A method and system for generating engineering inspection trajectories based on panoramic images

Abstract

The present invention belongs to the field of engineering inspection technology, and specifically relates to a method and system for generating engineering inspection trajectories based on panoramic images. Panoramic camera equipment is used to perform automatic inspections along a preset path or a path dynamically generated by an intelligent planning algorithm, thereby reducing the need for manual operation. A panoramic video data set with geographic reference is formed through spatial position and plane map attachment technology, ensuring that each photo and each frame of the image has precise spatial positioning, facilitating subsequent detailed analysis and visual display. The digitally collected data is transmitted to a central processing system in real time. Combined with the movement trajectory and timestamp information of the inspection personnel, the method of the present invention not only improves the automation level and data accuracy of engineering inspections, but also provides strong support for data analysis and decision-making after the inspection, thereby promoting the development of engineering management and monitoring to a higher level of intelligence and automation.

Description

Technical Field

[0001] The present invention belongs to the technical field of engineering inspection, and in particular relates to a method and system for generating an engineering inspection trajectory based on panoramic images. Background Art

[0002] In the current field of engineering inspections, traditional inspection methods rely primarily on manual on-site inspections. This method typically involves inspectors carrying paper record sheets or portable devices to manually record information, take photos, and mark key areas at the construction site. While this method can meet inspection needs to a certain extent, it has obvious limitations:

[0003] Inefficiency: Manual inspections require a lot of time and human resources, are slow, and are difficult to cover large or complex engineering sites.

[0004] Susceptible to human factors: The quality and accuracy of inspection results are highly dependent on the professional level and personal status of the inspectors, and omissions or misjudgments are prone to occur.

[0005] Lack of comprehensive coverage: Due to manpower limitations, certain hard-to-reach locations or hidden corners may not be fully inspected, resulting in incomplete inspections.

[0006] In summary, the most significant problem with existing technologies lies in the inefficiency and accuracy of data collection and analysis. Traditional manual inspections are not only time-consuming and labor-intensive, but also make it difficult to ensure that all inspection points are thoroughly inspected, especially for large or complex engineering projects. Furthermore, manually recorded data often lacks precise spatial location and timestamp information, making subsequent data processing and analysis difficult and impacting the effectiveness and timeliness of decision-making. Summary of the Invention

[0007] The purpose of the present invention is to provide a method and system for generating engineering inspection trajectories based on panoramic images, which realizes the comprehensive automation and digitization of engineering inspections through a series of innovative technical steps to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention provides a method for generating an engineering inspection trajectory based on panoramic images, comprising the following steps:

[0009] Deploy at least one panoramic camera device at the construction site, the panoramic camera device equipped with multiple cameras for capturing 360-degree environmental images without blind spots, and use the panoramic camera device to automatically patrol along a preset path or a path dynamically generated by an intelligent planning algorithm, and collect panoramic image data in real time;

[0010] During the automated inspection process, mobile devices are used to capture photos of key and problematic areas on the construction site. These photos are then automatically linked to the inspection trajectory plan for the corresponding locations. Based on the spatial location information of the photos and panoramic images, a geo-referenced panoramic video dataset is generated using spatial location and plan linking technology.

[0011] The obtained panoramic video data set is digitally collected, and the original images and photos are converted into digital format. At the same time, the movement trajectory and timestamp information of the inspection personnel are simultaneously collected. The digitally collected data and movement trajectory information are transmitted to the central processing system in real time through the AI ​​workstation;

[0012] Combining digitally collected data and movement trajectory information, the inspection trajectory is displayed in the form of lines or marks on the floor plan. Panoramic images and high-definition photos are integrated with the inspection path data for display. Based on the integrated display, users can explore different perspectives and areas through an interactive interface.

[0013] Preferably, the deployment of at least one panoramic camera device at the construction site includes:

[0014] Determine the locations of multiple observation points at the project site. Each observation point corresponds to one or more panoramic camera installation locations. Install panoramic cameras at the selected locations. For any two adjacent cameras, their field of view overlap ratio is at least P%, where P represents the ratio of the intersection of the fields of view between the two devices to the total field of view of a single device.

[0015] Based on the adjusted device settings, the panoramic camera device is started to capture images. The device shoots at a fixed frame rate F, where F represents the number of images captured per second. At the same time, the geographic coordinates (X, Y, Z) of each device are recorded for spatial positioning of subsequent images.

[0016] During the image acquisition process, the device posture data corresponding to each frame of image is synchronously recorded, including the rotation angle θ, tilt angle φ, and yaw angle ψ, and the orientation matrix M = R(θ)*T(φ)*P(ψ) of each device relative to the reference coordinate system is calculated, where R, T, and P represent the rotation operations around the three axes, respectively.

[0017] Preferably, the use of the panoramic camera device to perform automatic inspection along a preset path or a path dynamically generated by an intelligent planning algorithm and to collect panoramic image data in real time includes:

[0018] Set the inspection starting point and target end point, and plan at least one inspection route based on the map information of the construction site. This route consists of consecutive coordinate points (X_i, Y_i, Z_i), where i represents the i-th coordinate point on the path;

[0019] Start the panoramic camera device and move it along the defined route. At the same time, begin capturing images at a fixed frame rate F. For each coordinate point (X_i, Y_i, Z_i), calculate the time required for the device to reach that point T_i = D_i / V, where D_i is the distance from the previous point to the current point, and V is the device's movement speed.

[0020] During the movement, when the device approaches the next coordinate point, the device's posture parameters (θ_i, φ_i, ψ_i) are adjusted so that the center of the device's field of view is aligned with the direction of the next coordinate point. The posture adjustment is based on the formula M_i = R(θ_i)*T(φ_i)*P(ψ_i), where R, T, and P represent the operation of rotation around the three axes respectively;

[0021] After completing the posture adjustment, it immediately starts recording the panoramic image data of the location and saves the data together with the current position coordinates (X_i, Y_i, Z_i), timestamp T_i and posture parameters (θ_i, φ_i, ψ_i). Then, it continues to move to the next coordinate point until the inspection route ends.

[0022] Preferably, the step of automatically attaching the photos to the inspection track plan at the corresponding position includes:

[0023] When the inspection device reaches the preset key or problem area, it triggers the shooting command of the mobile terminal device to determine the coordinates (X_j, Y_j, Z_j) of each shooting position P_j;

[0024] At the designated location P_j, the mobile terminal device captures a high-definition photo I_j at the highest resolution. At the same time, it records the timestamp T_j of the capture moment and the device posture parameters (θ_j, φ_j, ψ_j). For each photo I_j, the spatial distance D_j to the most recent panoramic image acquisition point is calculated as sqrt((X_j-X_i)^2+(Y_j-Y_i)^2+(Z_j-Z_i)^2), where (X_i, Y_i, Z_i) are the coordinates of the most recent panoramic image acquisition point.

[0025] Based on the timestamp T_j and spatial distance D_j, associate the high-definition photo I_j with the corresponding inspection track location, and use the formula W_j = exp(-α*D_j^2) to calculate the weight W_j, where α is an adjustment parameter;

[0026] The high-definition photos I_j and their associated weights W_j, timestamps T_j, coordinates (X_j, Y_j, Z_j), and posture parameters (θ_j, φ_j, ψ_j) are integrated into a dataset of inspection trajectory planar maps. This dataset allows users to browse photos by clicking or dragging the interactive interface and intuitively see the relationship between the photos and the actual location.

[0027] Preferably, forming a panoramic video dataset with a geographical reference includes:

[0028] Collect all high-definition photos I_j and their corresponding spatial position coordinates (X_j, Y_j, Z_j), timestamps T_j and posture parameters (θ_j, φ_j, ψ_j). At the same time, obtain the continuous image frames F_i recorded by the panoramic camera device in the same time period and their corresponding geographical location coordinates (X_i, Y_i, Z_i) and timestamps T_i;

[0029] For each high-definition photo I_j, find the panoramic image frame F_i closest to its timestamp T_j, calculate the time difference ΔT = |T_j - T_i| between the two to ensure the time synchronization between the photo and the image frame, and select the photo and image frame pairs that meet the threshold ΔT < τ, where τ is a preset time tolerance;

[0030] Based on the selected photo and image frame pair, use the spatial position coordinates (X_j, Y_j, Z_j) and (X_i, Y_i, Z_i) to calculate the spatial distance between them D_ij = sqrt((X_j - X_i)^2 + (Y_j - Y_i)^2 + (Z_j - Z_i)^2). Calculate the association weight W_ij using the formula W_ij = exp(-β*D_ij^2), where β is an adjustment parameter.

[0031] The determined photo and image frame pairs, along with their weights Wij, timestamps Tj / Ti, coordinates (Xj, Yj, Zj) / (Xi, Yi, Zi) and pose parameters (θj, φj, ψj), are integrated into a unified dataset.

[0032] Preferably, the process of digitally collecting the obtained panoramic video data set, converting the original images and photos into digital format, and synchronously collecting the movement trajectory and timestamp information of the inspection personnel includes:

[0033] Receive a georeferenced panoramic video dataset. For each pair of high-definition photo I_j and panoramic image frame F_i, extract the binary data streams B_j and B_i of the original image and photo, and obtain the associated timestamps T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and posture parameters (θ_j, φ_j, ψ_j).

[0034] Convert the binary data streams B_j and B_i into compressed formats C_j and C_i using the formula C_x = Compress(B_x,r), where x represents j or i, Compress indicates the compression operation, and r is the compression ratio. At the same time, the original precision of the timestamps T_j / T_i, coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and attitude parameters (θ_j, φ_j, ψ_j) is preserved.

[0035] During data processing, the patrol personnel's movement trajectory is recorded synchronously. The positioning device integrated in the patrol equipment regularly samples the patrol personnel's position coordinates (X_p, Y_p, Z_p) and timestamp T_p. The time interval ΔT_p = T_p - T_(p-1) between two adjacent samples is calculated to ensure trajectory continuity and time synchronization. For each timestamp T_p, the corresponding patrol personnel posture parameters (θ_p, φ_p, ψ_p) are recorded to form a complete movement trajectory record.

[0036] The compressed image data C_j / C_i, timestamp T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), posture parameters (θ_j, φ_j, ψ_j) and the inspection personnel's movement trajectory (X_p, Y_p, Z_p), timestamp T_p and posture parameters (θ_p, φ_p, ψ_p) are integrated into a data set.

[0037] Preferably, the digitally collected data and movement trajectory information are transmitted to the central processing system in real time via the AI ​​workstation, including:

[0038] Extract the compressed image data C_j / C_i, timestamp T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), posture parameters (θ_j, φ_j, ψ_j), and the inspection personnel's movement trajectory (X_p, Y_p, Z_p), timestamp T_p and posture parameters (θ_p, φ_p, ψ_p) from the integrated data set;

[0039] Pack the data into transmission packets D. Each transmission packet D contains one or more image frames and their associated information, and is accompanied by a checksum (D) = Hash (Summary (D)), where Summary (D) represents the summary of the transmission packet content and Hash is a hash function.

[0040] The generated transmission packet D is sent to the central processing system in real time. After each transmission, the sending time and the receiving confirmation time are recorded, and the transmission delay ΔT_transmit = T_receive - T_send is calculated. For each transmission packet D, the corresponding transmission delay ΔT_transmit is recorded to form a transmission log. T_send is the timestamp of the data packet transmission from the sender, and T_receive is the timestamp of the data packet arriving at the receiver and being successfully received.

[0041] After the central processing system receives the transmission packet D, it unpacks it and verifies the checksum (D). Then, based on the timestamps T_j / T_i and T_p, the image data C_j / C_i and the inspection personnel's movement trajectory (X_p, Y_p, Z_p) are synchronized and integrated into the database. The timestamp T_x is adjusted using the formula Sync(T_x,ΔT_transmit) = Adjust(T_x+ΔT_transmit), where x represents j / i or p, and T_x represents the original timestamp.

[0042] Preferably, the fusion display of the panoramic image and high-definition photo with the inspection route data includes:

[0043] All data transmitted and integrated from the central processing system is obtained. Based on the inspection personnel's movement trajectory (X_p, Y_p, Z_p), the inspection route is drawn as a line on the project site plan. For each trajectory point P_p, the formula P_line = Line(P_(p-1), P_p) is used, where Line represents the line segment connecting two consecutive trajectory points, forming a coherent inspection route map. At the same time, the timestamp T_p of each trajectory point is recorded;

[0044] According to the inspection route map, the panoramic image frame F_i and the high-definition photo I_j are associated with the corresponding trajectory point P_p using the spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i) and timestamp T_j / T_i;

[0045] Find the closest trajectory point P_p using the formula D_associate=min(Distance(P_p,(X_x,Y_x,Z_x))), where x represents j or i, and Distance represents the distance between two points.

[0046] Markers are added to the corresponding positions on the inspection path map to display the panoramic image frame F_i and the high-definition photo I_j. For each marker M_x, the timestamp T_x of the shooting time and the posture parameters (θ_x, φ_x, ψ_x) are displayed, providing users with interactive browsing functions. The formula Overlay(M_x, P_line) = Merge(M_x, Line(P_(p-1), P_p)) is used to merge the marker with the inspection path segment to form a comprehensive view.

[0047] Preferably, the fusion display enables users to explore different perspectives and areas through an interactive interface, including:

[0048] A map view is constructed based on the integration of the inspection route map and image data. In this map view, each inspection trajectory point P_p and its associated panoramic image frame F_i and high-definition photo I_j are marked. For each marker M_x, its visible range V_x is defined as a sphere with a radius R and a center at (X_x, Y_x, Z_x). The formula is expressed as V_x = Sphere((X_x, Y_x, Z_x), R).

[0049] Based on the established map view, zoom and pan operations are provided to allow users to focus on the area of ​​interest. When the user selects a specific marker M_x, the system calculates the distance D_view between the marker and the user's current viewpoint and uses the formula ZoomLevel = f(D_view), where f is a mapping function, to determine the appropriate zoom level.

[0050] According to user operation, the multi-view browsing mode is activated. For the selected marker M_x, its posture parameters (θ_x, φ_x, ψ_x) are used to construct the transformation matrix T_x to adjust the angle of image display;

[0051] Introducing an instant message prompt box. When the user hovers over or clicks a marker M_x, relevant information is immediately displayed, including but not limited to the timestamp T_x, location coordinates (X_x, Y_x, Z_x), and a brief description.

[0052] On the other hand, the present invention proposes a system for generating engineering inspection trajectories based on panoramic images, comprising:

[0053] A panoramic camera deployment module is configured to deploy at least one panoramic camera at the construction site. The panoramic camera is equipped with multiple cameras for capturing 360-degree images of the environment without blind spots. The panoramic camera is used to perform automatic inspections along a preset path or a path dynamically generated by an intelligent planning algorithm, and to collect panoramic image data in real time.

[0054] The high-definition photo shooting and attachment module is used to capture photos of key and problematic areas of the construction site using mobile devices during the automated inspection process. These photos are then automatically attached to the inspection track plan at the corresponding locations. Based on the spatial location information of the photos and panoramic images, a panoramic video dataset with geo-referenced information is generated through spatial location and plan attachment technology.

[0055] The digital acquisition and synchronous recording module is used to digitally acquire the acquired panoramic video data set, convert the original images and photos into digital format, and simultaneously collect the movement trajectory and timestamp information of the inspection personnel. The digitally acquired data and movement trajectory information are transmitted to the central processing system in real time through the AI ​​workstation;

[0056] The inspection trajectory visualization display module is used to combine digitally collected data and movement trajectory information to display the inspection trajectory in the form of lines or marks on the floor plan. It integrates panoramic images and high-definition photos with the inspection path data for display. Based on the integrated display, the interactive interface enables users to explore different perspectives and areas.

[0057] Technical effects and advantages of the present invention: Compared with the existing technology, the method and system for generating engineering inspection trajectories based on panoramic images proposed in the present invention have the following advantages:

[0058] The present invention integrates panoramic camera equipment, high-definition cameras and positioning systems. The method realizes all-round, high-definition automatic inspection of the engineering site. The panoramic camera equipment performs automatic inspection along a preset path or a path dynamically generated by an intelligent planning algorithm, reducing the need for manual operation, ensuring the comprehensiveness and consistency of inspection coverage, and improving inspection efficiency. Through spatial position and plan view hooking technology, a panoramic video data set with geographical reference is formed to ensure that each photo and each frame of image has precise spatial positioning, which is convenient for subsequent detailed analysis and visual display. The digitally collected data is transmitted to the central processing system in real time. Combined with the movement trajectory and timestamp information of the inspection personnel, a comprehensive and interactive interface is provided. Users can quickly browse information from different perspectives and areas through this interface, supporting rapid analysis and decision-making, as well as the identification and response of complex problems. The method of the present invention not only improves the automation level and data accuracy of engineering inspections, but also provides strong support for data analysis and decision-making after inspections, thereby promoting the development of engineering management and monitoring to a higher level of intelligence and automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of a method for generating an engineering inspection trajectory based on panoramic images according to the present invention;

[0060] Figure 2This is a block diagram of a system for generating engineering inspection trajectories based on panoramic images according to the present invention. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0062] The present invention provides a method for generating an engineering inspection trajectory based on panoramic images, such as Figure 1 As shown, the following steps are included:

[0063] Step 1: Deploy at least one panoramic camera device at the construction site, wherein the panoramic camera device is equipped with multiple cameras for capturing 360-degree environmental images without blind spots; further comprising:

[0064] Determine the locations of multiple observation points at the construction site. Each observation point corresponds to one or more panoramic camera installation positions. Panoramic cameras are installed according to the selected locations. For any two adjacent devices, the overlap ratio of their fields of view is at least P%, where P represents the ratio of the intersection of the fields of view between the two devices to the total field of view of a single device. This setting not only avoids visual blind spots, but also provides redundant data, enhancing the accuracy of image stitching and 3D reconstruction.

[0065] Based on the adjusted device settings, the panoramic camera device is started to capture images. The device shoots at a fixed frame rate F, where F represents the number of images captured per second. At the same time, the geographic coordinates (X, Y, Z) of each device are recorded for the spatial positioning of subsequent images; this provides an accurate geographic reference for the subsequent spatial positioning of images, allowing each frame of the image to be accurately located at its actual location.

[0066] During image acquisition, the device posture data corresponding to each frame of image is synchronously recorded, including the rotation angle θ, tilt angle φ, and yaw angle ψ. The orientation matrix M = R(θ)*T(φ)*P(ψ) of each device relative to the reference coordinate system is calculated, where R, T, and P represent rotation operations around the three axes. By calculating the orientation matrix M = R(θ)*T(φ)*P(ψ) of each device relative to the reference coordinate system, the device's posture information at the time of capture can be accurately restored. This step is critical for subsequent data analysis because it ensures that images and photos are presented in the correct geospatial context, supporting detailed analysis of complex structures and environments.

[0067] Accurate geolocation coordinates and device posture data provide a solid foundation for subsequent data processing and visualization. This data can help construct georeferenced panoramic video datasets, supporting the application of virtual reality technology and providing users with an immersive inspection environment experience. It also facilitates advanced applications such as map updates, space planning, and asset management.

[0068] Through the careful deployment and parameter setting of panoramic camera equipment, the entire system has significantly improved inspection efficiency, data quality, and user experience. This has reduced the need for manual operation, lowered the risk of human error, and improved the reliability and repeatability of inspection results.

[0069] Step 2: Using the panoramic camera device to automatically patrol along a preset path or a path dynamically generated by an intelligent planning algorithm, and collecting panoramic image data in real time; further comprising:

[0070] The inspection starting point and target destination are set, and at least one inspection route is planned based on the construction site map. This route consists of a series of coordinate points (X_i, Y_i, Z_i), where i represents the i-th coordinate point on the path. This map-based route planning ensures comprehensive inspection coverage, avoids missing important areas, and is flexible enough to handle complex terrain and structures. The panoramic camera equipment moves along the defined route, ensuring the automation and standardization of the inspection process, reducing the need for manual intervention and improving inspection efficiency.

[0071] The panoramic camera is started to move along the defined route, while simultaneously capturing images at a fixed frame rate F. For each coordinate point (X_i, Y_i, Z_i), the time required for the device to reach that point is calculated as T_i = D_i / V, where D_i is the distance from the previous point to the current point and V is the device's movement speed. This step ensures that the device can accurately control movement speed and time during the inspection process, ensuring the consistency and time synchronization of image capture. Each time a new coordinate point is reached, the current position coordinates (X_i, Y_i, Z_i) and timestamp T_i are recorded, providing a precise time and space reference for subsequent data processing, enhancing data reliability and availability.

[0072] During movement, as the device approaches the next coordinate point, it adjusts its attitude parameters (θ_i, φ_i, ψ_i) to align the center of its field of view in the direction of the next coordinate point. This attitude adjustment is based on the formula M_i = R(θ_i)*T(φ_i)*P(ψ_i), where R, T, and P represent rotations around the three axes. This ensures the device always faces the intended direction, improving the quality and accuracy of image capture. This precise attitude adjustment allows the device to capture the most relevant and valuable images at each coordinate point, reducing unnecessary retakes and improving inspection efficiency.

[0073] After completing posture adjustment, the device immediately begins recording panoramic image data for that location and saves the data along with the current position coordinates (X_i, Y_i, Z_i), timestamp T_i, and posture parameters (θ_i, φ_i, ψ_i). The device then continues moving to the next coordinate point until the inspection route is complete. This detailed metadata recording method ensures that each image frame has complete information support, facilitating subsequent data analysis and visualization. The device captures images at a fixed frame rate F, ensuring image continuity and clarity, while also ensuring data integrity throughout the inspection process.

[0074] The entire inspection process is highly automated, from route planning to image acquisition to data recording. This reduces the need for manual operation, lowers the risk of human error, and improves the reliability of inspection results. Detailed data recording and precise spatial positioning provide users with a comprehensive and intuitive inspection report, enabling rapid analysis and decision-making, enhancing the user experience and system practicality.

[0075] Step 3: During the automatic inspection process, use a mobile terminal device to take photos of key areas and problem areas on the construction site, and automatically attach the photos to the inspection track plan of the corresponding locations; further including:

[0076] When the inspection equipment reaches a predetermined critical or problematic area, it triggers a capture command on the mobile terminal device, determining the coordinates (X_j, Y_j, Z_j) of each capture location P_j. This process ensures focus on important locations, reduces unnecessary capture, and improves work efficiency. At the designated location P_j, the mobile terminal device captures a high-resolution photo I_j at the highest resolution, providing detailed local details that facilitate subsequent analysis and decision-making.

[0077] At the specified position P_j, the mobile terminal device captures a high-definition photo I_j at the highest resolution, and at the same time records the timestamp T_j of the shooting moment and the device posture parameters (θ_j, φ_j, ψ_j). For each photo I_j, the spatial distance D_j from the most recent panoramic image acquisition point is calculated as D_j = sqrt((X_j-X_i)^2+(Y_j-Y_i)^2+(Z_j-Z_i)^2), where (X_i, Y_i, Z_i) are the coordinates of the most recent panoramic image acquisition point; this step ensures that the spatial relationship between the photo and the panoramic image is clear and unambiguous.

[0078] Based on the timestamp T_j and spatial distance D_j, the high-definition photo I_j is associated with the corresponding inspection track location, and the weight W_j is calculated using the formula W_j = exp(-α*D_j^2), where α is an adjustment parameter. This weight calculation method takes into account time and space factors, allowing photos to be reasonably sorted and displayed according to their relevance, improving the efficiency of users browsing and finding specific information.

[0079] High-definition photos I_j and their associated weights W_j, timestamps T_j, coordinates (X_j, Y_j, Z_j), and pose parameters (θ_j, φ_j, ψ_j) are integrated into a dataset of inspection trajectory maps. This comprehensive dataset not only contains rich information but also allows users to browse photos by clicking or dragging on an interactive interface, visually visualizing the relationship between photos and actual locations. This dataset allows users to browse photos by clicking or dragging on an interactive interface, visually visualizing the relationship between photos and actual locations. Through this interactive interface, users can easily explore different perspectives and areas, obtain more detailed information, and thus gain a more comprehensive understanding of the inspection situation, supporting rapid analysis and decision-making.

[0080] This approach helps engineers and managers quickly identify and respond to complex issues, improving inspection quality and decision-making efficiency. By capturing and accurately capturing high-definition photos of key and problematic areas, it provides strong data support, facilitates data-driven decision-making, and enhances project management and monitoring capabilities.

[0081] Step 4: Based on the spatial location information of the photos and panoramic images, a panoramic video dataset with geographical reference is formed by linking the spatial location with the plan view. This further includes:

[0082] Collect all high-definition photos I_j and their corresponding spatial position coordinates (X_j, Y_j, Z_j), timestamps T_j and posture parameters (θ_j, φ_j, ψ_j). At the same time, obtain the continuous image frames F_i recorded by the panoramic camera device in the same time period and their corresponding geographic location coordinates (X_i, Y_i, Z_i) and timestamps T_i; this step ensures that all relevant data are fully collected, providing a rich information basis for subsequent processing.

[0083] For each high-definition photo I_j, find the panoramic image frame F_i closest to its timestamp T_j, calculate the time difference ΔT = |T_j-T_i| between the two to ensure the time synchronization between the photo and the image frame, and select the photo and image frame pairs that meet the ΔT<τ threshold, where τ is a preset time tolerance; this time synchronization mechanism ensures the matching of images and photos in the time dimension, improving the consistency and reliability of the data.

[0084] Based on the selected photo and image frame pair, use the spatial position coordinates (X_j, Y_j, Z_j) and (X_i, Y_i, Z_i) to calculate the spatial distance between the two:

[0085] D_ij = sqrt((X_j - X_i)^2 + (Y_j - Y_i)^2 + (Z_j - Z_i)^2). The association weight W_ij is calculated using the formula W_ij = exp(-β*D_ij^2), where β is a tuning parameter. This step ensures accurate spatial association between images and photos, enhancing the spatial consistency of the data. This weight calculation method comprehensively considers the influence of spatial distance, making the association between different images and photos more reasonable, which facilitates prioritization during subsequent data analysis and presentation.

[0086] The identified photo and image frame pairs, along with their weights W_ij, timestamps T_j / T_i, coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and pose parameters (θ_j, φ_j, ψ_j), are integrated into a unified dataset. This integration not only includes rich metadata but also ensures data integrity and traceability, facilitating subsequent query and analysis. The resulting georeferenced panoramic video dataset accurately reflects the location and time of each image and photo at the actual construction site, supporting advanced spatial analysis and visualization applications such as virtual reality inspection and 3D modeling.

[0087] Through precise time synchronization and spatial correlation, as well as rational weight calculation, data accuracy and reliability are greatly improved, the need for manual intervention is reduced, and the risk of data errors is mitigated. Users can browse this data through an interactive interface, viewing the specific location and time relationships of high-definition photos and panoramic images, enabling rapid analysis and decision-making. This intuitive data presentation greatly enhances the user experience and system practicality, promoting efficient data-based management and decision-making.

[0088] Step 5: Digitally capture the acquired panoramic video data set, convert the original images and photos into digital format, and simultaneously collect the movement trajectory and timestamp information of the inspection personnel; further including:

[0089] Receive a panoramic video dataset with georeference. For each pair of high-definition photo I_j and panoramic image frame F_i, extract the binary data streams B_j and B_i of the original image and photo, and at the same time, obtain the associated timestamps T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i) and posture parameters (θ_j, φ_j, ψ_j); this step ensures that all relevant information is fully collected, providing a solid foundation for subsequent processing.

[0090] The binary data streams B_j and B_i are converted into compressed formats C_j and C_i using the formula C_x = Compress(B_x, r), where x represents j or i, Compress indicates the compression operation, and r is the compression ratio. This approach preserves the original precision of the timestamps T_j / T_i, coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and attitude parameters (θ_j, φ_j, ψ_j). During the data compression process, the original precision of the timestamps T_j / T_i, coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and attitude parameters (θ_j, φ_j, ψ_j) is preserved. This approach ensures data consistency and reliability before and after compression, avoiding information loss due to compression. The compressed data not only reduces storage requirements but also optimizes transmission efficiency, making large-scale data management and transmission more convenient.

[0091] During data processing, the inspector's movement trajectory is synchronously recorded. Using a positioning device integrated into the inspection equipment, the inspector's position coordinates (X_p, Y_p, Z_p) and timestamp T_p are regularly sampled. The time interval ΔT_p = T_p - T_(p-1) between two consecutive samplings is calculated to ensure trajectory continuity and time synchronization. This step ensures that the inspector's movement trajectory accurately reflects the actual inspection process. For each timestamp T_p, the corresponding inspector's posture parameters (θ_p, φ_p, ψ_p) are recorded to form a complete movement trajectory record. This data not only enhances the richness of the trajectory information but also provides support for subsequent spatial analysis and behavioral pattern recognition.

[0092] The compressed image data C_j / C_i, timestamps T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and posture parameters (θ_j, φ_j, ψ_j) are combined with the inspector's movement trajectory (X_p, Y_p, Z_p), timestamp T_p, and posture parameters (θ_p, φ_p, ψ_p) into a single dataset. This comprehensive dataset not only contains rich metadata but also ensures the relevance and consistency of all data.

[0093] This comprehensive dataset provides powerful support for subsequent data analysis, visualization, and decision-making. Users can browse this data through an interactive interface, viewing images and photos and their relationship to the movement paths of inspectors, gaining a more comprehensive understanding of the inspection situation and enabling rapid analysis and decision-making.

[0094] Step 6: Utilize digitally collected data and movement trajectory information and transmit it to the central processing system in real time via the AI ​​workstation; further including:

[0095] Extract the compressed image data C_j / C_i, timestamp T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), posture parameters (θ_j, φ_j, ψ_j), and the inspection personnel's movement trajectory (X_p, Y_p, Z_p), timestamp T_p and posture parameters (θ_p, φ_p, ψ_p) from the integrated data set; this step ensures that all relevant data are fully extracted, providing a complete data packet for subsequent transmission.

[0096] The data is packaged into transmission packets D. Each transmission packet D contains one or more image frames and their associated information, and is accompanied by a check code Checksum(D) = Hash(Summary(D)), where Summary(D) represents the summary of the transmission packet content and Hash is a hash function. This packaging method not only improves the security of data transmission, but also ensures data integrity through the check code.

[0097] The generated transmission packet D is sent to the central processing system in real time. After each transmission, the sending time and the reception confirmation time are recorded, and the transmission delay ΔT_transmit = T_receive - T_send is calculated. For each transmission packet D, the corresponding transmission delay ΔT_transmit is recorded to form a transmission log. This step ensures the timeliness and reliability of data transmission and provides a detailed transmission log for subsequent analysis. T_send is the timestamp of the packet transmission from the sender, and T_receive is the timestamp of the packet's arrival and successful reception at the receiver. By recording the transmission delay ΔT_transmit, network status can be monitored, potential transmission issues can be identified, and transmission paths can be optimized to ensure that data reaches the central processing system in the shortest possible time.

[0098] After the central processing system receives transmission packet D, it unpacks it and verifies the checksum (D). Then, based on the timestamps T_j / T_i and T_p, the image data C_j / C_i and the inspector's movement trajectory (X_p, Y_p, Z_p) are synchronized and integrated into the database. The timestamp T_x is adjusted using the formula Sync(T_x, ΔT_transmit) = Adjust(T_x + ΔT_transmit), where x represents j / i or p and T_x represents the original timestamp. This verification mechanism enhances the security and reliability of data transmission and reduces the risk of data loss or corruption. This step ensures the temporal consistency of all data, supporting precise time synchronization and event correlation.

[0099] The entire transmission process, from data extraction and packaging to transmission and verification, is highly automated, reducing the need for manual intervention and improving processing efficiency and accuracy. Through real-time transmission and synchronous integration, the central processing system can promptly access the latest inspection data, supporting rapid data analysis and decision-making. Users can browse this data through an interactive interface, viewing images and photos and their relationship to inspector movement trajectories, providing a more comprehensive understanding of inspection situations. Checksum verification and transmission delay monitoring ensure the security and reliability of data transmission, reduce the risk of data errors and loss, and enhance system stability and reliability.

[0100] Step 7: Combining the digitally collected data and the movement trajectory information, displaying the inspection trajectory in the form of lines or marks on the floor plan; further including:

[0101] All data transmitted and integrated from the central processing system is obtained. Based on the inspector's movement trajectory (X_p, Y_p, Z_p), an inspection route is drawn as lines on the project site plan. For each trajectory point P_p, the formula P_line = Line(P_(p-1), P_p) is used, where Line represents the line segment connecting two consecutive trajectory points. This forms a coherent inspection route map. At the same time, the timestamp T_p of each trajectory point is recorded; this step ensures the accurate drawing of the inspection route and provides clear visual guidance. Recording the timestamp T_p of each trajectory point also provides the basis for subsequent time synchronization and event correlation.

[0102] According to the inspection route map, the panoramic image frame F_i and the high-definition photo I_j are associated with the corresponding trajectory point P_p using the spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i) and timestamp T_j / T_i; this association method ensures that the image and photo can be displayed in the correct position, enhancing the spatial consistency of the data.

[0103] The closest trajectory point P_p is found using the formula D_associate = min(Distance(P_p,(X_x,Y_x,Z_x))), where x represents j or i and Distance represents the distance calculation between two points. This method ensures that images and photos can be accurately attached to the nearest patrol path point, improving the accuracy of the display.

[0104] Markers are added to the corresponding locations on the patrol route map to display the panoramic image frame F_i and high-definition photo I_j. For each marker M_x, the timestamp T_x and posture parameters (θ_x, φ_x, ψ_x) of the shooting time are displayed, and the user is provided with interactive browsing functions. This step allows users to easily view detailed information of specific locations and supports quick analysis and decision-making.

[0105] The formula Overlay(M_x,P_line)=Merge(M_x,Line(P_(p-1),P_p)) combines the markers with the inspection path segments to form a comprehensive view. This comprehensive view not only shows the inspection path but also provides the specific locations of images and photos, enhancing user understanding and operational convenience.

[0106] By displaying inspection tracks as lines or markers on a floor plan, users can intuitively see inspection paths and photos of key areas, enabling quick analysis and decision-making. This intuitive display significantly enhances the user experience and system practicality. The comprehensive view allows users to browse photos by clicking or dragging on the interactive interface and view the relationship between photos and actual locations, supporting efficient data query and analysis. Users can quickly locate problem areas and conduct detailed inspections, improving work efficiency.

[0107] Step 8: Fusing the panoramic image and high-definition photos with the inspection route data for display. Based on the fusion display, the user is enabled to explore different perspectives and areas through an interactive interface. This further includes:

[0108] A map view based on the integration of patrol route map and image data is constructed. In this map view, each patrol trajectory point P_p and its associated panoramic image frame F_i and high-definition photo I_j are marked. For each marker M_x, its visible range V_x is defined as a sphere with a radius R and its center at (X_x, Y_x, Z_x), expressed as V_x = Sphere((X_x, Y_x, Z_x), R). This setting allows users to intuitively understand the coverage of each marker, enhancing the information richness of the map view.

[0109] Based on the established map view, zoom and pan operations are provided to allow users to focus on areas of interest. When the user selects a specific marker M_x, the system calculates the distance D_view between the marker and the user's current viewpoint and uses the formula ZoomLevel = f(D_view), where f is a mapping function to determine the appropriate zoom level. This approach ensures that users can obtain the most appropriate visual experience during browsing, satisfying both global overview and detailed viewing.

[0110] According to user operations, the multi-view browsing mode is activated. For the selected marker M_x, its posture parameters (θ_x, φ_x, ψ_x) are used to construct the transformation matrix T_x to adjust the angle of image display; this allows users to freely change the perspective and observe the same location from different angles, enhancing the user's immersion and the comprehensiveness of information acquisition.

[0111] Introducing an instant information prompt box. When a user hovers over or clicks a marker M_x, relevant information is immediately displayed, including but not limited to the timestamp T_x, location coordinates (X_x, Y_x, Z_x), and a brief description. This instant feedback mechanism greatly improves the speed and accuracy of user information acquisition, supporting rapid decision-making and problem analysis.

[0112] Through these technologies, users can more intuitively explore different perspectives and locations within the inspection area through the interactive interface. This interactive approach not only increases user engagement but also promotes a deeper understanding of inspection data.

[0113] Users can efficiently query and analyze data at specific locations through an interactive interface, enabling faster problem discovery and resolution, thereby improving work efficiency. By integrating displays and interactive interfaces, managers can remotely monitor inspection progress, stay informed of on-site conditions, and make necessary adjustments, thereby enhancing management efficiency. The system's multi-view browsing mode and detailed information prompts are also suitable for training and education scenarios, helping new employees quickly get started and understand workflows and technical requirements.

[0114] On the other hand, the present invention proposes a system for generating engineering inspection tracks based on panoramic images, such as Figure 2 Shown, including:

[0115] A panoramic camera deployment module is configured to deploy at least one panoramic camera at the construction site. The panoramic camera is equipped with multiple cameras for capturing 360-degree images of the environment without blind spots. The panoramic camera is used to perform automatic inspections along a preset path or a path dynamically generated by an intelligent planning algorithm, and to collect panoramic image data in real time.

[0116] The high-definition photo shooting and attachment module is used to capture photos of key and problematic areas of the construction site using mobile devices during the automated inspection process. These photos are then automatically attached to the inspection track plan at the corresponding locations. Based on the spatial location information of the photos and panoramic images, a panoramic video dataset with geo-referenced information is generated through spatial location and plan attachment technology.

[0117] The digital acquisition and synchronous recording module is used to digitally acquire the acquired panoramic video data set, convert the original images and photos into digital format, and simultaneously collect the movement trajectory and timestamp information of the inspection personnel. The digitally acquired data and movement trajectory information are transmitted to the central processing system in real time through the AI ​​workstation;

[0118] The inspection trajectory visualization display module is used to combine digitally collected data and movement trajectory information to display the inspection trajectory in the form of lines or marks on the floor plan. It integrates panoramic images and high-definition photos with the inspection path data for display. Based on the integrated display, the interactive interface enables users to explore different perspectives and areas.

[0119] In addition, the above-mentioned panoramic camera equipment deployment module, high-definition photo shooting and mounting module, and digital acquisition and synchronous recording module are also used to implement the other steps of the above-mentioned method for generating an engineering inspection trajectory based on panoramic images when executed, as follows:

[0120] Automated inspections with panoramic cameras: Automated inspections with panoramic cameras are a key first step in intelligent engineering inspections. Panoramic cameras are equipped with multiple high-resolution cameras, capturing a 360-degree, comprehensive view of the environment. These cameras use spherical fisheye lenses to provide a wide viewing angle. Inspection routes can be preset or dynamically generated using intelligent planning algorithms to adapt to varying inspection needs and environmental changes.

[0121] High-definition photo acquisition module: When inspecting key areas and problem areas at the construction site, use the mobile phone app to take photos and save them in high definition. The photos are automatically attached to the inspection track plan of the current corresponding location to obtain more detailed local information and accurately locate the problem location.

[0122] Spatial Position and Floor Plan Linking: Spatial position and floor plan linking is the process of combining the spatial position information of panoramic images with a two-dimensional floor plan. This creates a spatially referenced panoramic video. Captured panoramic video data is combined with precise spatial position information to form a georeferenced panoramic video dataset. Using virtual reality technology, video data is integrated with map data to achieve a visual display of spatial data. Spatial position linking allows users to intuitively understand the correspondence between video content and actual locations, improving the readability and usability of the data. This process is achieved through virtual reality mapping technology, which enables the restoration, processing, and analysis of geospatial data. In this step, the panoramic image's position, orientation, and other spatial information are used to precisely locate the image on the floor plan. This linking not only provides geographic context for the image but can also be used for map updates, spatial planning, and asset management. Captured panoramic video data is linked to the corresponding spatial position information.

[0123] Digital acquisition: Digital acquisition is the process of converting the raw image and photo data captured by the panoramic camera into a digital format. This step involves image compression, encoding, and storage to facilitate subsequent analysis and processing. Digital acquisition includes not only image data, but may also include metadata related to the image, such as shooting time, camera position, environmental parameters, etc. The acquisition and processing of this data requires efficient algorithms and powerful computing power to ensure the integrity and availability of the data. Synchronously collect the movement trajectory and timestamp information of the inspection personnel. Use virtual reality positioning technology to record the changes in the movement trajectory of the inspection personnel in real time. Combined with timestamp information, ensure the synchronization and accuracy of the data. Through the AI ​​workstation, the collected data is transmitted to the central processing system in real time to reduce the risk of data loss. Encryption technology is used to protect the security of data during transmission and prevent unauthorized access.

[0124] Fusion of panoramic images and inspection tracks: The fusion of panoramic images and inspection tracks combines panoramic images and high-definition photos with inspection route data to provide comprehensive and intuitive inspection information. On a floor plan, the inspection route can be displayed as lines or markers, overlaid with the panoramic images and photos, allowing users to clearly see every key point on the inspection route and related visual information. This fusion display is achieved through an interactive interface, allowing users to explore different perspectives and areas and obtain more detailed information by clicking or dragging.

[0125] The effect is as follows:

[0126] Automation: Automated inspections using panoramic cameras not only reduce the need for manual labor but also ensure comprehensive and consistent inspection coverage through pre-set intelligent path planning. This automated process reduces human error and improves inspection reliability and repeatability.

[0127] Digital collection: Digital collection not only ensures data accuracy and storability, but also optimizes data transmission and storage efficiency through advanced data compression and encoding technologies. Furthermore, metadata collection provides rich contextual information for data analysis, enhancing data usability and the depth of analysis.

[0128] Scene Restoration: Scene Restoration technology processes video analysis into high-definition images for each frame, providing users with an immersive inspection environment experience. This technology not only enhances users' understanding of the inspection environment but also enables detailed analysis of complex structures and environments, providing strong visual support for engineering decision-making.

[0129] Precise alignment of spatial locations with the floor plan: Utilizing advanced virtual reality mapping technology, the precise alignment of spatial locations with the floor plan provides a precise geographic reference framework. This not only helps identify specific locations and paths during inspections, but also supports geospatial analysis and tracking.

[0130] Fusion of panoramic images and inspection tracks: The fusion of panoramic images and inspection tracks provides an intuitive overview of inspection information through an integrated visualization platform. This fusion not only facilitates rapid analysis and decision-making, but also supports the rapid identification and response of complex issues.

[0131] The intelligent engineering inspection method based on panoramic images, high-definition photos and track attachment of the present invention realizes the full automation and digitization of engineering inspections through a series of innovative technical steps. Automated panoramic camera inspections improve the efficiency and consistency of inspections, while digital acquisition ensures the accuracy and efficient management of data. Scene restoration technology provides users with an in-depth understanding of the inspection environment, while the precise attachment of spatial positions provides a solid foundation for geospatial analysis. The fusion display of panoramic images and inspection tracks not only enhances the visualization of information, but also facilitates rapid decision-making. Overall, this method provides an innovative, efficient and user-friendly solution for the field of intelligent engineering inspections, with broad application prospects and far-reaching industry impact. With the continuous advancement of technology, this method is expected to play a greater role in future engineering management and monitoring, and promote engineering inspections to a higher level of intelligence and automation.

[0132] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating engineering inspection trajectories based on panoramic images, characterized in that: The following steps are involved: Deploy at least one panoramic camera device at the construction site, the panoramic camera device equipped with multiple cameras for capturing 360-degree environmental images without blind spots, and use the panoramic camera device to automatically patrol along a preset path or a path dynamically generated by an intelligent planning algorithm, and collect panoramic image data in real time; During the automatic inspection process, mobile terminal devices are used to take photos of key and problem areas at the construction site, and the photos are automatically attached to the inspection trajectory plan of the corresponding positions. Specifically, when the inspection device reaches the preset key or problem area, the shooting instruction of the mobile terminal device is triggered, and the coordinates (X_j, Y_j, Z_j) of each shooting position P_j are determined; at the specified position P_j, the mobile terminal device captures a high-definition photo I_j with the highest resolution, and at the same time, records the timestamp T_j of the shooting moment and the device posture parameters (θ_j, φ_j, ψ_j). For each photo I_j, its spatial distance D_j to the most recent panoramic image acquisition point is calculated as sqrt((X_j-X_i)^2+(Y _j-Y_i)^2+(Z_j-Z_i)^2), where (X_i, Y_i, Z_i) are the coordinates of the most recent panoramic image acquisition point; based on the timestamp T_j and spatial distance D_j, the high-definition photo I_j is associated with the corresponding inspection track location, and the weight W_j is calculated using the formula W_j=exp(-α*D_j^2), where α is an adjustment parameter; the high-definition photo I_j and its associated weight W_j, timestamp T_j, coordinates (X_j, Y_j, Z_j) and posture parameters (θ_j, φ_j, ψ_j) are integrated into the inspection track plan map dataset. This dataset allows users to browse photos by clicking or dragging the interactive interface and intuitively see the relationship between the photo and the actual location; Based on the spatial location information of photos and panoramic images, a panoramic video dataset with geographical reference is formed through spatial location and plan view connection technology; The obtained panoramic video data set is digitally collected, and the original images and photos are converted into digital format. At the same time, the movement trajectory and timestamp information of the inspection personnel are simultaneously collected. The digitally collected data and movement trajectory information are transmitted to the central processing system in real time through the AI ​​workstation; Combining digitally collected data and movement trajectory information, the inspection trajectory is displayed in the form of lines or marks on the floor plan. Panoramic images and high-definition photos are integrated with the inspection path data for display. Based on the integrated display, users can explore different perspectives and areas through an interactive interface.

2. The method for generating an engineering inspection trajectory based on panoramic images according to claim 1, characterized in that: The step of deploying at least one panoramic camera device at the construction site includes: Determine the locations of multiple observation points at the project site. Each observation point corresponds to one or more panoramic camera installation locations. Install panoramic cameras at the selected locations. For any two adjacent cameras, their field of view overlap ratio is at least P%, where P represents the ratio of the intersection of the fields of view between the two devices to the total field of view of a single device. Based on the adjusted device settings, the panoramic camera device is started to capture images. The device shoots at a fixed frame rate F, where F represents the number of images captured per second. At the same time, the geographic coordinates (X, Y, Z) of each device are recorded for spatial positioning of subsequent images. During the image acquisition process, the device posture data corresponding to each frame of image is synchronously recorded, including the rotation angle θ, tilt angle φ, and yaw angle ψ, and the orientation matrix M = R(θ)*T(φ)*P(ψ) of each device relative to the reference coordinate system is calculated, where R, T, and P represent the rotation operations around the three axes, respectively.

3. The method for generating an engineering inspection trajectory based on panoramic images according to claim 2, characterized in that: The method of using the panoramic camera device to automatically inspect along a preset path or a path dynamically generated by an intelligent planning algorithm and collecting panoramic image data in real time includes: Set the inspection starting point and target end point, and plan at least one inspection route based on the map information of the construction site. This route consists of consecutive coordinate points (X_i, Y_i, Z_i), where i represents the i-th coordinate point on the path; Start the panoramic camera device and move it along the defined route. At the same time, begin capturing images at a fixed frame rate F. For each coordinate point (X_i, Y_i, Z_i), calculate the time required for the device to reach that point T_i = D_i / V, where D_i is the distance from the previous point to the current point, and V is the device's movement speed. During the movement, when the device approaches the next coordinate point, the device's posture parameters (θ_i, φ_i, ψ_i) are adjusted so that the center of the device's field of view is aligned with the direction of the next coordinate point. The posture adjustment is based on the formula M_i = R(θ_i)*T(φ_i)*P(ψ_i), where R, T, and P represent the operation of rotation around the three axes respectively; After completing the posture adjustment, it immediately starts recording the panoramic image data of the location and saves the data together with the current position coordinates (X_i, Y_i, Z_i), timestamp T_i and posture parameters (θ_i, φ_i, ψ_i). Then, it continues to move to the next coordinate point until the inspection route ends.

4. The method for generating an engineering inspection trajectory based on panoramic images according to claim 3, characterized in that: The forming of a panoramic video dataset with a geographical reference includes: Collect all high-definition photos I_j and their corresponding spatial position coordinates (X_j, Y_j, Z_j), timestamps T_j and posture parameters (θ_j, φ_j, ψ_j). At the same time, obtain the continuous image frames F_i recorded by the panoramic camera device in the same time period and their corresponding geographical location coordinates (X_i, Y_i, Z_i) and timestamps T_i; For each high-definition photo I_j, find the panoramic image frame F_i closest to its timestamp T_j, calculate the time difference ΔT = |T_j - T_i| between the two to ensure the time synchronization between the photo and the image frame, and select the photo and image frame pairs that meet the threshold ΔT < τ, where τ is a preset time tolerance; Based on the selected photo and image frame pairs, use the spatial position coordinates (X_j, Y_j, Z_j) and (X_i, Y_i, Z_i) to calculate the spatial distance between them: D_ij = sqrt((X_j - X_i)^2 + (Y_j - Y_i)^2 + (Z_j - Z_i)^2). Calculate the association weight Wij using the formula Wij = exp(-β*D_ij^2), where β is a tuning parameter. The identified photo and image frame pairs, along with their weights Wij, timestamps Tj / Ti, coordinates (Xj, Yj, Zj) / (Xi, Yi, Zi) and pose parameters (θj, φj, ψj), are integrated into a unified dataset.

5. The method for generating an engineering inspection trajectory based on panoramic images according to claim 4, characterized in that: The obtained panoramic video data set is digitally collected, the original images and photos are converted into digital format, and the movement trajectory and time stamp information of the inspection personnel are simultaneously collected, including: Receive a georeferenced panoramic video dataset. For each pair of high-definition photo I_j and panoramic image frame F_i, extract the binary data streams B_j and B_i of the original image and photo, and obtain the associated timestamps T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and posture parameters (θ_j, φ_j, ψ_j). Convert the binary data streams B_j and B_i into compressed formats C_j and C_i using the formula C_x = Compress(B_x,r), where x represents j or i, Compress indicates the compression operation, and r is the compression ratio. At the same time, the original precision of the timestamps T_j / T_i, coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), and attitude parameters (θ_j, φ_j, ψ_j) is preserved. During data processing, the patrol personnel's movement trajectory is recorded synchronously. The positioning device integrated in the patrol equipment regularly samples the patrol personnel's position coordinates (X_p, Y_p, Z_p) and timestamp T_p. The time interval ΔT_p = T_p - T_(p-1) between two adjacent samples is calculated to ensure trajectory continuity and time synchronization. For each timestamp T_p, the corresponding patrol personnel posture parameters (θ_p, φ_p, ψ_p) are recorded to form a complete movement trajectory record. The compressed image data C_j / C_i, timestamp T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), posture parameters (θ_j, φ_j, ψ_j) and the inspection personnel's movement trajectory (X_p, Y_p, Z_p), timestamp T_p and posture parameters (θ_p, φ_p, ψ_p) are integrated into a data set.

6. The method for generating an engineering inspection trajectory based on panoramic images according to claim 5, characterized in that: The digitally collected data and movement trajectory information are transmitted to the central processing system in real time through the AI ​​workstation, including: Extract the compressed image data C_j / C_i, timestamp T_j / T_i, spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i), posture parameters (θ_j, φ_j, ψ_j), and the inspection personnel's movement trajectory (X_p, Y_p, Z_p), timestamp T_p and posture parameters (θ_p, φ_p, ψ_p) from the integrated data set; Pack the data into transmission packets D. Each transmission packet D contains one or more image frames and their associated information, and is accompanied by a checksum (D) = Hash (Summary (D)), where Summary (D) represents the summary of the transmission packet content and Hash is a hash function. The generated transmission packet D is sent to the central processing system in real time. After each transmission, the sending time and the receiving confirmation time are recorded, and the transmission delay ΔT_transmit = T_receive - T_send is calculated. For each transmission packet D, the corresponding transmission delay ΔT_transmit is recorded to form a transmission log. T_send is the timestamp of the data packet transmission from the sender, and T_receive is the timestamp of the data packet arriving at the receiver and being successfully received. After the central processing system receives the transmission packet D, it unpacks it and verifies the checksum (D). Then, based on the timestamps T_j / T_i and T_p, the image data C_j / C_i and the inspection personnel's movement trajectory (X_p, Y_p, Z_p) are synchronized and integrated into the database. The timestamp T_x is adjusted using the formula Sync(T_x,ΔT_transmit) = Adjust(T_x+ΔT_transmit), where x represents j / i or p, and T_x represents the original timestamp.

7. The method for generating an engineering inspection trajectory based on panoramic images according to claim 6, characterized in that: The fusion display of panoramic images and high-definition photos with inspection route data includes: All data transmitted and integrated from the central processing system is obtained. Based on the inspection personnel's movement trajectory (X_p, Y_p, Z_p), the inspection route is drawn as a line on the project site plan. For each trajectory point P_p, the formula P_line = Line(P_(p-1), P_p) is used, where Line represents the line segment connecting two consecutive trajectory points, forming a coherent inspection route map. At the same time, the timestamp T_p of each trajectory point is recorded; According to the inspection route map, the panoramic image frame F_i and the high-definition photo I_j are associated with the corresponding trajectory point P_p using the spatial position coordinates (X_j, Y_j, Z_j) / (X_i, Y_i, Z_i) and timestamp T_j / T_i; Find the closest trajectory point P_p using the formula D_associate=min(Distance(P_p,(X_x,Y_x,Z_x))), where x represents j or i, and Distance represents the distance between two points. Markers are added to the corresponding positions on the inspection path map to display the panoramic image frame F_i and the high-definition photo I_j. For each marker M_x, the timestamp T_x of the shooting time and the posture parameters (θ_x, φ_x, ψ_x) are displayed, providing users with interactive browsing functions. The formula Overlay(M_x, P_line) = Merge(M_x, Line(P_(p-1), P_p)) is used to merge the marker with the inspection path segment to form a comprehensive view.

8. The method for generating an engineering inspection trajectory based on panoramic images according to claim 7, characterized in that: The above-mentioned integrated display enables users to explore different perspectives and areas through an interactive interface, including: A map view is constructed based on the integration of the inspection route map and image data. In this map view, each inspection trajectory point P_p and its associated panoramic image frame F_i and high-definition photo I_j are marked. For each marker M_x, its visible range V_x is defined as a sphere with a radius R and a center at (X_x, Y_x, Z_x). The formula is expressed as V_x = Sphere((X_x, Y_x, Z_x), R). Based on the established map view, zoom and pan operations are provided to allow users to focus on the area of ​​interest. When the user selects a specific marker M_x, the system will calculate the distance D_view between the marker and the user's current viewpoint and use the formula ZoomLevel = f(D_view), where f is a mapping function, to determine the appropriate zoom level; According to user operation, the multi-view browsing mode is activated. For the selected marker M_x, its posture parameters (θ_x, φ_x, ψ_x) are used to construct the transformation matrix T_x to adjust the angle of image display; Introducing an instant message prompt box. When the user hovers over or clicks a marker M_x, relevant information is immediately displayed, including but not limited to the timestamp T_x, location coordinates (X_x, Y_x, Z_x), and a brief description.

9. A system for generating engineering inspection tracks based on panoramic images, characterized in that: include: A panoramic camera deployment module is configured to deploy at least one panoramic camera at the construction site. The panoramic camera is equipped with multiple cameras for capturing 360-degree images of the environment without blind spots. The panoramic camera is used to perform automatic inspections along a preset path or a path dynamically generated by an intelligent planning algorithm, and to collect panoramic image data in real time. The high-definition photo shooting and attachment module is used to use a mobile terminal device to take photos of key parts and problem parts of the engineering site during the automatic inspection process, and automatically attach the photos to the inspection trajectory plan of the corresponding position, including: when the inspection device reaches the preset key or problem part, triggering the shooting instruction of the mobile terminal device, determining the coordinates (X_j, Y_j, Z_j) of each shooting position P_j; at the specified position P_j, the mobile terminal device captures a high-definition photo I_j with the highest resolution, and at the same time, records the timestamp T_j of the shooting moment and the device posture parameters (θ_j, φ_j, ψ_j); for each photo I_j, calculate its spatial distance D_j to the most recent panoramic image acquisition point = sqrt((X_j-X_i)^2+(Y_j-Y_i)^2+(Z_j-Z_i )^2), where (X_i, Y_i, Z_i) are the coordinates of the most recent panoramic image acquisition point; based on the timestamp T_j and spatial distance D_j, the high-definition photo I_j is associated with the corresponding inspection track position, and the weight W_j is calculated using the formula W_j=exp(-α*D_j^2), where α is an adjustment parameter; the high-definition photo I_j and its associated weight W_j, timestamp T_j, coordinates (X_j, Y_j, Z_j) and posture parameters (θ_j, φ_j, ψ_j) are integrated into the inspection track plan map dataset. This dataset allows users to browse photos by clicking or dragging the interactive interface and intuitively see the relationship between the photos and the actual location; based on the spatial location information of the photos and panoramic images, a panoramic video dataset with geographical reference is formed through the spatial location and plan map attachment technology; The digital acquisition and synchronous recording module is used to digitally acquire the acquired panoramic video data set, convert the original images and photos into digital format, and simultaneously collect the movement trajectory and timestamp information of the inspection personnel. The digitally acquired data and movement trajectory information are transmitted to the central processing system in real time through the AI ​​workstation; The inspection trajectory visualization display module is used to combine digitally collected data and movement trajectory information to display the inspection trajectory in the form of lines or marks on the floor plan. It integrates panoramic images and high-definition photos with the inspection path data for display. Based on the integrated display, the interactive interface enables users to explore different perspectives and areas.

Images