Machine room ar positioning correction system and method and medium

Abstract

The application discloses a kind of machine room AR positioning correction systems, the rack feature library establishment module establishes the rack feature library of information machine room;The number identification module of rack identifies the number of rack for positioning calculation;The feature library matching module of rack obtains the position of each rack for positioning calculation in machine room map and corresponding rack contour size;AR glasses positioning module determines the position of AR glasses in information machine room;Positioning correction module corrects SLAM positioning in information machine room according to the position of AR glasses in information machine room.The present application utilizes SLAM technology to combine target detection, feature library matching and positioning technology, according to the position of AR glasses first visual angle image back, SLAM positioning is corrected according to back position, and the positioning problem in similar environment of feature point is solved.

Description

Technical Field

[0001] This invention relates to the field of information room positioning correction technology, specifically to a positioning correction system, method and medium based on AR (Augmented Reality) in a computer room. Background Technology

[0002] SLAM (Simultaneous Localization and Mapping) technology is used to construct environmental information and estimate one's own position while moving, using designated sensors when environmental information is unknown. When a camera is used as a sensor, it is called "visual SLAM".

[0003] Object detection, as one of the most fundamental and challenging problems in computer vision, has received extensive research and exploration in recent years. As an important task in computer vision, object detection typically aims to provide the precise location of a certain type of visual object (such as a human, animal, or car) in a digital image. Object detection is an image segmentation method based on the geometric and statistical features of objects, combining object segmentation and recognition. Its accuracy and real-time performance are crucial capabilities of the entire system. Especially in complex scenes where multiple objects need to be processed in real time, automatic object extraction and recognition become particularly important. From an application perspective, object detection can be divided into two research topics: object detection in general scenarios and object detection for specific categories. The difference is that the former is similar to simulating human vision and cognition, primarily aiming to explore methods for detecting different categories of objects within a unified framework; while the latter refers to detection in specific application scenarios, such as face detection, pedestrian detection, and vehicle detection. In recent years, the rapid development of deep learning technology has brought significant breakthroughs to object detection. Object detection is now widely used in many real-world scenarios, such as autonomous driving, robot vision, and video surveillance.

[0004] With the development of positioning technology, GPS and base station positioning technologies have achieved accurate positioning in most outdoor scenarios. However, in some outdoor scenarios or indoors, due to the presence of various obstacles, GNSS signals attenuate rapidly or even disappear, making accurate positioning impossible in these areas. Meanwhile, indoor scenarios such as data centers have relatively uniform environmental characteristics, thus placing higher demands on positioning. Currently, technologies used for indoor positioning mainly include Wi-Fi positioning, Bluetooth positioning, infrared positioning, ultrasonic positioning, geomagnetic positioning, RFID positioning, and ultra-wideband positioning, among others. However, these technologies each have different shortcomings in terms of positioning accuracy, coverage, reliability, power consumption, and cost. For example, Wi-Fi positioning hotspots are greatly affected by the surrounding environment, resulting in low positioning accuracy and high maintenance costs; RFID positioning has a small coverage area and lacks communication capabilities; ultra-wideband positioning technology is costly and has complex network deployment. Summary of the Invention

[0005] The purpose of this invention is to provide an AR positioning correction system, method, and medium for computer rooms. This invention utilizes SLAM technology, combining target detection, feature library matching, and positioning techniques, to infer the position based on the first-view image of the AR glasses. SLAM positioning is then corrected based on the inferred position, solving the positioning problem in environments with similar feature points.

[0006] To achieve this objective, the computer room AR positioning correction system designed in this invention is characterized by comprising a cabinet feature library establishment module, a cabinet number identification module, a cabinet feature library matching module, an AR glasses positioning module, and a positioning correction module.

[0007] The cabinet feature library establishment module is used to establish a cabinet feature library for the information data center;

[0008] The cabinet number recognition module is used in an information data center scenario. When there is a complete cabinet in the first field of view of the AR glasses, it uses target detection technology to perform target detection on each cabinet in the first field of view of the AR glasses, and performs text recognition on each complete cabinet detected by the target. The recognized cabinet number is the number of the cabinet used for positioning calculation.

[0009] The cabinet feature library matching module is used to match each positioning computing cabinet in the cabinet feature library according to the number of each cabinet, and associate them to obtain the location of each positioning computing cabinet in the data center map and the corresponding cabinet outline size.

[0010] The AR glasses positioning module is used to determine the position of the AR glasses relative to each positioning computing cabinet in the first-view image based on the position of each positioning computing cabinet in the data center map and the corresponding cabinet outline size, and to determine the position of the AR glasses in the data center based on the coordinates of each positioning computing cabinet and the position of the AR glasses relative to each positioning computing cabinet in the first-view image.

[0011] The positioning correction module is used to correct the SLAM positioning within the information room based on the location of the AR glasses in the information room.

[0012] The beneficial effects of this invention are:

[0013] This invention creates a feature library for server racks and equipment in a data center, integrating feature matching, target detection, and text recognition technologies. First, target detection is performed on the equipment within the data center. Then, text recognition is performed on each identified target. Finally, the text recognition results are matched against the feature library to obtain the association information of the racks and equipment. Based on this association information, the location of the AR glasses is inferred. SLAM positioning is corrected based on the inferred results to avoid positioning errors caused by similar environmental features. SLAM can easily locate to another similar location in similar environments. If the SLAM positioning and the inferred position from the AR glasses' first-view perspective are inconsistent, the inferred position takes precedence. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the structure of the present invention;

[0015] Figure 2 This is a flowchart of the method of the present invention;

[0016] Figure 3 This is a three-point perspective diagram of P3P (Perspective-Three-Point) for a complete server rack;

[0017] Figure 4 A three-point perspective schematic diagram of P3P for a complete server rack;

[0018] Figure 5 This is a three-point perspective diagram of P3P for equipment inside the cabinet;

[0019] Figure 6 A three-point perspective schematic diagram of P3P for equipment inside a server rack; Detailed Implementation

[0020] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0021] like Figure 1The AR positioning correction system for the computer room shown is characterized by comprising a cabinet feature library establishment module, a cabinet number recognition module, a cabinet feature library matching module, an AR glasses positioning module, and a positioning correction module.

[0022] The cabinet feature library establishment module is used to establish a cabinet feature library for the information data center. The cabinet feature library includes cabinet number (Nx), cabinet outline dimensions (S1, S2, S3), and outline dimensions mainly include the length, width, and diagonal length of the cabinet panel, the position of the cabinet centroid on the data center map (X1, Y1), the equipment installed in the cabinet (type, quantity), the centroid position of each equipment panel in the cabinet panel (X2, Y2), and the outline dimensions of each equipment (S1, S2, S3). The outline dimensions mainly include the length, width, and diagonal length of the equipment panel.

[0023] The cabinet number recognition module is used in an information data center scenario. When there is a complete cabinet in the first field of view of the AR glasses, it uses image recognition-based target detection technology to detect each cabinet in the first field of view of the AR glasses, performs text recognition on each complete cabinet detected by the target, and the recognized cabinet number is the number of the cabinet used for positioning calculation.

[0024] The cabinet feature library matching module is used to match the cabinet feature library according to the number of each positioning computing cabinet (the cabinet number is obtained by text recognition, and the cabinet number field is matched in the library. If the recognized cabinet number is consistent with a certain cabinet in the library, the information of that cabinet in the library is associated). The association yields the location of each positioning computing cabinet in the data center map and the corresponding cabinet outline size.

[0025] The AR glasses positioning module is used to determine the position of the AR glasses relative to each positioning calculation cabinet in the first-view image based on the position of each positioning calculation cabinet in the data center map and the corresponding cabinet outline dimensions. It also determines the position of the AR glasses in the data center based on the coordinates of each positioning calculation cabinet and the position of the AR glasses relative to each positioning calculation cabinet in the first-view image (if multiple cabinets are detected, their positions are reverse-calculated based on the first-view image of each cabinet, and then the reverse-calculated positions are added together and averaged. This averages out the error and prevents large individual calculation errors. The reverse positioning process using the device is similar below). The target detection technology detects all matching targets in the field of view and identifies which complete cabinets are present in the first-view image. It then performs text recognition on the labels of each complete cabinet, obtaining the cabinet number. Based on the complete cabinet number field, it matches the cabinet size, coordinates, and other information to obtain the complete cabinet's dimensions. Finally, it performs reverse positioning calculations based on this information.

[0026] The positioning correction module is used to correct the position of the AR glasses in the information room based on their location in the information room. If the AR glasses are mispositioned due to similar environmental features, the positioning will be corrected to the correct position.

[0027] The above technical solution also includes a device target detection module and a device feature library matching module. The device target detection module is used in an information data center scenario, where there is no complete server rack within the first field of view of the AR glasses, to perform target detection technology on the devices in each server rack within the field of view, determining the type and quantity of devices in each rack, as well as the centroid position of each device panel on the rack panel. During inspection, personnel gradually approach the server rack. Once close, the rack label becomes invisible in the first image of the AR glasses, making text recognition impossible and thus preventing the acquisition of the rack number. Therefore, when the rack number cannot be obtained through text recognition, the rack can be identified and its dimensions and other parameters can be obtained by matching the type, quantity, and location of the devices in the rack against a feature library.

[0028] The feature library matching module is used to match the category and quantity of equipment in the cabinet determined by target detection, as well as the centroid position of each equipment panel in the cabinet panel, with the cabinet feature library to obtain the cabinet number of each positioning computing cabinet in the first field of view of the AR glasses (the library contains the quantity, category, and location fields of each cabinet; if the quantity, identification, and location of equipment in a cabinet in the first view are consistent with a cabinet in the feature library, the cabinet number field is assigned to that cabinet). Based on the cabinet number of each positioning computing cabinet, the feature library is matched to obtain the coordinates of each positioning computing cabinet (the position of the positioning computing cabinet in the data center map) and the corresponding outline dimensions of the equipment in the cabinet.

[0029] The AR glasses positioning module is used to determine the position of the AR glasses relative to each positioning computing cabinet in the first-view image based on the coordinates of each positioning computing cabinet and the outline dimensions of the corresponding equipment inside the cabinet. It also determines the position of the AR glasses in the information room based on the coordinates of each positioning computing cabinet and the position of the AR glasses relative to each positioning computing cabinet in the first-view image.

[0030] In the above technical solution, the AR glasses positioning module determines the position of the AR glasses relative to the positioning computing cabinet in the first-view image based on the coordinates of the positioning computing cabinet and the corresponding cabinet outline dimensions. The specific method for determining the position of the AR glasses in the information room based on the coordinates of the positioning computing cabinet and the position of the AR glasses relative to the positioning computing cabinet in the first-view image is as follows:

[0031] Since ∠aOb=∠AOB; ∠aOc=∠AOC; ∠cOb=∠COB, using the Law of Cosines, we have, as follows: Figure 3 As shown:

[0032] OA 2 +OB 2 -2OA*OB*cosθAB=AB 2

[0033] OB 2 +OC 2 -2OB*OC*cosθBC=BC 2

[0034] OA 2 +OC 2 -2OA*OC*cosθAC=AC 2

[0035] Therefore, as Figure 4 As shown;

[0036]

[0037]

[0038]

[0039] Wherein, point O represents the position of the AR glasses' front-facing camera; points A, B, and C represent the three vertices of the real-world positioning calculation cabinet outline; points a, b, and c represent the three vertices of the positioning calculation cabinet outline in the first-person view image; OA, OB, and OC represent the distances between the three vertices of the real-world positioning calculation cabinet outline and the AR glasses' front-facing camera; AB, BC, and AC represent the side lengths between adjacent fixed points of the real-world calculation cabinet outline; cosθAB represents the cosine function of ∠AOB, cosθAC represents the cosine function of ∠AOC, cosθBC represents the cosine function of ∠BOC, S1 = AB, S2 = AC, S3 = BC, L A =OA,L B =OB,L C =OC;

[0040] The lengths of OA, OB, and OC are obtained. Combined with the coordinates of the server rack and the distance of the AR glasses from the rack, the location of the AR glasses in the data center is determined. After object detection and text recognition, and then feature database matching and association, the coordinates of the server rack on the data center map are obtained. The distances of the AR glasses from each vertex of the server rack outline are also known. Through calculation, it can be found that there is a point whose distance from each vertex of the server rack outline is consistent with OA, OB, and OC. This point is the location of the AR glasses.

[0041] In the above technical solution, the AR glasses positioning module determines the position of the AR glasses relative to the positioning computing cabinet in the first-view image based on the coordinates of a positioning computing cabinet and the outline dimensions of the corresponding equipment inside the cabinet. The specific method for determining the position of the AR glasses in the information room based on the coordinates of the positioning computing cabinet and the position of the AR glasses relative to the positioning computing cabinet in the first-view image is as follows:

[0042] Since ∠a1Ob1=∠A1OB1; ∠a1Oc1=∠A1OC1; ∠c1Ob1=∠C1OB1, using the Law of Cosines, we have, as follows: Figure 5 As shown:

[0043] OA1 2 +OB1 2 -2OA1*OB1*cosθA1B1=A1B1 2

[0044] OB1 2 +OC1 2 -2OB1*OC1*cosθB1C1=B1C1 2

[0045] OA1 2 +OC1 2 -2OA1*OC1*cosθA1C1=A1C1 2

[0046] Therefore, as Figure 6 As shown;

[0047] S11 2 =A1B1 2 =L1 A 2 +L1 B 2 -2L1 A L1 B cosθA1B1

[0048] S12 2 =A1C1 2 =L1 A 2 +L1 C 2 -2L1 A L1 C cosθA1C1

[0049] S13 2 =B1C1 2 =L1 B 2 +L1 C 2 -2L1B L1 C cosθB1C1

[0050] Wherein, point O represents the position of the front-facing camera of the AR glasses; points A1, B1, and C1 represent the three vertices of the outline of the equipment inside the positioning and computing cabinet in the real world; points a1, b1, and c1 represent the three vertices of the outline of the equipment inside the positioning and computing cabinet in the first-view image; OA1, OB1, and OC1 represent the distances between the three vertices of the outline of the equipment inside the positioning and computing cabinet in the real world and the front-facing camera of the AR glasses; A1B1, B1C1, and A1C1 represent the side lengths between adjacent fixed points of the outline of the equipment inside the computing cabinet in the real world; cosθA1B1 represents the cosine function of ∠A1OB1, cosθA1C1 represents the cosine function of ∠A1OC1, cosθB1C1 represents the cosine function of ∠B1OC1, S11 = A1B1, S12 = A1C1, S13 = B1C1, L1 A =OA1, L1 B =OB1,L1 C =OC1;

[0051] The lengths of OA1, OB1, and OC1 are obtained. Combined with the coordinates of the server rack and the position of the AR glasses relative to the equipment within the rack, the location of the AR glasses in the data center is determined. After object detection and text recognition, and then feature database matching and association, the coordinates of the server rack on the data center map are obtained. Given the distances of the AR glasses from the vertices of the equipment within the rack's outline, calculations show that there is a point whose distance from the vertices of the equipment within the rack's outline matches that of OA1, OB1, and OC1. This point is the location of the AR glasses.

[0052] In the above technical solution, the cabinet feature library establishment module is used to establish a cabinet feature library for the information room based on the construction drawings and equipment manuals.

[0053] In the above technical solution, the specific method by which the positioning correction module corrects the SLAM positioning in the information room based on the position of the AR glasses in the information room is as follows: if the AR glasses position located by SLAM is inconsistent with the reverse-engineered position of the AR glasses' first view, then the reverse-engineered position shall prevail.

[0054] A method for AR positioning correction in a computer room, such as Figure 2 As shown, it includes the following steps:

[0055] Step 1: Establish a cabinet feature library for the information data center. The cabinet feature library includes cabinet number (Nx), cabinet outline dimensions (S1, S2, S3), the location of the cabinet centroid on the data center map (X1, Y1), the equipment installed in the cabinet (type, quantity), the centroid location of each equipment panel in the cabinet panel (X2, Y2), and the outline dimensions of each equipment (S1, S2, S3).

[0056] Step 2: In the information data center scenario, when there is a complete server rack in the first field of view of the AR glasses, target detection technology based on image recognition is used to detect each server rack in the first field of view of the AR glasses. Text recognition is performed on each complete server rack detected by the target, and the identified server rack number is the number of the server rack used for positioning and computing.

[0057] The rack number of each positioning computing rack is matched in the rack feature library (the rack number is obtained by text recognition, and the rack number field is matched in the library. If the recognized rack number is consistent with a certain rack in the library, the information of that rack in the library is associated). The association yields the location of each positioning computing rack in the data center map and the corresponding rack outline dimensions.

[0058] The position of the AR glasses relative to each positioning computing cabinet in the first-view image is determined based on the position of each positioning computing cabinet in the data center map and the corresponding cabinet outline dimensions. The position of the AR glasses in the data center is determined based on the coordinates of each positioning computing cabinet and the position of the AR glasses relative to each positioning computing cabinet in the first-view image.

[0059] In the context of an information data center, when there is no complete server rack within the first field of view of the AR glasses, target detection technology is used to detect the devices in each server rack within the field of view, determine the type and quantity of devices in each server rack, and the centroid position of each device panel in the server rack panel.

[0060] The categories and quantities of equipment in the cabinets determined by target detection, as well as the centroid positions of each equipment panel in the cabinet, are matched with the cabinet feature library to obtain the cabinet number of each positioning and computing cabinet in the first field of view of the AR glasses. Based on each positioning and computing cabinet number, the cabinet feature library is matched to obtain the coordinates of each positioning and computing cabinet (the position of the positioning and computing cabinet in the data center map) and the corresponding outline dimensions of the equipment in the cabinet.

[0061] The position of the AR glasses relative to each positioning computing cabinet in the first-view image is determined based on the coordinates of each positioning computing cabinet and the outline dimensions of the corresponding equipment inside the cabinet. The position of the AR glasses in the information room is determined based on the coordinates of each positioning computing cabinet and the position of the AR glasses relative to each positioning computing cabinet in the first-view image.

[0062] Step 3: Correct the SLAM positioning in the information room based on the location of the AR glasses in the information room.

[0063] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method.

[0064] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A computer room AR positioning correction system, characterized in that, It includes a cabinet feature library establishment module, a cabinet number recognition module, a cabinet feature library matching module, an AR glasses positioning module, and a positioning correction module; The cabinet feature library establishment module is used to establish a cabinet feature library for the information data center; The cabinet number recognition module is used in an information data center scenario. When there is a complete cabinet in the first field of view of the AR glasses, it uses target detection technology to perform target detection on each cabinet in the first field of view of the AR glasses, and performs text recognition on each complete cabinet detected by the target. The recognized cabinet number is the number of the cabinet used for positioning calculation. The cabinet feature library matching module is used to match each positioning computing cabinet in the cabinet feature library according to the number of each cabinet, and associate them to obtain the location of each positioning computing cabinet in the data center map and the corresponding cabinet outline size. The AR glasses positioning module is used to determine the position of each positioning computing cabinet in the first-view image based on the position of each positioning computing cabinet in the data center map and the corresponding cabinet outline size. It also determines the position of the AR glasses in the data center based on the coordinates of each positioning computing cabinet and the position of the AR glasses in the first-view image. The target detection technology detects all matching targets in the field of view and detects complete cabinets in the first-view image. Then, it performs text recognition on the signs of each complete cabinet to obtain the number of each complete cabinet. Then, it matches the complete cabinet number field to obtain the size, coordinates and other information of each complete cabinet. Finally, it performs inverse positioning calculation based on this information. The positioning correction module is used to correct the SLAM positioning within the information room based on the location of the AR glasses in the information room; It also includes a device target detection module, which is used in an information data center scenario to detect the devices in each cabinet within the first field of view of the AR glasses using target detection technology when there is no complete cabinet, in order to determine the type and quantity of devices in each cabinet, as well as the centroid position of each device panel in the cabinet panel. It also includes a device feature library matching module, which is used to match the type and quantity of devices in the cabinet determined by target detection, as well as the centroid position of each device panel in the cabinet panel, with the cabinet feature library to obtain the cabinet number of each positioning computing cabinet in the first field of view of the AR glasses. Based on the cabinet number of each positioning computing cabinet, it matches in the cabinet feature library to obtain the coordinates of each positioning computing cabinet and the corresponding outline size of the devices in the cabinet. The AR glasses positioning module determines the position of the AR glasses relative to the first-view image of the positioning computing cabinet based on the coordinates of the cabinet and the corresponding cabinet outline dimensions. The specific method for determining the position of the AR glasses in the data center based on the coordinates of the positioning computing cabinet and the position of the AR glasses relative to the first-view image of the positioning computing cabinet is as follows: Since ∠aOb=∠AOB; ∠aOc=∠AOC; ∠cOb=∠COB, using the Law of Cosines, we have: OA 2 +OB 2 -2OA*OB*cosθAB=AB 2 OB 2 +OC 2 -2OB*OC*cosθBC=BC 2 OA 2 +OC 2 -2OA*OC*cosθAC=AC 2 therefore; Wherein, point O represents the position of the AR glasses' front-facing camera; points A, B, and C represent the three vertices of the real-world positioning calculation cabinet outline; points a, b, and c represent the three vertices of the positioning calculation cabinet outline in the first-person view image; OA, OB, and OC represent the distances between the three vertices of the real-world positioning calculation cabinet outline and the AR glasses' front-facing camera; AB, BC, and AC represent the side lengths between adjacent fixed points of the real-world calculation cabinet outline; cosθAB represents the cosine function of ∠AOB, cosθAC represents the cosine function of ∠AOC, cosθBC represents the cosine function of ∠BOC, S1 = AB, S2 = AC, S3 = BC, L A =OA,L B =OB,L C =OC; The lengths of OA, OB, and OC are obtained. Combined with the coordinates of the server rack and the distance of the AR glasses from the server rack, the location of the AR glasses in the information room is determined. After target detection and text recognition, and then feature database matching and association, the coordinates of the server rack in the server room map have been obtained. The distances of the AR glasses from each vertex of the server rack outline are also known. Through calculation, it can be found that there is a point whose distance from each vertex of the server rack outline is consistent with OA, OB, and OC. This point is the location of the AR glasses.

2. The computer room AR positioning correction system according to claim 1, characterized in that: The AR glasses positioning module is used to determine the position of the AR glasses relative to each positioning computing cabinet in the first-view image based on the coordinates of each positioning computing cabinet and the outline dimensions of the corresponding equipment inside the cabinet. It also determines the position of the AR glasses in the information room based on the coordinates of each positioning computing cabinet and the position of the AR glasses relative to each positioning computing cabinet in the first-view image.

3. The computer room AR positioning correction system according to claim 2, characterized in that: The AR glasses positioning module determines the position of the AR glasses relative to the first-view image of the positioning computing cabinet based on the coordinates of the cabinet and the outline dimensions of the equipment inside the cabinet. The specific method for determining the AR glasses' position within the data center based on the coordinates of the positioning computing cabinet and the position of the AR glasses relative to the first-view image is as follows: Since ∠a1Ob1=∠A1OB1; ∠a1Oc1=∠A1OC1; ∠c1Ob1=∠C1OB1, using the Law of Cosines, we have: OA1 2 +OB1 2 -2OA1*OB1*cosθA1B1=A1B1 2 OB1 2 +OC1 2 -2OB1*OC1*cosθB1C1=B1C1 2 OA1 2 +OC1 2 -2OA1*OC1*cosθA1C1=A1C1 2 therefore; S11 2 =A1B1 2 =L1 A 2 +L1 B 2 -2L1 A L1 B cosθA1B1 S12 2 =A1C1 2 =L1 A 2 +L1 C 2 -2L1 A L1 C cosθA1C1 S13 2 =B1C1 2 =L1 B 2 +L1 C 2 -2L1 B L1 C cosθB1C1 Wherein, point O represents the position of the front-facing camera of the AR glasses; points A1, B1, and C1 represent the three vertices of the outline of the equipment inside the positioning and computing cabinet in the real world; points a1, b1, and c1 represent the three vertices of the outline of the equipment inside the positioning and computing cabinet in the first-view image; OA1, OB1, and OC1 represent the distances between the three vertices of the outline of the equipment inside the positioning and computing cabinet in the real world and the front-facing camera of the AR glasses; A1B1, B1C1, and A1C1 represent the side lengths between adjacent fixed points of the outline of the equipment inside the computing cabinet in the real world; cosθA1B1 represents the cosine function of ∠A1OB1, cosθA1C1 represents the cosine function of ∠A1OC1, cosθB1C1 represents the cosine function of ∠B1OC1, S11 = A1B1, S12 = A1C1, S13 = B1C1, L1 A =OA1, L1 B =OB1,L1 C =OC1; The lengths of OA1, OB1, and OC1 are obtained. Combined with the coordinates of the server rack and the position of the AR glasses relative to the equipment inside the rack, the position of the AR glasses in the information room is determined. After target detection and text recognition, and then feature database matching and association, the coordinates of the server rack in the server room map have been obtained. The distances of the AR glasses from each vertex of the equipment within the rack outline are also known. Through calculation, it can be found that there is a point whose distance from each vertex of the equipment outline within the rack is consistent with OA1, OB1, and OC1. This point is the position of the AR glasses.

4. The computer room AR positioning correction system according to claim 1, characterized in that: The cabinet feature library creation module is used to create a cabinet feature library for the information room based on construction drawings and equipment manuals.

5. The computer room AR positioning correction system according to claim 1, characterized in that: The specific method by which the positioning correction module corrects the SLAM positioning in the information room based on the AR glasses' position in the information room is as follows: if the AR glasses' position determined by SLAM positioning is inconsistent with the AR glasses' position calculated from the first viewpoint, then the calculated position shall prevail.

6. A computer room AR positioning correction method based on the system of claim 1, characterized in that, It includes the following steps: Step 1: Establish a database of server rack characteristics for the data center; Step 2: In the information data center scenario, when there is a complete server rack in the first field of view of the AR glasses, target detection technology based on image recognition is used to detect each server rack in the first field of view of the AR glasses. Text recognition is performed on each complete server rack detected by the target, and the identified server rack number is the number of the server rack used for positioning and computing. Based on the number of each positioning computing cabinet, the cabinet feature database is matched to obtain the location of each positioning computing cabinet in the data center map and the corresponding cabinet outline dimensions. The position of the AR glasses relative to each positioning computing cabinet in the first-view image is determined based on the position of each positioning computing cabinet in the data center map and the corresponding cabinet outline dimensions. The position of the AR glasses in the data center is determined based on the coordinates of each positioning computing cabinet and the position of the AR glasses relative to each positioning computing cabinet in the first-view image. In the context of an information data center, when there is no complete server rack within the first field of view of the AR glasses, target detection technology is used to detect the devices in each server rack within the field of view, determine the type and quantity of devices in each server rack, and the centroid position of each device panel in the server rack panel. The categories and quantities of devices in the cabinet determined by target detection, as well as the centroid positions of each device panel in the cabinet, are matched with the cabinet feature library to obtain the cabinet number of each positioning computing cabinet in the first field of view of the AR glasses. Based on the cabinet number of each positioning computing cabinet, the cabinet feature library is matched to obtain the coordinates of each positioning computing cabinet and the corresponding outline dimensions of the devices in the cabinet. The position of the AR glasses relative to each positioning computing cabinet in the first-view image is determined based on the coordinates of each positioning computing cabinet and the outline dimensions of the corresponding equipment inside the cabinet. The position of the AR glasses in the information room is determined based on the coordinates of each positioning computing cabinet and the position of the AR glasses relative to each positioning computing cabinet in the first-view image. Step 3: Correct the SLAM positioning in the information room based on the location of the AR glasses in the information room.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in claim 6.

Images