A cabinet internal state detection system, detection method and detection device based on a limited field of view image, and a cabinet

Abstract

The application provides a cabinet internal state detection system based on a limited field of view image, a detection method and a detection device, and a cabinet, and relates to the technical field of industrial automation and intelligent operation and maintenance. The system comprises a mobile detection unit, a lifting driving unit, an edge computing device, a plurality of temperature and humidity sensors, a touch display screen and a cloud platform. The lifting driving unit drives the mobile detection unit to vertically ascend and descend through a steel wire rope winding and unwinding, and drives an image acquisition device to move along a horizontal guide shaft through a first motor, so that the image acquisition device completes image acquisition in the cabinet by reciprocating in the vertical direction. Meanwhile, the temperature and humidity sensors distributed in different temperature zones of the cabinet collect environmental data, and the image information and the temperature and humidity information are jointly input into the edge computing device, and the cabinet internal state detection result is output and uploaded to the cloud platform and the touch display screen. Without modifying the existing cabinet structure, the application can completely collect the overall appearance of the cabinet, and improve the integrity and detection accuracy of the cabinet state perception.

Description

Technical Field

[0001] This application relates to the fields of industrial automation and intelligent operation and maintenance technology, and in particular to a cabinet internal status detection system, detection method and detection device, and cabinet based on limited field of view images. Background Technology

[0002] With the rapid development of new infrastructure such as new energy power generation, artificial intelligence computing centers, and smart grids, various server racks, as key carriers for core equipment such as power conversion equipment, servers, communication equipment, and computing power equipment, are experiencing a continuous increase in quantity and complexity. In these application scenarios with high reliability requirements, the equipment inside the racks operates under high load, high density, and high temperature conditions for extended periods, making them highly susceptible to hazards such as equipment overheating, abnormal indicator lights, fan failure, missing equipment, or foreign object intrusion. If these issues are not detected and addressed in a timely manner, they can lead to minor malfunctions, or even serious consequences such as equipment burnout, system downtime, or fires, severely impacting energy supply, computing power services, or grid security.

[0003] Currently, the traditional methods for detecting the internal status of server racks mainly rely on regular manual inspections or the deployment of a small number of sensors (such as temperature and smoke sensors). However, manual inspections suffer from problems such as low efficiency, strong subjectivity, high missed detection rate, and inability to respond in real time. Sensors can only monitor a limited number of physical quantities and are difficult to fully perceive visual information such as the appearance status of equipment, indicator light color, and equipment type. In addition, the wiring is complex, the scalability is poor, and the maintenance cost is high.

[0004] In recent years, machine vision-based intelligent inspection technology, integrating optical imaging, image processing, pattern recognition, and artificial intelligence, has demonstrated strong potential in industrial automation and intelligent operation and maintenance. Its core components include image acquisition, preprocessing, feature extraction, target recognition, and decision output: high-quality images are acquired through high-resolution industrial cameras and light source systems, and image quality is improved through denoising, correction, and enhancement. Then, edge detection, contour extraction, template matching, or deep learning models are used to achieve high-precision real-time analysis of parameters such as equipment position, size, shape, color, and indicator light status. However, due to the highly compact space inside server racks, cameras can only be placed on top of the rack or in a fixed area, resulting in a severely limited field of view, making it difficult to fully capture the entire interior space. Under these limited field-of-view conditions, conventional image recognition methods struggle to accurately determine key parameters such as equipment position, size, type, and indicator light color, preventing existing systems from achieving reliable and comprehensive perception of the server rack's internal status. Summary of the Invention

[0005] To address the problem that existing technologies struggle to acquire a complete view of the internal space of a server rack under limited field of view conditions, thus hindering accurate monitoring of the rack's internal status, this invention proposes a rack internal status detection device, method, and detection system based on limited field of view images. This system can completely capture the entire internal space of the rack without modifying the existing rack structure, improving the completeness and accuracy of rack status perception and further enhancing the accuracy of rack internal status detection.

[0006] To achieve the above-mentioned technical effects, the technical solution of the present invention is as follows: A cabinet interior status detection system based on limited field-of-view images, comprising: The mobile detection unit includes a first motor, an image acquisition device, a horizontal transmission mechanism, and a horizontal guide shaft. The output shaft of the first motor is connected to one end of the horizontal transmission mechanism, and the other end of the horizontal transmission mechanism is connected to the image acquisition device. The image acquisition device is slidably mounted on the horizontal guide shaft. The two ends of the horizontal guide shaft are respectively provided with a first motor fixing component and a terminal fixing component. The first motor fixing component is fixedly installed on the first motor fixing component. The first motor fixing component and the terminal fixing component are slidably engaged with the original U-position columns extending vertically on both sides of the cabinet. The original U-position columns serve as vertical guide rails for the lifting and lowering of the mobile detection unit. The lifting drive unit includes a second motor, a second transmission mechanism, a rotating shaft, and a take-up reel. The second motor is mounted on the cabinet door, and its output shaft is connected to the rotating shaft via the second transmission mechanism. The rotating shaft is mounted on the top of the cabinet and rotates horizontally. Two take-up reels are fixedly installed at axial intervals along the shaft, and each take-up reel winds a steel wire rope. One end of each steel wire rope is connected to the first motor fixing component and the terminal fixing component, respectively. The rotating shaft drives the take-up reels to wind and unwind the steel wire rope, thereby causing the first motor fixing component and the terminal fixing component to move vertically up and down along the two sides of the cabinet's original U-shaped support columns. An edge computing device includes a control module, a communication module, and a processing module. The control module controls the operation of the first motor and the second motor. The communication module transmits image information of the cabinet interior acquired by the image acquisition device to the processing module. The processing module is equipped with a pre-trained state detection model. Based on the image information, the state detection model is used to obtain the state detection results of the cabinet interior. The image information includes: the location, dimensions, type, shape, color, and status indicator color of the equipment within the cabinet.

[0007] Preferably, it further includes: a plurality of temperature and humidity sensors, which are respectively deployed in different temperature zones within the cabinet, and collect temperature and humidity information inside the cabinet in their respective corresponding temperature zones. The temperature and humidity information inside the cabinet and the image information are jointly input into the state detection model to obtain the state detection result inside the cabinet.

[0008] Preferably, it further includes: a cloud platform, which is communicatively connected to the edge computing device and receives and analyzes the cabinet internal status detection results transmitted by the edge computing device.

[0009] Preferably, it also includes a touch display screen, which is deployed on the front door of the cabinet and connected to the edge computing device.

[0010] Preferably, the moving detection unit and the lifting drive unit are located in the original reserved gap between the chassis inside the cabinet to be tested and the cabinet door.

[0011] This invention also proposes a method for detecting the internal status of a server rack based on a limited field-of-view image, comprising the following steps: The control module of the edge computing device controls the operation of the second motor, which drives the first motor fixing component and the terminal fixing component to rise or fall a distance d in the vertical direction along the two sides of the cabinet's original U-position columns. The distance d is determined by the vertical field of view of the image acquisition device. The first motor drives a horizontal transmission mechanism via its output shaft, and the horizontal transmission mechanism drives the image acquisition device, causing the image acquisition device to move horizontally. The image acquisition device is repeatedly moved in the vertical or horizontal direction until the image acquisition of the entire cabinet is completed. The acquired image information is then transmitted to the edge computing device, which is used to perform state detection inside the cabinet to obtain the state detection result of the cabinet.

[0012] Preferably, it further includes: using edge computing devices to acquire temperature and humidity information inside the cabinet collected by temperature and humidity sensors; Based on the image information and the temperature and humidity information, the pre-trained state detection model deployed in the edge computing device (3) is used to obtain the state detection results inside the cabinet and transmit the detection results to the cloud platform in real time.

[0013] Preferably, the expression that satisfies the relationship between the distance d and the number of motor control pulses P of the second motor is:

[0014]

[0015] Where PPR represents the number of pulses required per revolution of the second motor. The step angle of the second motor is represented by N, the microstepping of the driver is represented by R, the effective winding radius of the take-up reel is represented by i, and the belt drive ratio is represented by i.

[0016] The present invention also proposes a detection device, including the above-mentioned cabinet internal state detection system based on limited field of view images, wherein the detection device utilizes the detection system to implement the steps of the above-mentioned cabinet internal state detection method based on limited field of view images.

[0017] The present invention also proposes a cabinet, including a cabinet body and a cabinet internal status detection system based on a limited field of view image disposed within the cabinet body.

[0018] Compared with the prior art, the beneficial effects of the technical solution of the present invention are: This invention proposes a cabinet internal status detection system, method, and device based on limited field-of-view images, as well as a cabinet itself. The system includes a mobile detection unit, a lifting drive unit, and an edge computing device. Through coordinated control of the mobile detection unit and the lifting drive unit, the image acquisition device within the mobile detection unit moves in two dimensions, capturing image information that comprehensively covers all areas within the cabinet, effectively solving the problem of blind spots inherent in fixed cameras. The edge computing device integrates image acquisition and multi-view image stitching functions to identify and monitor the cabinet's internal status in real time. This invention, without modifying the existing cabinet structure, completely captures the entire internal space, improving the completeness and accuracy of cabinet status perception, and further enhancing the accuracy of cabinet internal status detection. Attached Figure Description

[0019] Figure 1 This diagram illustrates the structure of a cabinet internal status detection system based on a limited field-of-view image, as proposed in an embodiment of the present invention. Figure 2 This diagram illustrates the structure of the edge computing device proposed in an embodiment of the present invention. Figure 3 This diagram illustrates the position of the touchscreen as proposed in an embodiment of the present invention. The reference numerals in the attached figures are explained as follows: 1. Motion detection unit; 2. Lifting drive unit; 3. Edge computing device; 4. Touch screen; 101. First motor; 102. Image acquisition device; 103. Horizontal transmission mechanism; 104. Horizontal guide shaft; 105. First motor fixing component; 106. Terminal fixing component; 201. Second motor; 202. Second transmission mechanism; 203. Rotating shaft; 204. Take-up reel. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0021] Example 1 This embodiment proposes a cabinet interior status detection system based on limited field-of-view images. (See also...) Figure 1 ,include: The mobile detection unit 1 includes a first motor 101, an image acquisition device 102, a horizontal transmission mechanism 103, and a horizontal guide shaft 104. The output shaft of the first motor 101 is connected to one end of the horizontal transmission mechanism 103, and the other end of the horizontal transmission mechanism 103 is connected to the image acquisition device 102. The image acquisition device 102 is slidably mounted on the horizontal guide shaft 104. The two ends of the horizontal guide shaft 104 are respectively provided with a first motor fixing component 105 and a terminal fixing component 106. The first motor 101 is fixedly installed on the first motor fixing component 105. The first motor fixing component 105 and the terminal fixing component 106 are slidably engaged with the original U-position columns extending vertically on both sides of the cabinet. The original U-position columns serve as vertical guide rails for the lifting and lowering of the mobile detection unit 1. The lifting drive unit 2 includes a second motor 201, a second transmission mechanism 202, a rotating shaft 203, and a take-up reel 204. The second motor 201 is mounted on the cabinet door, and the output shaft of the second motor 201 is connected to the rotating shaft 203 via the second transmission mechanism 202. The rotating shaft 203 is mounted on the top of the cabinet in a horizontal direction. Two take-up reels 204 are fixedly installed at axial intervals on the shaft of the rotating shaft 203, and each take-up reel 204 is wound with a steel wire rope. One end of the two steel wire ropes is respectively connected to the first motor fixing member 105 and the terminal fixing member 106. The rotating shaft 203 drives the take-up reels 204 to take up and release the steel wire rope, thereby driving the first motor fixing member 105 and the terminal fixing member 106 to move up and down in the vertical direction along the two sides of the cabinet's original U-shaped columns. The edge computing device 3 includes a control module, a communication module, and a processing module. The control module controls the operation of the first motor 101 and the second motor 201. The communication module transmits the image information of the cabinet interior collected by the image acquisition device 102 to the processing module. The processing module is equipped with a pre-trained state detection model. Based on the image information, the state detection model is used to obtain the state detection results of the cabinet interior. The image information includes: the location, dimensions, type, shape, color, and status indicator color of the equipment inside the cabinet.

[0022] In this embodiment, in the motion detection unit 1, the horizontal guide shaft 104 provides horizontal guidance constraint for the image acquisition device 102 and increases the stability of the motion detection unit 1, ensuring that the image acquisition device 102 maintains a stable posture and clear imaging during inspection. The horizontal guide shaft 104 is made of stainless steel. The image acquisition device 102 performs multiple image acquisitions on the chassis within the cabinet, and combined with image stitching and deep learning technology, automatically identifies and monitors parameters such as chassis position, dimensions, and status light colors in real time. Components such as the first motor fixing component 105 and the terminal fixing component 106 are made of polyester material, effectively reducing the overall weight of the motion detection unit 1, thus facilitating precise control and stable operation of the second motor 201. The terminal fixing component 106 is used to restrict the lateral movement of the horizontal guide shaft 104.

[0023] In the lifting drive unit 2, the rotating shaft 203 is stably installed via a P-type bearing housing, ensuring good coaxiality and rotational stability during operation. Simultaneously, the second motor 201 is fixed to the uniform cabinet door surface via a dedicated motor bracket. The output shaft of the second motor 201 is connected to the rotating shaft 203 via a second transmission mechanism 202, driving the rotating shaft 203 to rotate. The symmetrically mounted take-up reels 204 at both ends of the rotating shaft 203 simultaneously wind up and unwind the wire rope. The other end of the wire rope is connected to the motion detection unit 1, thus converting the rotational motion of the rotating shaft 203 into a smooth vertical lifting motion of the motion detection unit 1. The rotating shaft 203 is made of stainless steel, and the take-up reels 204 are made of insulating material.

[0024] Edge computing device 3 is the core processing unit for cabinet internal status detection. It employs an ARM-based architecture with multi-core heterogeneous computing capabilities, enabling it to efficiently support real-time edge-side sensing, control, and interaction tasks using a high-performance embedded processor. Edge computing device 3 achieves comprehensive perception of the cabinet's internal operating environment through interfaces such as GPIO, I²C, UART, and RS485. Edge computing device 3 integrates complete motor drivers to drive the motors in motion detection unit 1 and lifting drive unit 2, completing automatic scanning of the cabinet's internal space. Edge computing device 3, through its embedded processor, storage unit, and communication module, deploys a lightweight chassis size, type, and status indicator light detection model. This model can efficiently infer and recognize chassis image data acquired in real-time by a local camera.

[0025] See Figure 2The edge computing device integrates multiple interfaces and network communication modules, responsible for data processing, device control, and cloud interaction. Specifically, the temperature and humidity sensor communicates with the edge computing device via an IIC interface to collect environmental temperature and humidity data; the image acquisition device connects to the edge computing device via a USB interface to collect image data. The first and second motors both communicate with the edge device via RS485 interfaces. The edge computing device has a built-in communication module that enables bidirectional data interaction with the cloud platform, allowing it to upload collected device and environmental data and receive control commands from the cloud. An HDMI interface connects to an external monitor for outputting the display interface and facilitating human-computer interaction; the RS485 interface connects to the electronic lock to control the cabinet door's opening and closing. The power supply module provides power to the entire edge computing device.

[0026] In an optional embodiment, it further includes: a plurality of temperature and humidity sensors, which are respectively deployed in different temperature zones within the cabinet, and collect temperature and humidity information inside the cabinet in their respective corresponding temperature zones. The temperature and humidity information inside the cabinet and the image information are jointly input into the state detection model to obtain the state detection result inside the cabinet.

[0027] In this embodiment, temperature and humidity sensors deployed in different temperature zones are connected to the input of edge computing device 3 via an IIC bus to achieve comprehensive monitoring of temperature and humidity inside the cabinet. First, edge computing device 3 sends initialization commands to each temperature and humidity sensor to complete the configuration and startup of the sensors. Then, according to a set time interval, the edge device periodically sends measurement commands to start the temperature and humidity sensors to synchronously collect temperature and humidity data. Finally, the temperature and humidity sensors transmit the collected temperature and humidity data back to the edge device via the IIC bus for subsequent processing and display.

[0028] In an optional embodiment, it further includes: a cloud platform, which is communicatively connected to the edge computing device 3, and receives and analyzes the cabinet internal status detection results transmitted by the edge computing device 3.

[0029] In this embodiment, the cloud platform supports multiple application layer protocols such as MQTT, RTMP, and IEC61850 to complete data interaction with the edge computing device 3. Regarding the communication architecture, the network layer between the cloud platform and the edge computing device 3 adopts the standard TCP / IP protocol stack as the underlying transmission mechanism to ensure data transmission reliability. To further enhance the system's protocol adaptability, a multi-threaded communication management module is deployed inside the edge device. This module handles data interaction tasks for different application layer protocols through independent threads. For example, for lightweight message transmission, the system supports the MQTT (Message Queuing Telemetry Transport) protocol; for audio and video stream data, it is compatible with the RTMP (Real-Time Messaging Protocol). Through the multi-threaded concurrency mechanism, each protocol channel does not interfere with each other, and different types of data streams can be received, parsed, and forwarded, thereby achieving flexible support for diverse business scenarios. The cloud platform provides a web-based graphical user interface (GUI), which uses 2D visualization technology to construct a digital twin model of the server rack. After logging into the cloud platform, users can intuitively view the current status information of the internal environment of the server rack in the browser. At the same time, the interface accurately presents the internal layout of the server rack in a scaled manner, clearly marking the type, physical size and current working status of the chassis deployed in each installation location. The status information is dynamically updated, making it easy for maintenance personnel to quickly identify abnormal chassis.

[0030] In an alternative embodiment, it further includes: a touch display screen 4, see [link to previous embodiment]. Figure 3 The touch display 4 is deployed on the front door of the cabinet and connected to the edge computing device 3.

[0031] In this embodiment, the touch display screen is connected to the edge computing device 3 via HDMI and USB cables. It embeds a human-machine interface developed based on the QT framework, primarily used to display the internal environment of the cabinet (temperature, humidity, door lock status, etc.), the type of chassis installed inside, and the current operating status of each chassis. Furthermore, users can configure system parameters through touch operation. For example, they can set temperature and humidity alarm thresholds and motor control modes on the interface. The touch display screen is preferably a capacitive multi-touch screen with a resolution of at least 1920*1080, supporting industrial-grade wide-temperature operating environments. The touch display screen is deployed in an embedded mounting slot in the front door of the cabinet for easy on-site viewing and operation by maintenance personnel.

[0032] In an optional embodiment, the moving detection unit 1 and the lifting drive unit 2 are disposed in the original reserved gap between the chassis inside the cabinet to be tested and the cabinet door.

[0033] Example 2 This embodiment proposes a method for detecting the internal status of a server rack based on a limited field-of-view image, including the following steps: S1: The control module of the edge computing device 3 controls the second motor 201 to operate, driving the first motor fixing part 105 and the terminal fixing part 106 to rise or fall a distance d in the vertical direction along the two sides of the cabinet's vertical original U-position columns. The distance d is determined by the vertical field of view of the image acquisition device 102. S2: The first motor 101 drives the horizontal transmission mechanism 103 through the output shaft, and the horizontal transmission mechanism 103 drives the image acquisition device 102, so that the image acquisition device 102 moves in the horizontal direction; S3: Repeatedly control the image acquisition device 102 to move repeatedly in the vertical or horizontal direction until the image acquisition of the entire cabinet is completed, and transmit the acquired image information to the edge computing device 3. Use the edge computing device 3 to perform state detection inside the cabinet and obtain the state detection result inside the cabinet.

[0034] It also includes: using the edge computing device 3 to acquire temperature and humidity information inside the cabinet collected by the temperature and humidity sensor; based on the image information and the temperature and humidity information, using the pre-trained state detection model deployed in the edge computing device 3 to acquire the state detection results inside the cabinet, and transmitting the detection results to the cloud platform in real time.

[0035] The expression that satisfies the relationship between the distance d and the number of motor control pulses P of the second motor 201 is:

[0036]

[0037] Wherein, PPR represents the number of pulses required per revolution of the second motor 201. The step angle of the second motor 201 is represented by N, the microstepping of the driver is represented by R, the effective winding radius of the take-up reel is represented by i, and the belt drive ratio is represented by i.

[0038] In this embodiment, the expression for the number of pulses PPR required per revolution of the motor is: The expression for the displacement of the moving detection unit 1 driven by each rotation of the shaft, i.e., the length L of the wire rope wound / unwound by the take-up reel 204, is: The expression for the number of pulses M per revolution of the shaft is: The quantitative relationship between the lifting distance d of the moving detection unit 1 and the number of motor control pulses P can be calculated using the above formula.

[0039] Example 3 This embodiment proposes a cabinet internal status detection device based on limited field of view images, including a cabinet internal status detection system based on limited field of view images proposed in Embodiment 1, which is used to implement a cabinet internal status detection method based on limited field of view images proposed in Embodiment 2.

[0040] This embodiment also proposes a cabinet, including a cabinet body and a cabinet internal status detection system based on a limited field-of-view image proposed in Embodiment 1, which is disposed inside the cabinet body.

[0041] The embodiments described are merely examples to clearly illustrate the present invention and are not intended to limit the implementation of the invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively describe all possible implementations. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A cabinet interior status detection system based on limited field-of-view images, characterized in that, include: The mobile detection unit (1) includes a first motor (101), an image acquisition device (102), a horizontal transmission mechanism (103), and a horizontal guide shaft (104). The output shaft of the first motor (101) is connected to one end of the horizontal transmission mechanism (103), and the other end of the horizontal transmission mechanism (103) is connected to the image acquisition device (102). The image acquisition device (102) is slidably mounted on the horizontal guide shaft (104). The two ends of the horizontal guide shaft (104) are respectively provided with a first motor fixing component (105) and a terminal fixing component (106). The first motor (101) is fixedly installed on the first motor fixing component (105). The first motor fixing component (105) and the terminal fixing component (106) are respectively slidably engaged with the original U-position columns extending vertically on both sides of the cabinet. The original U-position columns serve as vertical guide rails for the lifting and lowering of the mobile detection unit (1). The lifting drive unit (2) includes a second motor (201), a second transmission mechanism (202), a rotating shaft (203), and a take-up reel (204). The second motor (201) is mounted on the cabinet door. The output shaft of the second motor (201) is connected to the rotating shaft (203) via the second transmission mechanism (202). The rotating shaft (203) is mounted on the top of the cabinet in a horizontal direction. Two take-up reels (204) are fixedly installed on the shaft of the rotating shaft (203) at axial intervals. Each take-up reel (204) is wound with a steel wire rope. One end of the two steel wire ropes is connected to the first motor fixing member (105) and the terminal fixing member (106) respectively. The rotating shaft (203) drives the take-up reel (204) to take up and release the steel wire rope, thereby driving the first motor fixing member (105) and the terminal fixing member (106) to move up and down in the vertical direction along the two sides of the cabinet's original U-shaped columns. The edge computing device (3) includes a control module, a communication module, and a processing module. The control module is used to control the operation of the first motor (101) and the second motor (201). The communication module is used to transmit the image information inside the cabinet collected by the image acquisition device (102) to the processing module. The processing module is equipped with a pre-trained state detection model. Based on the image information, the state detection model is used to obtain the state detection results inside the cabinet. The image information includes: the location, dimensions, type, shape, color, and status indicator color of the equipment inside the cabinet.

2. The cabinet interior status detection system based on limited field-of-view images according to claim 1, characterized in that, It also includes: several temperature and humidity sensors, which are deployed in different temperature zones within the cabinet to collect internal temperature and humidity information in their respective temperature zones. The internal temperature and humidity information and the image information are input together into the state detection model to obtain the internal state detection result of the cabinet.

3. The cabinet interior status detection system based on limited field-of-view images according to claim 1, characterized in that, Also includes: The cloud platform is connected to the edge computing device (3) and receives and analyzes the cabinet internal status detection results transmitted by the edge computing device (3).

4. The cabinet internal status detection system based on limited field-of-view images according to claim 1, characterized in that, Also includes: A touch display screen (4) is deployed at the front door of the cabinet and connected to the edge computing device (3).

5. The cabinet internal status detection system based on limited field-of-view images according to claim 1, characterized in that, The mobile detection unit (1) and the lifting drive unit (2) are located in the original reserved gap between the chassis inside the cabinet to be tested and the cabinet door.

6. A method for detecting the internal status of a server rack based on a limited field-of-view image, characterized in that, Includes the following steps: The control module of the edge computing device (3) controls the operation of the second motor (201), which drives the first motor fixing part (105) and the terminal fixing part (106) to rise or fall a distance d in the vertical direction along the two sides of the cabinet's vertical original U-position column. The distance d is determined by the vertical field of view of the image acquisition device (102). The first motor (101) drives the horizontal transmission mechanism (103) through the output shaft, and the horizontal transmission mechanism (103) drives the image acquisition device (102) to move in the horizontal direction. The image acquisition device (102) is controlled to move repeatedly in the vertical or horizontal direction until the image acquisition of the entire cabinet is completed, and the acquired image information is transmitted to the edge computing device (3). The edge computing device (3) is used to perform state detection inside the cabinet and obtain the state detection result inside the cabinet.

7. The method for detecting the internal status of a server rack based on a limited field-of-view image according to claim 6, characterized in that, It also includes: using edge computing devices (3) to obtain temperature and humidity information inside the cabinet collected by temperature and humidity sensors; Based on the image information and the temperature and humidity information, the pre-trained state detection model deployed in the edge computing device (3) is used to obtain the state detection results inside the cabinet and transmit the detection results to the cloud platform in real time.

8. The method for detecting the internal status of a server rack based on a limited field-of-view image according to claim 6, characterized in that, The expression that satisfies the relationship between the distance d and the number of motor control pulses P of the second motor (201) is: Wherein, PPR represents the number of pulses required per revolution of the second motor (201). The step angle of the second motor (201) is represented by N, the microstepping of the driver is represented by R, the effective winding radius of the take-up reel is represented by i, and the belt drive ratio is represented by i.

9. A detection device, characterized in that, The detection device includes the detection system described in claims 1 to 5, and uses the detection system to implement the steps of the cabinet internal status detection method based on limited field-of-view images as described in claim 6.

10. A server rack, characterized in that, It includes the cabinet body and the detection system as described in claims 1 to 5, which is disposed within the cabinet body.

Images