Unmanned inspection method for equipment in property and equipment and facility industries

Abstract

The invention relates to an unmanned inspection method for equipment in property and equipment and facility industries, and belongs to the technical field of equipment management. According to the method, an unmanned aerial vehicle routing inspection route is formulated, and the unmanned aerial vehicle passes through each waypoint regularly according to the determined routing inspection frequency to obtain routing inspection data; the inspection data is sent to a user side through a network to generate an inspection report, and an alarm is given when the inspection data exceeds a preset range, so that real pure unmanned inspection, automatic inspection record generation and abnormal alarm functions can be realized. And the method disclosed by the invention is wide in application range and relatively low in application cost.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment management, and in particular to the technical field of equipment inspection methods, and specifically refers to an unmanned inspection method for equipment in the property and equipment facility industries. Background Art

[0002] In the existing property industry, such as power distribution rooms and data centers, technicians are required to conduct manual inspections and meter readings at least 3-4 times a day, and inspect the voltage, current, power factor, load, pressure, temperature, and water leakage displayed by the equipment. In this process, labor costs are the main cost component. Therefore, how to reduce manual intervention and realize more automatic inspections is the main demand of the property management industry.

[0003] There are some similar unmanned equipment and products on the market that can take photos and videos of relevant places, but they cannot achieve true unmanned operation. Because backstage personnel are required to view the photos and videos taken and to operate the drone on site, all on-site conditions need to be manually generated by post-processing to form relevant inspection records and status lists, and post-processing to judge the equipment status and generate alarms to report for maintenance.

[0004] In addition, the existing building automatic control systems are generally completed through later renovation, and the related sensors cannot necessarily guarantee 100% normal operation. Personnel inspections are still required, and the cost of investing in building intelligent control systems is not low. It only reduces the demand for some personnel. Most companies still use traditional personnel to conduct intermittent inspections in different time periods.

[0005] In summary, there is currently no equipment on the market that can achieve truly unmanned inspection, automatically generate inspection records, and remotely alarm for pre-emptive abnormal conditions of the equipment. Summary of the invention

[0006] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and to provide an unmanned inspection method for equipment in the property and equipment facility industries that can achieve true pure unmanned inspection and automatically generate inspection records and abnormal alarm functions.

[0007] In order to achieve the above-mentioned purpose, the unmanned inspection method for equipment in the property and equipment facilities industry of the present invention comprises:

[0008] Formulate a drone inspection route, draw a route through inspection waypoints on a map including the inspected device, the waypoints correspond to the locations of the inspected device, and determine the inspection frequency;

[0009] The drone passes through the waypoints in sequence according to the inspection route and obtains corresponding inspection data at the waypoints;

[0010] The drone sends the inspection data to the user end via the network;

[0011] The user terminal generates an inspection report based on the inspection data and issues an alarm when the inspection data exceeds a preset range.

[0012] In the unmanned inspection method for property and equipment facilities industry equipment, the waypoints include at least one environmental monitoring waypoint and at least one equipment monitoring waypoint, the route includes a drone flight route that passes through all environmental monitoring waypoints and equipment monitoring waypoints, and the inspection data includes environmental data and equipment data.

[0013] In the unmanned inspection method for property and equipment facilities industry equipment, the route also includes the drone's flight altitude, speed, and camera action information, and each waypoint also includes the drone's altitude, speed, hovering time, camera orientation, camera action, gimbal pitch angle, and data content corresponding to the photo taken corresponding to the waypoint.

[0014] In the unmanned inspection method for property and equipment facilities industry equipment, the corresponding inspection data is obtained at the waypoint as follows:

[0015] At the environmental monitoring waypoints, the drone uses infrared temperature measurement to detect and obtain temperature information; uses a combustible gas sensor to monitor whether there is combustible gas leakage information; and uses a camera to capture information about water leakage;

[0016] At the equipment monitoring waypoint, the drone uses a camera to capture pictures containing equipment instruments, and obtains instrument data by analyzing the pictures. The instruments include one or more of a temperature gauge, a pressure gauge, a voltmeter, an ammeter, a power factor meter, and an illumination meter.

[0017] In the unmanned inspection method for property and equipment facilities industry equipment, the acquisition of instrument data by analyzing pictures specifically includes:

[0018] Image data extraction, including shooting time, shooting location, camera model and shooting parameters;

[0019] Text recognition, using OCR technology to identify the text content in the picture, and generate instrument data based on the information corresponding to the data content of the photograph;

[0020] Image content analysis, using image recognition technology to identify objects, scenes or activities in pictures and generate content data;

[0021] Report generation organizes the extracted image data, instrument data, and content data into structured data according to logical relationships, and automatically generates reports including charts, tables, and text based on the structured data.

[0022] In the unmanned inspection method for property and equipment facilities industry equipment, the drone sends pictures to the server through the network, and the server implements the picture data extraction, text recognition, image content analysis and report generation processing.

[0023] In the unmanned inspection method for property and equipment facilities industry equipment, the server also stores and backs up data.

[0024] In the unmanned inspection method for property and equipment facilities industry equipment, the user terminal is connected to the server through the network, obtains the reports and data, generates and displays the inspection report, and issues an alarm when the inspection data exceeds a preset range.

[0025] In the unmanned inspection method for equipment in the property and equipment facilities industry, the server interacts with the AI ​​platform for data, and the AI ​​platform analyzes the environmental data and equipment data to generate energy-saving and emission-reduction setting suggestions for the inspected equipment.

[0026] The unmanned inspection method for equipment in the property and equipment facility industry adopts the invention. By formulating a drone inspection route, the drone is regularly passed through each waypoint according to the determined inspection frequency to obtain inspection data. The inspection data is then sent to the user end through the network to generate an inspection report, and an alarm is issued when the inspection data exceeds a preset range. This can achieve true unmanned inspection and automatic generation of inspection records and abnormal alarm functions, and has a wide range of applications and relatively low application costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of the unmanned inspection method for equipment in the property and equipment facilities industry of the present invention;

[0028] Figure 2 This is a schematic diagram of the inspection route of the power distribution room when the unmanned inspection method for equipment in the property and equipment facility industry of the present invention is applied.

[0029] Figure 3 It is a flow chart of the unmanned inspection method for equipment in the property and equipment facilities industry of the present invention in practical application. DETAILED DESCRIPTION

[0030] In order to more clearly understand the technical content of the present invention, the following embodiments are given in detail.

[0031] See also Figure 1 Shown is a flow chart of the unmanned inspection method for equipment in the property and equipment facility industry of the present invention.

[0032] In one embodiment, the unmanned inspection method for equipment in the property and equipment facilities industry includes:

[0033] Formulate a drone inspection route, draw a route through inspection waypoints on a map including the inspected device, the waypoints correspond to the locations of the inspected device, and determine the inspection frequency;

[0034] The drone passes through the waypoints in sequence according to the inspection route and obtains corresponding inspection data at the waypoints;

[0035] The drone sends the inspection data to the user end via the network;

[0036] The user terminal generates an inspection report based on the inspection data and issues an alarm when the inspection data exceeds a preset range.

[0037] In a preferred embodiment, the waypoints include at least one environmental monitoring waypoint and at least one equipment monitoring waypoint, the route includes a drone flight route passing through all environmental monitoring waypoints and equipment monitoring waypoints, and the inspection data includes environmental data and equipment data.

[0038] The route also includes the drone's flight altitude, speed, and camera action information, and each waypoint also includes information on the drone's altitude, speed, hovering time, camera orientation, camera action, gimbal pitch angle, and data content corresponding to the photos taken.

[0039] The obtaining of corresponding inspection data at the waypoint is specifically as follows:

[0040] At the environmental monitoring waypoints, the drone uses infrared temperature measurement to detect and obtain temperature information; uses a combustible gas sensor to monitor whether there is combustible gas leakage information; and uses a camera to capture information about water leakage;

[0041] At the equipment monitoring waypoint, the drone uses a camera to capture pictures containing equipment instruments, and obtains instrument data by analyzing the pictures. The instruments include one or more of a temperature gauge, a pressure gauge, a voltmeter, an ammeter, a power factor meter, and an illumination meter.

[0042] In a further preferred embodiment, the obtaining of instrument data by analyzing pictures specifically includes:

[0043] Image data extraction, including shooting time, shooting location, camera model and shooting parameters;

[0044] Text recognition, using OCR technology to identify the text content in the picture, and generate instrument data based on the information corresponding to the data content of the photograph;

[0045] Image content analysis, using image recognition technology to identify objects, scenes or activities in pictures and generate content data;

[0046] Report generation organizes the extracted image data, instrument data, and content data into structured data according to logical relationships, and automatically generates reports including charts, tables, and text based on the structured data.

[0047] The drone sends the image to the server via the network, and the server implements the image data extraction, text recognition, image content analysis and report generation processing.

[0048] In a more preferred embodiment, the server also stores and backs up data.

[0049] The user terminal is connected to the server through the network, obtains the reports and data, generates and displays the inspection report, and issues an alarm when the inspection data exceeds a preset range.

[0050] The server interacts with the AI ​​platform for data exchange, and the AI ​​platform analyzes the environmental data and equipment data to generate energy-saving and emission-reduction setting suggestions for the inspected equipment.

[0051] In practical applications, the present invention is implemented using a low-altitude drone, and corresponding functions are added, such as a laser radar, additional cameras, communication modules, etc., and related algorithms are developed to analyze, record and transmit relevant information to achieve the purpose of the invention.

[0052] The specific implementation process is as follows Figure 3 shown.

[0053] Before implementing drone inspections, it is necessary to model the inspection locations and generate maps. This can be achieved using a variety of environmental modeling techniques, and the methods described below are only examples.

[0054] 1. Conduct 3D modeling of indoor environment

[0055] First, use a monocular or binocular camera to collect image sequences of the indoor environment;

[0056] Start the ORB-SLAM system, extract features from the first frame, use ORB (Oriented FAST and Rotated BRIEF) features for matching, estimate the initial pose of the camera, and start tracking subsequent frames.

[0057] As the camera moves, ORB-SLAM continuously updates the camera's pose through visual odometer, while building and optimizing the map. The map contains keyframes and 3D map points that reflect the geometric structure of the indoor environment.

[0058] The system uses a closed-loop detection mechanism to discover repeated access points in the scene, correct accumulated errors, and ensure the consistency and accuracy of the map.

[0059] 2. Instrument detection and positioning

[0060] Use the YOLO model for real-time instrument detection. First, YOLO needs to be trained using an image dataset containing various types of instruments to ensure that the model can recognize different types of instruments.

[0061] Based on the 3D map constructed by ORB-SLAM, combined with the detection results of YOLO, the instruments in the indoor space are identified and located. YOLO is responsible for detecting instruments from each frame of the image and outputting the bounding box of the instrument and its position in the image.

[0062] Using the camera posture information provided by ORB-SLAM, the two-dimensional instrument position detected by YOLO is mapped into three-dimensional space to obtain the precise position of the instrument in the three-dimensional map.

[0063] Through feedback from the visual odometry and loop closure detection, the positioning accuracy of the instrument is continuously optimized to ensure that the instrument can be accurately identified and tracked even when moving.

[0064] 3. System integration and optimization

[0065] The environment modeling of ORB-SLAM and the instrument detection module of YOLO are seamlessly integrated to form a unified indoor navigation and instrument detection system.

[0066] Optimizing algorithms and data processing flows to ensure that the entire system achieves a balance between real-time performance and computing resource consumption is crucial, especially in mobile robot applications.

[0067] In view of the complexity of indoor environments, such as light changes, occlusions, similar textures, etc., the YOLO model is fine-tuned, and an adaptation strategy for dynamic environments is added to ORB-SLAM to improve the robustness of the system.

[0068] Inspection routes are planned based on the completed modeling.

[0069] Through the flight control system, it supports the setting of fixed flight paths, such as fixed take-off altitude and path to implement fixed-path flight, hover and take photos to record relevant data for specific equipment to be monitored, and set the frequency of daily inspections. Use the corresponding software to burn the correct firmware into the drone's flight control system.

[0070] Flight control parameter settings include: Calibrate sensors such as accelerometers and compasses. Set remote control channels, including roll, pitch, throttle and heading sticks. Calibrate and set output channels, such as the ground steering wheel controlled by the servo. Flight mode configuration: Configure different flight modes as needed, such as manual mode, stabilization mode, fixed-point mode, etc.

[0071] Use the ground control software to plan the route. Set waypoints on the map. Each waypoint can include information such as altitude, speed, camera movement, etc. Then edit each waypoint in detail, including camera movement, flight altitude, speed, nose direction, gimbal pitch angle, etc. You can further set points of interest so that the drone can shoot at specific points during flight.

[0072] Specifically, path planning includes the following:

[0073] 1. Data collation and target sorting

[0074] All instrument points detected by YOLO and located by ORB-SLAM are marked in the 3D map and organized into a list of target points.

[0075] According to the current position of the UAV and the distribution of the target points, the target points are logically sorted, such as from the closest to the farthest, or according to some optimization criteria (such as minimizing the total flight distance) to develop an efficient access sequence.

[0076] 2. Path planning algorithm selection

[0077] A* algorithm: This is a widely used heuristic algorithm for path searching that can effectively find the shortest path from the starting point to the end point. By defining a suitable heuristic function, such as Euclidean distance, the path can be planned while considering obstacles.

[0078] RRT / RRT* (Fast Random Tree / Fast Random Tree Star): Suitable for path planning in dynamic environments and high-dimensional spaces. It generates a tree structure from the starting point to the end point through random sampling, and continuously optimizes to find the optimal path.

[0079] Dijkstra or its variants: If the environment is relatively simple and computing resources are sufficient, the Dijkstra algorithm can also be used to find the shortest path, but this algorithm is less efficient in large-scale complex environments.

[0080] 3. Path smoothing and obstacle avoidance

[0081] The planned path may contain sharp turns. Through path smoothing algorithms (such as B-spline curve fitting, path optimization, etc.), the flight trajectory of the drone can be made smoother, reducing energy consumption and improving flight stability.

[0082] Real-time obstacle avoidance needs to be considered, and an obstacle avoidance system based on LiDAR or vision should be adopted to ensure that dynamic and static obstacles can be effectively avoided during flight.

[0083] Check and confirm that all waypoints and routes are set correctly. Start route execution and the drone will automatically fly according to the preset route. In terms of safety and monitoring, continuously monitor the status of the drone and the surrounding environment when executing a fixed flight path. Be prepared for possible emergencies such as obstacle avoidance or out-of-control behavior. Save the planned route for future reuse or fine-tuning. Load the saved route when necessary and fly the mission. And make necessary adjustments and optimizations to the route based on the results of the flight execution.

[0084] Drone navigation can include the following:

[0085] 1. Navigation control, task execution and result output

[0086] According to the planned path, the navigation control system of the UAV needs to perform path tracking and maintain the accuracy of the flight route through PID controller or other advanced control strategies.

[0087] When approaching each instrument target point, more sophisticated positioning and hovering strategies can be enabled to ensure that the drone can stay stably and perform the corresponding detection or reading tasks.

[0088] Set the collection terminal to start taking photos regularly;

[0089] The image is uploaded to the server via wireless network;

[0090] Through image recognition algorithms, photo readings are converted into numerical results;

[0091] Realize upper-level application services such as remote meter reading, data analysis, and fee management.

[0092] 2. Dynamic replanning

[0093] During actual flight, if new obstacles or environmental changes are encountered, the path planning module should have the ability to dynamically replan and quickly adjust the flight path to adapt to the new situation.

[0094] Furthermore, it is necessary to control the drone to return to the landing pad

[0095] 1. Final target point setting

[0096] First, add the helipad location as the last target point in the path planning. The location of the helipad should be clear in the initial settings of the drone, or accurately calibrated through the SLAM process before the mission begins.

[0097] 2. Path planning adjustment

[0098] If the current location of the drone is far from the landing pad, the existing path needs to be adjusted or replanned to ensure that the drone can return to the landing pad directly and efficiently from the last instrument point. This may require the application of the previously mentioned path planning algorithms (such as A*, RRT* or Dijkstra) again to ensure the safety and efficiency of the path.

[0099] 3. Safety strategy and obstacle avoidance

[0100] On the return journey, continuously monitor the surrounding environment, especially when approaching the apron, as there may be dynamic obstacles in the area (such as other drones, people, or temporary objects). Use obstacle avoidance systems (such as vision- or radar-based perception modules) to ensure that the flight path is clear.

[0101] Landing preparation and precise landing: When approaching the apron, the drone should switch to a more precise control mode, slow down and adjust its attitude in preparation for landing.

[0102] Use vision or sensors (such as ultrasonic sensors, infrared sensors) to accurately measure altitude and position to ensure a smooth landing. Some drones may also use preset landmarks or specific landing marks for precise positioning.

[0103] During the landing process, factors such as wind speed, the vertical speed and horizontal speed of the drone must also be considered, and stability must be maintained through automatic control algorithms until the drone lands safely.

[0104] Subsequent tasks may also include task completion and status checks.

[0105] After the drone lands, it performs routine end-of-mission procedures, including shutting down unnecessary systems to save energy and uploading flight data to a ground station for analysis and archiving.

[0106] Check the drone status, including battery power, sensor status, mechanical component wear, etc., to prepare for the next mission.

[0107] Furthermore, the photo data is read.

[0108] The technology can take pictures, perform infrared temperature measurement and detect combustible gas leaks at a preset hovering position, and use computer vision and machine learning algorithms to analyze and understand the content of the picture. Specifically, it can analyze and calculate the relevant data such as thermometers, pressure gauges, voltmeters, ammeters, power factor meters, illumination, infrared temperature measurement, etc., and automatically generate inspection records. It can also implement remote mobile phone alarms for parameters that are higher or lower than the system presets, including pre-emptive remote alarms for combustible gas leaks and water leaks through combustible gas sensors and related radar wave alarms.

[0109] There are many ways to obtain relevant information in the image, including:

[0110] Use specialized software or tools to read the image data to extract metadata. This data usually includes shooting time, camera model, shooting parameters, etc.

[0111] Using OCR (Optical Character Recognition) technology, you can recognize the text content in the image. There are many online services and software that provide OCR functions.

[0112] Use image recognition techniques to identify objects, scenes, or activities in pictures. This usually requires the use of machine learning models such as convolutional neural networks (CNNs).

[0113] Furthermore, the extracted data is organized into a structured format, such as JSON, XML, or a database table, according to logical relationships.

[0114] Use report generation tools or libraries to generate the required reports based on structured data. Reports can be in Excel, PDF or HTML format and contain charts, tables and text.

[0115] Automate the above steps and create a workflow to process large numbers of images and generate reports.

[0116] This enables the automatic generation of reports for unmanned inspections.

[0117] Since the drone is equipped with a monocular or binocular camera that can transmit image information in real time, instrument detection and identification can be carried out at the same time, or the information on the transmitted images can be processed when the drone arrives at the apron. The specific processing time depends on actual usage needs or personal preferences.

[0118] Before the drone successfully returns to the landing pad and is ready for image data processing, preprocessing the collected instrument reading images is a critical step to ensure that subsequent image analysis and recognition can be accurate and efficient. Using OpenCV for image denoising is a common method in image preprocessing. The specific operation can be divided into the following steps:

[0119] First, load the instrument image taken by the drone through the OpenCV library. This usually involves reading the image file using the cv2.imread() function.

[0120] Depending on the image characteristics and denoising requirements, you may need to convert the image from BGR color space to grayscale space (using cv2.cvtColor() function), because many denoising algorithms perform better on grayscale images.

[0121] Gaussian blur: Use a Gaussian filter to perform preliminary denoising on the image, which can reduce the Gaussian noise in the image. This is achieved through the cv2.GaussianBlur() function, and you need to set a suitable kernel size (such as (5,5)) and standard deviation.

[0122] Median filtering: Median filtering is an effective way to deal with salt and pepper noise. It removes isolated noise points by replacing the central pixel value with the median value of the surrounding pixels.

[0123] Bilateral filtering: If you want to keep edge information while denoising, bilateral filtering is a good choice. It takes into account spatial proximity and pixel value similarity. Use the cv2.bilateralFilter() function and adjust the diameter size, the standard deviation of the color space, and the standard deviation of the spatial Gaussian weight.

[0124] To improve the contrast of the image and make the meter readings clearer, histogram equalization can be applied (using cv2.equalizeHist()). This is especially effective for images with uneven brightness or low contrast.

[0125] Since the current instruments are marked with text labels indicating the instrument types, we use OCR technology to extract key text information from the collected instrument images and classify the instrument types based on the extracted text information.

[0126] The steps of instrument identification include: marking the zero scale point; extracting the dial; fitting the scale line contour to a straight line and finding the intersection point as the center of the circle; finding a straight line based on the pointer contour and extracting multiple straight lines; fitting the contours of multiple straight lines into a straight line, i.e., the pointer; and evaluating based on the graduation value.

[0127] Finally, automatic table generation is performed.

[0128] Connect to the wireless network through the data transmission module and insert the operator's mobile phone card to realize data transmission and remote mobile phone alarm and inspection work order report functions and send them to the developed APP application.

[0129] Mobile data communication emphasizes the non-fixed location of the two communicating parties and the wireless nature of the information transmission link. It is a communication method that transmits data through radio waves. It can be a cellular or quasi-cellular network, and the user terminal can be a smart terminal (such as a mobile phone that supports WAP function).

[0130] You can also use servers and cloud platforms to save data and store it in the developed APP applications. Use the remote servers of cloud service providers for data storage. Cloud storage facilitates data backup and cross-device access

[0131] Furthermore, it can be linked with the AI ​​platform and put forward relevant suggestions for energy conservation and emission reduction based on the analysis of relevant parameters by the AI ​​platform.

[0132] The unmanned inspection method for equipment in the property and equipment facilities industry of the present invention truly realizes all-round unmanned operation. Unmanned automatic inspection; unmanned background data analysis, such as viewing photos or videos to form inspection reports, and pre-setting related alarm work; unmanned automatic flight operation, and fully replacing relevant technicians to perform unmanned inspections according to the requirements of relevant facilities and equipment.

[0133] Take the application of the method of the present invention to the inspection of a distribution room as an example.

[0134] References Figure 2 The inspection route of the power distribution room shown in the figure, using the method of the present invention to implement an inspection includes the following steps:

[0135] Step 1: The drone automatically takes off from position 1, flies to a certain height and hovers, uses the camera to take pictures and related sensors to record the environment, temperature and humidity of the distribution room, and checks for abnormal conditions, such as high temperature and high humidity. Once they exceed the set values, and whether there is any water leakage, etc., if so, an alarm will be sent immediately through the APP.

[0136] Step 2: The drone flies to position 2 and repeats the steps in step 1.

[0137] Step 3: The drone flies to position 3 and repeats the steps in step 1.

[0138] Step 4: The drone flies to position 4 and repeats the steps in step 1.

[0139] Step 5: The drone flies to the No. 5 position, and gradually goes up or down according to the set height. It hovers in front of specific instruments at different heights to take pictures and record relevant parameters such as current, voltage, load, and temperature. If there are any abnormal conditions, the parameters will be recorded in the APP. If the reading exceeds the preset value, an alarm will be sent immediately through the APP.

[0140] Step 6: The drone flies to position 6 and repeats step 5.

[0141] Step 6: The drone flies to position 7 and repeats step 5.

[0142] Step 7: The drone flies to position 8 and repeat step 5.

[0143] Step 8: The drone flies to position 9 and repeat step 5.

[0144] Step 9: The drone will fly to the No. 1 position for automatic charging, and all inspection records will be sent to the client through the APP to form an inspection record.

[0145] Due to the different power loads in the distribution rooms, the sizes of the distribution rooms are also different, and the number of electrical cabinets is also different. The circular digital marks are waypoints, and the routes are digital marks from small to large.

[0146] If manual inspection is adopted in the power distribution room, it is necessary to visually complete the inspection of the instruments at different heights and complete the meter reading records, such as voltage 400+5%V, power factor between 0.9-1, current is recorded according to the actual load, load conditions, check whether there is no water leakage in the room, the ambient temperature is within 40 degrees Celsius, the relative humidity is less than 50%, and other related information, and according to different industry requirements and standards, the daily inspection is not less than 4 times. It is time-consuming and labor-intensive, with high labor costs and poor accuracy.

[0147] The unmanned inspection method for equipment in the property and equipment facility industry adopts the invention. By formulating a drone inspection route, the drone is regularly passed through each waypoint according to the determined inspection frequency to obtain inspection data. The inspection data is then sent to the user end through the network to generate an inspection report, and an alarm is issued when the inspection data exceeds a preset range. This can achieve true unmanned inspection and automatic generation of inspection records and abnormal alarm functions, and has a wide range of applications and relatively low application costs.

[0148] In this specification, the present invention has been described with reference to specific embodiments thereof. However, it is apparent that various modifications and variations may be made without departing from the spirit and scope of the present invention. Therefore, the specification and drawings should be regarded as illustrative rather than restrictive.

Claims

1. An unmanned inspection method for equipment in the property and equipment facilities industry, characterized in that: include: Formulate a drone inspection route, draw a route through inspection waypoints on a map including the inspected device, the waypoints correspond to the locations of the inspected device, and determine the inspection frequency; The drone passes through the waypoints in sequence according to the inspection route and obtains corresponding inspection data at the waypoints; The drone sends the inspection data to the user end via the network; The user terminal generates an inspection report based on the inspection data and issues an alarm when the inspection data exceeds a preset range.

2. The unmanned inspection method for property and equipment facilities industry equipment according to claim 1 is characterized in that: The waypoints include at least one environmental monitoring waypoint and at least one equipment monitoring waypoint, the route includes an unmanned aerial vehicle flight route passing through all environmental monitoring waypoints and equipment monitoring waypoints, and the inspection data includes environmental data and equipment data.

3. The unmanned inspection method for property and equipment facilities industry equipment according to claim 2 is characterized in that: The route also includes the drone's flight altitude, speed, and camera action information, and each waypoint also includes information on the drone's altitude, speed, hovering time, camera orientation, camera action, gimbal pitch angle, and data content corresponding to the photos taken.

4. The unmanned inspection method for property and equipment facilities industry equipment according to claim 3 is characterized in that: The obtaining of corresponding inspection data at the waypoint is specifically as follows: At the environmental monitoring waypoints, the drone uses infrared temperature measurement to detect and obtain temperature information; uses a combustible gas sensor to monitor whether there is combustible gas leakage information; and uses a camera to capture information about water leakage; At the equipment monitoring waypoint, the drone uses a camera to capture pictures containing equipment instruments, and obtains instrument data by analyzing the pictures. The instruments include one or more of a temperature meter, a pressure gauge, a voltmeter, an ammeter, a power factor meter, and an illumination meter.

5. The unmanned inspection method for property and equipment facilities industry equipment according to claim 4 is characterized in that: The method of obtaining instrument data by analyzing pictures specifically includes: Image data extraction, including shooting time, shooting location, camera model and shooting parameters; Text recognition, using OCR technology to identify the text content in the picture, and generate instrument data based on the information corresponding to the data content of the photograph; Image content analysis, using image recognition technology to identify objects, scenes or activities in pictures and generate content data; Report generation organizes the extracted image data, instrument data, and content data into structured data according to logical relationships, and automatically generates reports including charts, tables, and text based on the structured data.

6. The unmanned inspection method for property and equipment facilities industry equipment according to claim 5 is characterized in that: The drone sends the image to the server via the network, and the server implements the image data extraction, text recognition, image content analysis and report generation processing.

7. The unmanned inspection method for property and equipment facilities industry equipment according to claim 6 is characterized in that: The server also stores and backs up data.

8. The unmanned inspection method for property and equipment facilities industry equipment according to claim 7 is characterized in that: The user terminal is connected to the server through the network, obtains the reports and data, generates and displays the inspection report, and issues an alarm when the inspection data exceeds a preset range.

9. The unmanned inspection method for property and equipment facilities industry equipment according to claim 7 is characterized in that: The server interacts with the AI ​​platform for data exchange, and the AI ​​platform analyzes the environmental data and equipment data to generate energy-saving and emission-reduction setting suggestions for the inspected equipment.

Images