worm gear and a worm wheel to mesh, which can effectively achieve precise adjustment of the angle of the mounting base and has a good self-locking function to prevent the camera position from shifting during use. This helps to maintain the consistency of image recognition direction in multiple application scenarios and improve the recognition accuracy and positioning reliability of the device.

[0018] (3) Setting up a laser distance sensor on the camera can enable precise distance measurement between the camera and the target, assist in completing functions such as image depth perception and abnormal position judgment, improve the system's ability in spatial analysis, stereo positioning and three-dimensional recognition, and enhance the device's perception dimension and analysis accuracy.

[0019] Employing a six-degree-of-freedom robotic arm as the actuator, and equipped with an electric gripper and pressure sensor, it can complete complex motion path planning and precise operation tasks; the pressure sensor monitors the gripping force in real time to ensure safety and stability during the process, and is suitable for various scenarios such as troubleshooting, button operation, or picking up and placing items.

[0020] (4) The method is based on the image acquisition and comparison process, which realizes automatic identification, anomaly judgment and branch response of equipment operation status. It has complete automatic inspection closed-loop control logic, which can reduce the frequency of manual operation, improve inspection efficiency and accuracy, and support the actual needs of intelligent operation and maintenance of large-scale communication equipment rooms.

[0021] By using abnormal image localization, laser ranging, attitude analysis, and clamping pressure control, the system achieves intelligent response and safe handling of abnormal equipment. It completes autonomous interactive operation while ensuring that the equipment is not damaged, thereby improving the autonomous decision-making ability and fault tolerance of the automatic inspection device and enhancing the system's ability to cope with complex scenarios and its practicality. Attached Figure Description

[0022] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are 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.

[0023] Figure 1 A schematic diagram of a computer vision-based communication equipment room inspection device provided for an embodiment of the present invention; Figure 2 An exploded view of the structure of a computer vision-based communication equipment room inspection device provided for an embodiment of the present invention; Figure 3 An exploded view of the adjustment components of a computer vision-based communication equipment room inspection device provided for an embodiment of the present invention; Figure 4This is a schematic diagram of a data center inspection system according to an embodiment of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] Example 1: Please see Figures 1-3 As shown, the present invention provides a communication equipment room inspection device based on computer vision, including an inspection vehicle 1. A buffer seat 2 is fixedly connected to the top of the inspection vehicle 1. Two slide grooves 3 are opened on the top of the buffer seat 2. Two sliders 4 are slidably arranged in each slide groove 3. A spring damper 7 is provided on the opposite side of the two sliders 4 and the spring damper 7 is arranged in the slide groove 3. A movable frame 5 is rotatably connected to the adjacent side of the two sliders 4. A connector 6 is rotatably connected between the two movable frames 5. A loading seat 8 is fixedly connected to the top of the connector 6. A camera 24 is arranged on the loading seat 8.

[0026] A mechanical buffer and shock absorption structure is formed by installing a buffer seat 2 on the top of the inspection vehicle 1, along with a slide 3, a slider 4, a movable frame 5, a connector 6, and a spring damper 7. When the inspection vehicle 1 vibrates while running on the complex floor of the machine room, the slider 4 slides in the slide 3 and compresses the spring damper 7 to absorb the vibration energy, thereby reducing the transmission of vibration to the loading seat 8 and ensuring the operational stability of the camera 24.

[0027] In a preferred embodiment, the inspection vehicle 1 is used for automatic movement within the communication equipment room. A buffer seat 2 is mounted on its top, and the buffer seat 2 has multiple grooves 3. Sliding sliders 4 are installed within the grooves 3. The side of the slider 4 furthest from the loading end is connected to a spring damper 7, which is installed within the groove 3 to absorb vibrations and impacts during vehicle operation. The side of the slider 4 is rotatably connected to a connecting member 6 via a movable frame 5. A loading seat 8 is fixed above the connecting member 6, and a camera 24 is mounted on the upper part of the loading seat 8 to acquire images of the communication equipment room.

[0028] In this embodiment, an actuator 9 is fixedly connected to the top of the loading seat 8, and a placement slot 10 is opened on the top of the loading seat 8. A first motor 11 is fixedly connected in the placement slot 10, and a turntable 12 is fixedly connected to the output end of the first motor 11. An adjustment component is provided on the top of the turntable 12, and a mounting base 19 is connected to the top of the adjustment component. A camera 24 is provided on the top of the mounting base 19.

[0029] The upper functional structure of the loading base 8 has been further expanded by adding an electric drive control component and a mounting module. The first motor 11 drives the turntable 12 to rotate, thereby rotating the upper adjustment component to achieve horizontal rotation adjustment of the camera 24. At the same time, the camera 24 is mounted on the top of the mounting base 19 for easy vertical or tilt angle adjustment.

[0030] In a further preferred embodiment, an electric actuator 9 for performing operations is mounted on the top of the loading seat 8. The upper part of the loading seat 8 has a placement slot 10 for mounting a transmission device, and a first motor 11 is installed in the slot. The motor output is connected to a turntable 12. The turntable 12 drives the upper adjustment mechanism to rotate, thereby enabling the camera 24 to have horizontal viewing angle adjustment capability. A mounting base 19 is fixed above the adjustment assembly, facilitating subsequent modular setup and installation of image acquisition equipment.

[0031] In this embodiment, the adjustment assembly includes a fixed base 13 and a second motor 17. Both the fixed base 13 and the second motor 17 are fixedly connected to the turntable 12. A rotating rod 14 is rotatably connected through the fixed base 13. Two connecting plates 15 are fixedly connected to the outer ring of the rotating rod 14. The connecting plates 15 are disposed on both sides of the fixed base 13. The tops of the two connecting plates 15 are fixedly connected to the mounting base 19. A worm gear 16 is fixedly connected to the right end of the rotating rod 14. A worm 18 is fixedly connected to the output end of the second motor 17. The worm 18 is disposed below the worm gear 16 and is meshed with the worm gear 16.

[0032] The above describes a mechanism for adjusting the pitch angle of the camera 24, namely a worm gear drive-based tilt adjustment device. This structure is compact, provides stable transmission, has self-locking characteristics, and can precisely control the tilt angle of the camera 24, preventing it from rotating or shifting during operation.

[0033] In a further preferred embodiment, the adjustment assembly includes a fixed base 13 and a second motor 17 fixedly mounted on the turntable 12. A rotatable rotating rod 14 is inserted inside the fixed base 13, and connecting plates 15 for supporting the mounting base 19 are connected to both sides of the rotating rod 14. The connecting plates 15 transmit the rotation of the rotating rod 14 to the mounting base 19, thereby enabling the mounting base 19 to flip. To achieve power transmission, a worm gear 16 is provided at the end of the rotating rod 14, and a worm 18 meshing with it is connected to the output end of the second motor 17. When the motor rotates, it drives the worm 18 to mesh with the worm gear 16, thereby achieving precise control of the pitch angle of the camera 24.

[0034] In this embodiment, the top of the mounting base 19 is provided with a mounting groove 20, and two clamping blocks 22 are slidably arranged in the mounting groove 20. A third motor 25 is fixedly connected to the outside of the mounting base 19, and the output end of the third motor 25 is fixedly connected to a bidirectional threaded rod 21. The outer ring of the bidirectional threaded rod 21 is threadedly connected to the two clamping blocks 22 on both sides. A limit groove 23 is provided on the adjacent side of the two clamping blocks 22, and the camera 24 is arranged in the limit groove 23 between the two clamping blocks 22.

[0035] The above describes a clamping structure for fixing the camera 24. The bidirectional threaded rod 21 drives the clamping block 22 to clamp the camera 24, and the limiting groove 23 is used to fix the position. The structure is simple, easy to disassemble and assemble, convenient for maintenance or replacement of the camera 24, and can ensure the positional stability when clamped.

[0036] In a further preferred embodiment, the mounting base 19 has an open mounting groove 20 on its upper part, and a rotatable bidirectional threaded rod 21 passes through the groove. Two clamping blocks 22 are connected to both ends of the threaded rod for clamping the camera 24. A third motor 25 is provided on one side of the mounting base 19. The motor drives the threaded rod to rotate, causing the two clamping blocks 22 to move towards or away from each other. The clamping blocks 22 have a limiting structure inside for precise positioning of the camera 24, ensuring stability and consistency during image acquisition.

[0037] In this embodiment, a buffer pad is provided in the limiting groove 23, and anti-slip texture is provided on the contact surface of the buffer pad that is in contact with the camera 24.

[0038] The clamping mechanism has been enhanced with a buffer and anti-slip design, which further improves the shock resistance and installation stability of the camera 24, prevents damage to the camera 24 housing due to clamping or slippage due to vibration, and improves the reliability of clamping.

[0039] In a further preferred embodiment, a flexible cushioning material is affixed to the limiting groove 23 on the clamping block 22 to absorb minor impacts between the camera 24 and the clamping block 22. Simultaneously, the surface of the cushioning pad has several raised structures, providing an anti-slip effect and enhancing the stability of the camera 24 during clamping, preventing slippage or deflection due to vibration or tilting.

[0040] In this embodiment, a fixing frame 26 is fixedly connected to the front end of the loading seat 8, and a lighting lamp 27 is fixedly connected to the outside of the fixing frame 26.

[0041] An auxiliary lighting module is described to meet supplemental lighting needs, which enhances the imaging quality of the camera in low-light or shadow areas, expands the applicable scenarios for inspection, and improves recognition stability.

[0042] In a further preferred embodiment, the front end of the mounting base 8 is provided with a fixed support structure for mounting a lighting source. The lighting lamp 27 can be automatically turned on according to the inspection task to provide auxiliary lighting for the camera 24 in low-light or complex lighting environments, ensuring the brightness and clarity of image acquisition and improving recognition accuracy.

[0043] In this embodiment, a laser distance sensor 28 is provided on the camera 24, and the optical axis of the laser distance sensor 28 is arranged parallel to the optical axis of the camera 24.

[0044] By setting a laser distance sensor 28 on the camera 24 and keeping it parallel to the optical axis of the camera 24, distance calibration and spatial coordinate acquisition can be achieved when acquiring images, providing data support for target positioning and depth recognition.

[0045] In a further preferred embodiment, a laser ranging module is provided above the camera 24, and the direction of its emitted light is consistent with the imaging optical path of the camera 24. This sensor can collect distance information between the camera 24 and the object being measured in real time, providing auxiliary support for 3D positioning and image correction, and improving the system's spatial perception capability.

[0046] In this embodiment, the actuator 9 is a six-degree-of-freedom robotic arm. An end effector is provided at the end of the robotic arm. The end effector is an electric gripper. A pressure sensor is provided between the electric gripper and the robotic arm. The two ends of the pressure sensor are respectively in contact with the electric gripper and the robotic arm.

[0047] This section introduces a robotic arm with multi-degree-of-freedom control capabilities to perform auxiliary operations after image recognition. By setting a pressure feedback device between the gripper and the body, the gripping force can be monitored in real time to ensure operational safety, making it suitable for performing physical intervention tasks.

[0048] In a further preferred embodiment, the actuator 9 adopts a six-axis robotic arm structure, possessing multi-dimensional motion capabilities, and can flexibly reach various detection or intervention positions within the communication equipment room. The end of the robotic arm is equipped with an electrically controlled gripper for gripping or pushing the target object. A pressure detection element is installed between the gripper and the robotic arm to detect the contact pressure during the gripping process in real time, preventing damage to the target equipment due to excessive force and achieving safe control.

[0049] Example 2: This invention provides a method for inspecting a communication equipment room, the method comprising: Step S1: Control the inspection vehicle to move at a constant speed in the computer room along the preset path; When vibrations occur during the movement of the inspection vehicle or when the inspection vehicle crosses an obstacle, the vibration is transmitted to the camera by the inspection vehicle, buffer seat, spring damper, slider, and movable frame. The spring damper physically absorbs the vibration during this process. During image acquisition, the camera's horizontal rotation is achieved by driving the turntable to rotate via the first motor according to the shooting requirements, and the camera's pitch angle is adjusted by driving the worm gear mechanism via the second motor. Meanwhile, images are captured in real time using a camera mounted on the top of the inspection vehicle as it moves. The inspection vehicle moves at a constant speed within the communication equipment room along a pre-set route. During the movement, cameras continuously capture images of the front of the equipment, ports, indicator lights, and other parts, ensuring comprehensive inspection of the equipment's status. The path planning and control module ensures the inspection route is controllable and repeatable, facilitating subsequent data comparison and reproduction. In a preferred embodiment, the control system drives the inspection vehicle to move at a constant speed within the communication equipment room according to a pre-set inspection path, avoiding the impact of speed fluctuations on image acquisition stability. The camera mounted on the inspection vehicle continuously acquires images throughout the entire journey, and the acquired images serve as the raw input for subsequent image recognition and comparison. The camera's shooting angle, acquisition frequency, and aperture parameters can be automatically adjusted according to a preset program to adapt to different equipment room scenarios and equipment layouts. Step S2: After the acquired image is preprocessed, it is compared with a preset comparison image; if an anomaly is found, proceed to step S3. The system processes the images acquired in step S1 through an image preprocessing module, performing noise reduction, enhancement, and alignment to improve image clarity and analysis results. The preprocessed image is then compared with a preset standard or baseline image to determine if the device is in normal working order. If no anomalies are found during the comparison, the subsequent inspection path continues; if a suspected anomaly is detected, the process jumps to the next processing stage. After acquiring an image, the image acquisition module immediately transmits it to the data processing module. The data processing module performs standardization processing on the image, including image grayscale correction, edge extraction, and noise filtering. The processed image is then compared with preset normal equipment image samples within the inspection system to determine if the current equipment has any abnormal status or appearance deviation. If the image feature matching degree is within the set range, the system determines that the current equipment is normal and controls the system to continue executing subsequent inspection tasks; otherwise, it enters the anomaly analysis process. Step S3: Compare the abnormal image with the images in the preset abnormal operation table. If a comparison result exists, proceed to step S4; otherwise, generate the corresponding inspection alarm signal and send it to the remote monitoring center or maintenance personnel terminal. At the same time, generate an inspection report and store it in the historical database in the system. After the alarm is sent, proceed to step S5. If the comparison reveals abnormal features in the image, it is further compared with the system's preset abnormal image samples to identify the type and location of the anomaly. If a match is found, the anomaly handling process begins; if the specific type cannot be identified, the system immediately generates a corresponding anomaly alarm signal and uploads it to the remote center, while simultaneously generating and archiving an inspection report to provide data support for subsequent operation and maintenance. When the system detects an abnormal image, it further inputs the image into the anomaly comparison module for one-to-one matching and analysis with standard abnormal images in the built-in abnormal image database. If the comparison result clearly points to a certain type of fault, such as equipment disconnection, abnormal indicator light, or panel damage, the system enters the preset response process; if the comparison fails, the system immediately sends the abnormal status as an alarm message to the remote monitoring center or maintenance personnel via the communication module, and generates a detailed inspection report, including image timestamps, descriptions of abnormal features, and equipment location information. Simultaneously, the record is stored in the historical database for tracking. Step S4: Read the corresponding preset control signal in the abnormal operation table, send the corresponding preset control signal to the actuator to handle the abnormal situation, and then jump to step S5; Based on the successfully matched anomaly type, the system extracts the corresponding control commands from the anomaly operation table and issues them to the actuator. The actuator performs actions including reset, button operation, interface plugging / unplugging, and obstruction removal, achieving preliminary automatic elimination of some common faults, improving the system's autonomous response capability, and reducing manual intervention. When an abnormal image is successfully matched to a specific fault category, the inspection system automatically reads the corresponding operation command and sends it to the actuator mounted on the inspection vehicle. After the actuator is activated, it performs the specific operation according to the control logic, such as operating the electric gripper to press a button or reconnect a power-off interface, thus completing the initial intervention or restoration of the abnormal equipment. After the operation is completed, the system records the process and marks it as "processed," preparing to continue to the next inspection step. Step S5: Continue the inspection until the inspection task along the preset path is completed. Ensure that after the system completes a certain abnormal response or information upload, it can continue to execute the remaining inspection tasks, ensuring that all devices within the preset path are inspected and covered, thereby improving task integrity and process continuity. After completing anomaly handling or alarm generation, the system automatically switches the control flow back to normal inspection mode and continues to perform the next stage of equipment image acquisition and analysis along the preset path. The system continuously monitors the current progress through the path management module until all set inspection points have been processed, at which point it automatically terminates the task or waits for the start command of the next cycle. The entire process achieves a fully automated closed-loop inspection without manual intervention, improving inspection efficiency and coverage accuracy.

[0050] In this embodiment, after an anomaly is detected, the inspection vehicle is stopped and an image of the anomaly is captured. When the system detects any abnormal equipment status or changes in appearance in the image, the inspection vehicle immediately stops moving to facilitate subsequent operations, and the camera remains stationary, focusing on the abnormal area to take a fixed-point image. This operation ensures the acquisition of high-quality, stable images of the abnormality, providing a clear basis for subsequent anomaly identification and processing.

[0051] In a preferred embodiment, when the inspection system determines that the currently acquired image is inconsistent with the reference image and exhibits abnormal characteristics, it immediately issues a stop command to the inspection vehicle, causing the vehicle to stop moving at its current position. The control system retrieves images from the camera for local magnification or precise focusing, acquiring multi-angle, continuous images of the abnormal area. The captured images are then used as key inputs to the anomaly identification module to support subsequent anomaly type comparison and processing.

[0052] The abnormal image is compared with the image in the preset abnormal operation table. When a comparison result is found, the camera optical axis is aligned with the abnormal point, and the laser distance sensor collects the distance information between the camera and the object being measured in real time. The system uses an image comparison algorithm to match and identify the current abnormal image with abnormal sample images in a preset abnormal operation table. If the match is successful, the system adjusts the camera orientation to ensure that the optical axis is aligned with the abnormal device's location, and simultaneously activates a laser rangefinder to collect the spatial distance between the camera and the target object, providing spatial reference for subsequent precise positioning and operational decisions.

[0053] When an abnormal image is determined by the comparison module to match a certain anomaly type within the system, the system controls the camera to adjust its angle based on the target's feature information in the image, ensuring its optical axis is precisely aligned with the target anomaly point. Simultaneously, the laser ranging module is activated to measure the straight-line distance between the camera and the target device in real time. These operations facilitate subsequent location of the anomaly's 3D coordinates, enabling accurate identification and spatial modeling of the anomaly.

[0054] By acquiring distance information and camera angle information of the abnormal location, the system can calculate the abnormal location information, find the closest abnormal location information in the abnormal operation table, and execute the operation of the actuator. Based on the camera attitude angle and laser ranging results, the system establishes the spatial relative relationship between the camera and the target object, and combines the image information to deduce the spatial coordinates of the abnormal point. Then, it finds the operation task closest to the coordinates in the abnormal operation table, generates control commands to drive the actuator to perform the corresponding operation, such as pushing, pressing, clearing, etc., to complete the initial handling task.

[0055] The system analyzes the current shooting angle of the camera, the direction of the optical axis, and the ranging data provided by the laser distance sensor to establish a spatial geometric model and calculate the spatial location of the malfunctioning device. It then calls the malfunction operation table and matches the calculated location information with the corresponding processing action, such as controlling a robotic arm to move to the target location and performing operations like removing foreign objects or reconnecting interfaces. This approach ensures a precise correspondence between the executed actions and the target point, improving the automation level of malfunction handling.

[0056] During the operation, the pressure value of the pressure sensor is acquired in real time. When the pressure value exceeds the pressure threshold, the execution operation of the actuator is canceled, and a corresponding inspection alarm signal is generated and sent to the remote monitoring center or the terminal of the operation and maintenance personnel. At the same time, an inspection report is generated and stored in the historical database in the system. After the alarm is sent, the process jumps to step S5.

[0057] During the handling of anomalies, the actuator monitors the applied force in real time using pressure sensors. When the detected pressure exceeds the system's set safety threshold, indicating a risk of abnormality such as excessive clamping or target blockage, the system immediately suspends the operation and notifies the remote monitoring terminal via an alarm mechanism. This ensures the safety of the equipment and the target and prevents secondary failures caused by misoperation.

[0058] During the actual operation of the actuator, the system continuously acquires pressure sensor output signals from the gripper or end effector to monitor the contact force on the manipulated object. When the system detects that the pressure value exceeds a preset safety threshold, it immediately interrupts the current execution action, controls the actuator to return to its initial position, and generates an alarm message. The alarm message includes the reason for the operation failure, the abnormal pressure value, the operation time, and object information. Simultaneously, the system uploads the relevant data to a remote platform and records it in a local database for subsequent inspection history analysis and maintenance decision-making. After processing, the system automatically returns to the main inspection process to continue executing subsequent tasks.

[0059] Example 3: Please see Figure 4As shown, this invention provides a data center inspection system, which includes a data processing module, a control module, a storage module, and a communication module. The data processing module includes image preprocessing, feature extraction, and pattern recognition. Image preprocessing is used to reduce noise and enhance the acquired images to improve the accuracy of subsequent processing. Feature extraction uses deep learning algorithms to extract features from the images, such as edge, texture, and color information. Pattern recognition inputs the extracted features into a pre-trained classifier to identify the equipment status, such as normal, abnormal, or faulty. The control module includes inspection report generation and alarm signal generation. Inspection report generation allows the control module to generate a detailed inspection report based on the results of the data processing module, including equipment status, abnormal locations, and recommendations. The maintenance measures include alarm signal generation. When the data processing module detects an equipment abnormality, the control module generates an alarm signal and sends it to the remote monitoring center or designated maintenance personnel via the communication module. The storage module includes image information storage, inspection report storage, and historical record storage. Image information storage stores all collected image information, inspection report storage stores the report generated for each inspection, and historical record storage stores historical inspection records for trend analysis and fault prediction. The communication module includes data transmission and protocol support. Data transmission sends inspection reports and alarm signals to a remote server or designated maintenance personnel via a selected communication protocol. Protocol support supports multiple communication protocols to ensure normal communication in different network environments.

[0060] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A computer vision-based communication equipment room inspection device, comprising an inspection vehicle (1), characterized in that: The inspection vehicle (1) has a buffer seat (2) fixedly connected to its top. The buffer seat (2) has two slide grooves (3) on its top. Two sliders (4) are slidably arranged in each slide groove (3). Spring dampers (7) are provided on opposite sides of the two sliders (4). The spring dampers (7) are located in the slide grooves (3). Movable frames (5) are rotatably connected to adjacent sides of the two sliders (4). Connectors (6) are rotatably connected between the two movable frames (5). Loading seat (8) is fixedly connected to the top of the connector (6). Camera (24) is installed on the loading seat (8).

2. The computer vision-based communication equipment room inspection device according to claim 1, characterized in that, An actuator (9) is fixedly connected to the top of the loading seat (8). A placement slot (10) is opened on the top of the loading seat (8). A first motor (11) is fixedly connected in the placement slot (10). A turntable (12) is fixedly connected to the output end of the first motor (11). An adjustment component is provided on the top of the turntable (12). A mounting base (19) is connected to the top of the adjustment component. A camera (24) is provided on the top of the mounting base (19).

3. The computer vision-based communication equipment room inspection device according to claim 2, characterized in that, The adjustment assembly includes a fixed base (13) and a second motor (17). Both the fixed base (13) and the second motor (17) are fixedly connected to the turntable (12). A rotating rod (14) is rotatably connected through the fixed base (13). Two connecting plates (15) are fixedly connected to the outer ring of the rotating rod (14). The connecting plates (15) are located on both sides of the fixed base (13). The tops of the two connecting plates (15) are fixedly connected to the mounting base (19). A worm gear (16) is fixedly connected to the right end of the rotating rod (14). A worm (18) is fixedly connected to the output end of the second motor (17). The worm (18) is located below the worm gear (16), and the worm (18) meshes with the worm gear (16).

4. The computer vision-based communication equipment room inspection device according to claim 2, characterized in that, The mounting base (19) has a mounting groove (20) on its top. Two clamping blocks (22) are slidably arranged in the mounting groove (20). A third motor (25) is fixedly connected to the outside of the mounting base (19). The output end of the third motor (25) is fixedly connected to the bidirectional threaded rod (21). The outer ring of the bidirectional threaded rod (21) is threaded to the two clamping blocks (22) on both sides respectively. A limit groove (23) is opened on the adjacent side of the two clamping blocks (22). The camera (24) is set in the limit groove (23) between the two clamping blocks (22).

5. The computer vision-based communication equipment room inspection device according to claim 5, characterized in that, A buffer pad is provided inside the limiting groove (23), and anti-slip ridges are provided on the contact surface between the buffer pad and the camera (24).

6. The computer vision-based communication equipment room inspection device according to claim 1, characterized in that, The loading seat (8) is fixedly connected to a mounting frame (26) at the front end, and a lighting lamp (27) is fixedly connected to the outside of the mounting frame (26).

7. The computer vision-based communication equipment room inspection device according to claim 3, characterized in that, A laser distance sensor (28) is provided on the camera (24), and the optical axis of the laser distance sensor (28) is set parallel to the optical axis of the camera (24).

8. The computer vision-based communication equipment room inspection device according to claim 2, characterized in that, The actuator (9) is a six-degree-of-freedom robotic arm. An end effector is provided at the end of the robotic arm. The end effector is an electric gripper. A pressure sensor is provided between the electric gripper and the robotic arm. The two ends of the pressure sensor are respectively in contact with the electric gripper and the robotic arm.

9. A method for inspecting a communication equipment room, characterized in that, The method using the inspection device according to any one of claims 1-8 includes: Control the inspection vehicle to move at a constant speed along a preset path in the computer room; Meanwhile, images are captured in real time using a camera mounted on the top of the inspection vehicle as it moves. During image acquisition, the camera's horizontal rotation is achieved by driving the turntable to rotate via the first motor according to the shooting requirements, and the camera's pitch angle is adjusted by driving the worm gear mechanism via the second motor. When vibrations occur during the movement of the inspection vehicle or when the inspection vehicle crosses an obstacle, the vibration is transmitted to the camera. During this process, the vibration is physically absorbed by the spring damper as it is transmitted from the inspection vehicle to the buffer seat, spring damper, slider, and movable frame.

10. The communication equipment room inspection method according to claim 9, characterized in that, After the acquired images are preprocessed, they are compared with preset comparison images. When an anomaly is found, the abnormal image is compared with the image in the preset abnormal operation table. If a comparison result is found, the corresponding preset control signal in the abnormal operation table is read and sent to the actuator to handle the abnormal situation. If no comparison result is found, a corresponding inspection alarm signal is generated and sent to the remote monitoring center or the maintenance personnel terminal. At the same time, an inspection report is generated and stored in the historical database in the system. After the alarm is completed, the inspection continues until the inspection task of the preset path is completed. Upon detecting an anomaly, the inspection vehicle was brought to a stop, and images of the anomaly were captured. The abnormal image is compared with the image in the preset abnormal operation table. When a comparison result is found, the camera optical axis is aligned with the abnormal point, and the laser distance sensor collects the distance information between the camera and the object being measured in real time. By acquiring distance information and camera angle information of abnormal location, abnormal location information can be calculated, and the closest abnormal location information can be found in the abnormal operation table to perform the execution operation of the actuator. During operation, the pressure value of the pressure sensor is acquired in real time. When the pressure value exceeds the pressure threshold, the execution operation of the actuator is canceled, and a corresponding inspection alarm signal is generated and sent to the remote monitoring center or the terminal of the operation and maintenance personnel. At the same time, an inspection report is generated and stored in the historical database of the system.

Images