Construction Project Unmanned Aerial Vehicle Inspection and Monitoring AI Analysis System

Abstract

This application relates to the technical field of safety supervision and discloses an AI analysis system for construction engineering drone inspection and monitoring. The inspection management module of this invention includes a route management unit and an upload management unit. The upload management unit sets up an upload queue for data stored in batches and uploads it sequentially to a management center located on the ground. Based on the drone's flight position and camera angle when acquiring monitoring data, the upload management unit divides the data stored in batches into a first priority and a second priority, and puts the data with the first priority in the upload queue before the data with the second priority. By storing monitoring data in batches and setting up queues and priorities for uploading, this invention can delay the upload time of monitoring data from secondary locations during the inspection of foundation pits and ensure that monitoring data generated at the location of the foundation pit can be uploaded and analyzed in a timely manner, thereby improving the timeliness of inspection reports generated by unmanned drones for safety supervision.

Description

Technical Field

[0001] This application relates to the technical field of safety supervision, and in particular to an AI analysis system for unmanned aerial vehicle (UAV) inspection and monitoring of construction projects. Background Technology

[0002] Foundation pit construction is the foundation construction that provides the necessary space and conditions for the foundation construction of a building. The construction process involves building foundation columns and digging pits inside the foundation columns to meet the requirements of subsequent construction.

[0003] Currently, drones can be used for safety inspections of foundation pits, but many problems still exist in the process of unmanned drone inspections. For example, when the foundation pit is deep, the communication efficiency of the drone will be greatly reduced, posing challenges to both drone operation and data uploading.

[0004] The characteristic of drone inspection of foundation pits is that the inspection range is concentrated inside the foundation pit and the foundation pit buildings are concentrated at the edge of the foundation pit. This results in a large amount of secondary data that does not involve the foundation pit buildings during the comprehensive inspection process. Under normal circumstances, this secondary data may be cleaned up so as not to affect the foundation pit inspection process and the generation of inspection reports. However, if the drone's data upload speed is slow, this secondary data will seriously restrict the upload of effective data, which may lead to the untimely detection of foundation pit problems and affect the quality of inspection.

[0005] In view of this, the present invention proposes an AI analysis system for construction engineering drone inspection and monitoring. By adjusting the upload order of monitoring data, it ensures that valid data can be uploaded first, thereby improving the timeliness of inspection data. Summary of the Invention

[0006] To ensure timely detection of problems in the foundation pit, this application provides an AI analysis system for construction engineering drone inspection and monitoring by adjusting the upload order of monitoring data to ensure that valid data is uploaded first, thereby improving the timeliness of inspection data.

[0007] Firstly, this application provides an AI analysis system for unmanned aerial vehicle (UAV) inspection and monitoring of construction projects, employing the following technical solution:

[0008] The construction project drone inspection and monitoring AI analysis system includes a drone module, a data management module, an inspection management module, and a management center.

[0009] The drone module includes multiple drone bodies, a positioning unit mounted on the drone body, monitoring sensors, and a camera with an adjustable camera angle. The drone module is used to inspect foundation pit projects and record drone flight position and camera angle data, and to acquire monitoring data, including sensor data and video data.

[0010] The data management module includes a data storage unit, which stores the monitoring data in batches according to the drone's flight position and camera angle when the monitoring data is acquired;

[0011] The inspection management module includes a route management unit and an upload management unit. The upload management unit sets up an upload queue for the data stored in batches and uploads it sequentially to the management center located on the ground. Based on the UAV's flight position and camera angle when acquiring monitoring data, the upload management unit divides the data stored in batches into a first priority and a second priority, and puts the data with the first priority in the upload queue before the data with the second priority.

[0012] The management center is equipped with an analysis unit that generates inspection reports based on uploaded monitoring data. The analysis unit needs to preprocess the acquired monitoring data, including at least denoising, geometric correction, artifact removal, and image restoration. During flight, the drone captures a large number of raw images, which may be affected by uneven lighting, noise, blur, or distortion, thus requiring preprocessing. In this embodiment, the preprocessing steps include: image denoising (using Gaussian filtering or median filtering to eliminate noise), image enhancement (improving image quality through histogram equalization or contrast stretching), geometric correction (eliminating lens distortion and performing perspective transformation to correct image deformation), and image cropping and scaling (extracting regions of interest and adjusting image size to suit analytical needs). Preprocessed images can more clearly reflect the details of the foundation pit project, providing high-quality input data for subsequent crack detection, deformation analysis, or target recognition. Efficient image preprocessing can significantly improve the accuracy and efficiency of drone inspections.

[0013] The above technical solution provides a method for storing monitoring data in batches and setting queues and priorities for uploading. This allows for the postponement of uploading monitoring data from secondary locations during foundation pit inspections, while ensuring that monitoring data generated at the location of the foundation pit structure can be uploaded and analyzed in a timely manner. This improves the timeliness of inspection reports generated by unmanned drones for safety supervision.

[0014] Optionally, the process of storing monitoring data in batches includes:

[0015] A three-dimensional rectangular coordinate system is constructed within the foundation pit, and coordinate data is generated based on the positioning data of the UAV's positioning unit. At the same time, the camera angle data when the UAV is located at the coordinate data is obtained. The origin of the coordinate system is preferably selected from the center point or corner point of the foundation pit.

[0016] The storage unit is equipped with a read-only address and a write address. The monitoring data is stored in batches on the write address, and each write address is equipped with a corresponding read-only address. The read-only address stores coordinate data and camera angle data.

[0017] Optionally, the process of setting up an upload queue and uploading data stored in batches includes:

[0018] Obtain the priority of the current batch storage information, and set the upload queue according to the priority and the acquisition time;

[0019] Based on the upload speed and historical data of the inspection reports, the upload order of each batch of the first priority is adjusted. If there are no historical inspection reports for the current foundation pit, no adjustment is made.

[0020] The above technical solution provides a method for managing the upload order of each batch through a first priority and a second priority. Specifically, this invention sets the monitoring data generated by UAVs within the range close to the foundation pit structure as the first priority machine, thereby ensuring that the monitoring data containing the foundation pit structure can be uploaded first, ensuring that key data can be transmitted in a timely manner, providing data support for foundation pit safety monitoring and early warning. Through priority management, it can also avoid non-critical data occupying too much network bandwidth and improve network resource utilization.

[0021] Optionally, the process of dividing data stored in batches into first priority and second priority includes:

[0022] Pre-set verification conditions, access the write address bit corresponding to the read-only address before generating the queue, and determine whether the information stored at the write address can meet the verification conditions;

[0023] If the conditions are met, the monitoring data stored at the write address corresponding to the current read-only address is determined to be of the first priority; otherwise, the monitoring data stored at the write address corresponding to the current read-only address is determined to be of the second priority.

[0024] Optionally, the process of adjusting the upload order of each batch of the first priority includes:

[0025] Get the upload speed of the drone upload management unit. If the upload speed is lower than the preset base value, adjust the queue order of the first priority batches of stored data in the queue.

[0026] If the upload speed is not lower than the preset base value, no adjustment will be made.

[0027] Optionally, the process of adjusting the upload order of each batch of the first priority also includes:

[0028] Before the inspection, the inspection management module obtains historical data of the inspection report and counts the images that appear in the inspection report and the read-only address information corresponding to the batch of the image.

[0029] The priority coefficient is obtained based on the first occurrence count of read-only address information in multiple inspection reports and the second occurrence count of read-only address information in the most recent inspection report.

[0030] The batches with the highest priority are ranked according to their priority coefficients.

[0031] Optionally, the process of obtaining the priority coefficient includes:

[0032] Through the formula:

[0033] ;

[0034] Get priority coefficient ,in, It is the first occurrence. It is the second occurrence. It is the preset first standard value. It is the preset second standard value. It is the preset first weight value. It is the preset second weight value. It is a correction value set based on the read-only address information acquisition time corresponding to the current batch. The correction value is used to correct data when the inspection report only appears once. For example, if two data appear the same number of times in an inspection report, without a correction value, their priority coefficients will also be the same, which will make it impossible to sort them by priority value.

[0035] The above technical solution provides a way to sort batches with the highest priority. By obtaining a priority coefficient, the present invention adjusts the upload queue when the upload speed is poor, thereby further ensuring that critical data can be uploaded in a timely manner.

[0036] Optionally, the process by which the analysis unit generates an inspection report based on the uploaded monitoring data includes:

[0037] Preprocess the monitoring data to obtain the processed sensor data and video data;

[0038] Sensor data and video data are input into a trained neural network model, and the corresponding neural network model outputs the recognition result.

[0039] An inspection report is generated based on the recognition results.

[0040] The process of setting the verification conditions includes:

[0041] Obtain the working range of the drone and the outline of the edge of the foundation pit;

[0042] The verification conditions are set as follows: the current coordinate data of the drone is within the outline and the distance between the drone and the outline does not exceed the working range of the drone; and the camera direction, determined based on the camera angle, points to the vertical plane where the pit outline is located.

[0043] The system also includes an alarm notification module and a progress display module. The alarm notification module notifies relevant personnel when safety hazards such as cracks or deformation are detected in the identification results. The progress display module records the construction progress through drones, providing visualized data for project management.

[0044] Drones capture ground images using high-resolution or multispectral cameras and, in conjunction with lidar, acquire high-precision point cloud data by emitting laser pulses and receiving reflected signals. The collected images and sensor data are transmitted wirelessly to a ground station in real time. Photogrammetry software then stitches, corrects, and reconstructs the images in 3D to generate digital elevation models or 3D point cloud models. Based on these models, project management is achieved through visualized monitoring.

[0045] In summary, this application includes at least the following beneficial technical effects:

[0046] (1) This invention provides a technical solution for storing monitoring data in batches and setting queues and priorities for uploading. This allows the uploading time of monitoring data from secondary locations to be delayed during the inspection of the foundation pit, and enables the monitoring data generated at the location of the foundation pit building to be uploaded and analyzed in a timely manner, thereby improving the timeliness of the inspection reports generated by unmanned drones in safety supervision.

[0047] (2) By setting the monitoring data generated by the UAV in the range close to the foundation pit building as the first priority machine, the monitoring data containing the foundation pit building can be uploaded first, ensuring that key data can be transmitted in a timely manner, providing data support for foundation pit safety monitoring and early warning. Through priority management, it can also avoid non-critical data occupying too much network bandwidth and improve the utilization rate of network resources.

[0048] (3) By obtaining the priority coefficient, the upload queue is adjusted when the upload speed is poor, so as to further ensure that the key data can be uploaded in a timely manner. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the system modules.

[0050] Figure 2 This is a schematic diagram of the components of the inspection management module. Detailed Implementation

[0051] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0052] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0053] This application discloses an AI analysis system for unmanned aerial vehicle (UAV) inspection and monitoring of construction projects, referring to... Figure 1 and Figure 2 It includes a drone module, a data management module, an inspection management module, and a management center.

[0054] The drone module includes multiple drone bodies, a positioning unit mounted on the drone body, monitoring sensors, and a camera with an adjustable camera angle. The drone module is used to inspect foundation pit projects and record drone flight position and camera angle data, and to acquire monitoring data, including sensor data and video data. The drone body should integrate a GPS module, an IMU (inertial measurement unit), and a compass.

[0055] The data management module includes a data storage unit, which stores the monitoring data in batches based on the drone's flight position and camera angle when the monitoring data is acquired.

[0056] The inspection management module includes a route management unit and an upload management unit. The upload management unit sets up upload queues for data stored in batches and uploads it sequentially to the management center located on the ground. Based on the drone's flight position and camera angle when acquiring monitoring data, the upload management unit divides the data stored in batches into first and second priorities, ensuring that first-priority data is uploaded before second-priority data. Data upload can be achieved via Wi-Fi, 4G / 5G networks, or a dedicated radio link. Furthermore, the collected data is packaged according to a predetermined format and uploaded to a designated storage location in real-time or periodically via a communication link. Uploaded data includes GPS coordinates, flight altitude, speed, camera angle, image or video files, etc., which are used for subsequent analysis, report generation, or task optimization. To ensure data integrity and security, verification mechanisms (such as CRC checksums) and encryption measures (such as SSL / TLS encryption) can be added during the upload process. Through efficient data upload, real-time monitoring and remote management of drone missions can be achieved, providing strong support for tasks such as foundation pit inspection.

[0057] The management center is equipped with an analysis unit that generates inspection reports based on uploaded monitoring data. The analysis unit needs to preprocess the acquired monitoring data, including at least denoising, geometric correction, artifact removal, and image restoration. Image data preprocessing aims to improve image quality and extract useful information, laying the foundation for subsequent analysis. In this embodiment, the UAV captures a large number of raw images during flight. These images may be affected by factors such as uneven lighting, noise, blur, or distortion, thus requiring preprocessing. The preprocessing steps in this embodiment include: image denoising (using Gaussian filtering or median filtering to eliminate noise), image enhancement (improving image quality through histogram equalization or contrast stretching), geometric correction (eliminating lens distortion and performing perspective transformation to correct image deformation), and image cropping and scaling (extracting regions of interest and adjusting image size to suit analytical needs). Preprocessed images can more clearly reflect the details of the foundation pit project, providing high-quality input data for subsequent crack detection, deformation analysis, or target recognition. Efficient image preprocessing can significantly improve the accuracy and efficiency of UAV inspections.

[0058] This embodiment provides a technical solution for storing monitoring data in batches and setting queues and priorities for uploading. This allows the uploading time of monitoring data from secondary locations to be delayed during foundation pit inspections, while ensuring that monitoring data generated at the location of the foundation pit structure can be uploaded and analyzed in a timely manner. This improves the timeliness of inspection reports generated by unmanned drones for safety supervision.

[0059] The process of storing monitoring data in batches includes:

[0060] A three-dimensional rectangular coordinate system is constructed within the foundation pit, and coordinate data is generated based on the positioning data of the UAV's positioning unit. At the same time, the camera angle data when the UAV is located at the coordinate data is obtained. The origin of the coordinate system is preferably selected from the center point or corner point of the foundation pit.

[0061] The storage unit is equipped with read-only addresses and write addresses. The monitoring data is stored in batches on the write addresses, and each write address is equipped with a corresponding read-only address. The read-only address stores coordinate data and camera angle data.

[0062] The process of setting up an upload queue and uploading data stored in batches includes:

[0063] Obtain the priority of the current batch storage information, and set the upload queue according to the priority and the acquisition time;

[0064] Based on the upload speed and historical data of the inspection reports, the upload order of each batch of the first priority is adjusted. If there are no historical inspection reports for the current foundation pit, no adjustment is made.

[0065] This embodiment provides a technical solution for managing the upload order of each batch through a first priority and a second priority. Specifically, the present invention sets the monitoring data generated by the UAV within the range close to the foundation pit structure as the first priority machine, thereby ensuring that the monitoring data containing the foundation pit structure can be uploaded first, ensuring that key data can be transmitted in a timely manner, providing data support for foundation pit safety monitoring and early warning. Through priority management, it can also avoid non-critical data occupying too much network bandwidth and improve network resource utilization.

[0066] The process of dividing data stored in batches into first and second priorities includes:

[0067] Pre-set verification conditions, access the write address bit corresponding to the read-only address before generating the queue, and determine whether the information stored at the write address can meet the verification conditions;

[0068] If the conditions are met, the monitoring data stored at the write address corresponding to the current read-only address is determined to be of the first priority; otherwise, the monitoring data stored at the write address corresponding to the current read-only address is determined to be of the second priority.

[0069] The process of adjusting the upload order of each batch of the first priority includes:

[0070] Get the upload speed of the drone upload management unit. If the upload speed is lower than the preset base value, adjust the queue order of the first priority batches of stored data in the queue.

[0071] If the upload speed is not lower than the preset base value, no adjustment will be made.

[0072] The process of adjusting the upload order of each batch of the first priority also includes:

[0073] Before the inspection, the inspection management module obtains historical data of the inspection report and compiles statistics on the images that appear in the inspection report and the read-only address information corresponding to the batches in which the images are located.

[0074] The priority coefficient is obtained based on the first occurrence count of read-only address information in multiple inspection reports and the second occurrence count of read-only address information in the most recent inspection report.

[0075] The batches with the highest priority are ranked according to their priority coefficients.

[0076] In this embodiment, the process of obtaining the priority coefficient includes:

[0077] Through the formula:

[0078] ;

[0079] Get priority coefficient ,in, It is the first occurrence. It is the second occurrence. It is the preset first standard value. It is the preset second standard value. It is the preset first weight value. It is the preset second weight value. It is a correction value set based on the read-only address information acquisition time corresponding to the current batch. The correction value is used to correct data when the inspection report only appears once. For example, if two data appear the same number of times in an inspection report, without a correction value, their priority coefficients will also be the same, which will make it impossible to sort them by priority value.

[0080] This embodiment provides a technical solution for sorting batches with the first priority. By obtaining a priority coefficient, the present invention adjusts the upload queue when the upload speed is poor, thereby further ensuring that critical data can be uploaded in a timely manner.

[0081] The process by which the analysis unit generates an inspection report based on the uploaded monitoring data includes:

[0082] Preprocess the monitoring data to obtain the processed sensor data and video data;

[0083] Sensor data and video data are input into a trained neural network model, and the corresponding neural network model outputs the recognition result.

[0084] An inspection report is generated based on the recognition results.

[0085] The process of setting verification conditions includes:

[0086] Obtain the working range of the drone and the outline of the edge of the foundation pit;

[0087] The verification conditions are set as follows: the current coordinate data of the drone is within the outline and the distance between the drone and the outline does not exceed the working range of the drone; and the camera direction, determined based on the camera angle, points to the vertical plane where the pit outline is located.

[0088] The system provided by this invention also includes an alarm notification module and a progress display module. The alarm notification module notifies relevant personnel when safety hazards such as cracks or deformations are detected in the identification results. The progress display module records the construction progress using a drone, providing visualized data for project management. Specifically, in this embodiment, the drone uses a high-resolution camera or multispectral camera to capture ground images. Simultaneously, a GPS module records the precise location information of each image, and an IMU (Inertial Measurement Unit) provides the drone's attitude data to ensure spatial consistency of the images. Furthermore, in conjunction with a lidar system, high-precision point cloud data is acquired by emitting laser pulses and receiving reflected signals. The acquired images and sensor data are transmitted to a ground station in real time via wireless communication. Then, photogrammetry software is used to stitch, correct, and reconstruct the images in three dimensions, generating a digital elevation model or a three-dimensional point cloud model. Based on the acquired model, visualized monitoring is provided for project management.

[0089] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. An AI analysis system for unmanned aerial vehicle (UAV) inspection and monitoring in construction projects, characterized in that: Includes a drone module, a data management module, an inspection management module, and a management center: The drone module includes multiple drone bodies, a positioning unit mounted on the drone body, monitoring sensors, and a camera with an adjustable camera angle. The drone module is used to inspect foundation pit projects and record drone flight position and camera angle data, and to acquire monitoring data, including sensor data and video data. The data management module includes a data storage unit, which stores the monitoring data in batches according to the drone's flight position and camera angle when the monitoring data is acquired; The inspection management module includes a route management unit and an upload management unit. The upload management unit sets up an upload queue for the data stored in batches and uploads it sequentially to the management center located on the ground. Based on the UAV's flight position and camera angle when acquiring monitoring data, the upload management unit divides the data stored in batches into a first priority and a second priority, and puts the data with the first priority in the upload queue before the data with the second priority. The process of dividing data stored in batches into first and second priorities includes: Pre-set verification conditions, access the write address bit corresponding to the read-only address before generating the queue, and determine whether the information stored at the write address can meet the verification conditions; If the conditions are met, the first priority is to determine that the write address corresponding to the current read-only address is the stored monitoring data; otherwise, the second priority is to determine that the write address corresponding to the current read-only address is the stored monitoring data. The process of adjusting the upload order of each batch of the first priority also includes: The inspection management module obtains historical data of the inspection report before the inspection, and counts the images that appear in the inspection report and the read-only address information corresponding to the batch of the images. The priority coefficient is obtained based on the first occurrence count of read-only address information in multiple inspection reports and the second occurrence count of read-only address information in the most recent inspection report. The batches with the highest priority are ranked according to their priority coefficients. The process of obtaining the priority coefficient includes: Through the formula: ； Get priority coefficient ,in, It is the first occurrence. It is the second occurrence. It is the preset first standard value. It is the preset second standard value. It is the preset first weight value. It is the preset second weight value. This is a correction value set based on the read-only address information retrieval time setting corresponding to the current batch; The management center is equipped with an analysis unit, which generates inspection reports based on the uploaded monitoring data. The process of setting the verification conditions includes: Obtain the working range of the drone and the outline of the edge of the foundation pit; The verification conditions are set as follows: the current coordinate data of the drone is within the outline and the distance between the drone and the outline does not exceed the working range of the drone; and the camera direction, determined based on the camera angle, points to the vertical plane where the pit outline is located.

2. The construction project drone inspection and monitoring AI analysis system according to claim 1, characterized in that, The process of storing monitoring data in batches includes: A three-dimensional rectangular coordinate system is constructed within the foundation pit, and coordinate data is generated based on the positioning data of the UAV's positioning unit. At the same time, the camera angle data when the UAV is located at the coordinate data is obtained. The storage unit is equipped with a read-only address and a write address. The monitoring data is stored in batches on the write address, and each write address is equipped with a corresponding read-only address. The read-only address stores coordinate data and camera angle data.

3. The construction engineering drone inspection and monitoring AI analysis system according to claim 2, characterized in that, The process of setting up an upload queue and uploading data stored in batches includes: Obtain the priority of the current batch storage information, and set the upload queue according to the priority and the acquisition time; Based on upload speed and historical data from inspection reports, the upload order of each batch of the first priority is adjusted.

4. The construction engineering drone inspection and monitoring AI analysis system according to claim 3, characterized in that, The process of adjusting the upload order of each batch of the first priority includes: Get the upload speed of the drone upload management unit. If the upload speed is lower than the preset base value, adjust the queue order of the first priority batches of stored data in the queue. If the upload speed is not lower than the preset base value, no adjustment will be made.

5. The construction engineering drone inspection and monitoring AI analysis system according to claim 1, characterized in that, The process by which the analysis unit generates an inspection report based on the uploaded monitoring data includes: Preprocess the monitoring data to obtain the processed sensor data and video data; Sensor data and video data are input into a trained neural network model, and the corresponding neural network model outputs the recognition result. An inspection report is generated based on the recognition results.

6. The construction engineering drone inspection and monitoring AI analysis system according to claim 1, characterized in that, It also includes an alarm notification module and a progress display module. The alarm notification module notifies relevant personnel when cracks or deformations are detected in the identification results. The progress display module records the construction progress through drones, providing visualized data for project management.

Images