A method and device for designing a full-scene coverage of a track of an inspection robot, and a medium

Abstract

This invention relates to the field of intelligent inspection track design, specifically a method, device, and medium for designing full-scene coverage of inspection robot tracks. The method involves: 3D modeling of the steel structure space of the building to be inspected to identify the target inspection area; establishing a detection coverage radius model to determine the maximum coverage radius; determining and correcting the spacing between adjacent tracks to generate a track spacing scheme; planning and segmenting the track path to generate a track path scheme; setting track starting points and energy replenishment nodes to form a track layout scheme; importing the 3D model of the target inspection area for spatial interference verification, and correcting any errors; and finally, verifying the full-scene coverage integrity, outputting the final full-scene track coverage design scheme if no blind spots are found. This enables continuous, blind-spot-free coverage inspection of the target scene by the inspection robot, avoiding detection omissions caused by experience-based track layout or local optimization, and significantly improving the coverage integrity and engineering applicability of the inspection system.

Description

Technical Field

[0001] This invention relates to the field of intelligent inspection track design technology, specifically a method, device, and medium for designing inspection robot tracks that cover all scenarios. Background Technology

[0002] In public buildings and large industrial buildings, steel structure space frames, space trusses, and other structural forms are widely used, creating numerous concealed steel structures in their interior and upper areas that are difficult for people to access. These areas contain a large number of critical components and connection nodes, and their operational status directly affects structural safety and the quality of building maintenance.

[0003] Current steel structure inspection methods primarily rely on manual inspection or localized fixed testing, both of which have significant limitations. Manual inspection poses significant safety risks when working at heights and in concealed spaces. Due to limitations in human accessibility, the inspection coverage is limited, failing to effectively cover areas that are extremely difficult for humans to reach, such as high-altitude areas and the interior of space frames, resulting in numerous blind spots. Fixed testing equipment has limited deployment and coverage, only able to cover a limited spatial range, making it difficult to adapt to the complex spatial variations of large-span steel structure buildings and unable to achieve dynamic and continuous inspection of the entire scene.

[0004] With the development of inspection robot technology, track-based inspection robots are gradually being applied to steel structure inspection scenarios. However, in actual engineering, the track layout scheme lacks a systematic design method for full coverage of the inspection target, which can easily lead to problems such as unreasonable track spacing, overlapping or omissions in the detection coverage area, and improper placement of energy replenishment nodes. Ultimately, this results in blind spots in the inspection and affects the inspection effect.

[0005] Therefore, there is an urgent need for a track design method that focuses on full coverage of inspection targets to guide the rational layout of inspection robot tracks and ensure full-scenario coverage, engineering applicability, and economy. Summary of the Invention

[0006] (1) Technical problems to be solved

[0007] This invention provides a method, device, and medium for designing a full-scene coverage track for an inspection robot, which solves the problem of the lack of a systematic full-coverage design method for existing track layouts in concealed spaces of steel structures. It overcomes technical defects such as blind spots in detection, unreasonable track spacing, improper arrangement of energy replenishment nodes, and underutilization of robot posture extension capabilities caused by experience-based track layout or local optimization.

[0008] (2) Technical solution

[0009] To achieve the above objectives, the present invention provides a method for designing a full-scene coverage track for an inspection robot, the method comprising the following steps:

[0010] Step S1: Based on the structural drawings and on-site data of the building to be inspected, a three-dimensional model is created for the steel structure space area. The key inspection targets are then identified in the three-dimensional model to obtain a three-dimensional model of the target inspection area.

[0011] Step S2: Based on the light sensitivity, zoom level, field of view, and effective detection distance of the detection module on the inspection robot, establish a detection coverage radius model and determine the maximum coverage radius of a single effective detection.

[0012] Step S3: Based on the detection coverage radius model, determine the spacing between adjacent tracks by combining the spatial scale of the target inspection area and the distribution density of structural components. The spacing between adjacent tracks is determined with the maximum coverage radius as the upper limit. The spacing between adjacent tracks is then corrected based on the occlusion between structural components and the posture extension capability of the inspection robot to generate a track spacing scheme.

[0013] Step S4: Plan the track path according to the overall spatial orientation of the target inspection area, determine the maximum length of a single track, and generate a track path scheme after segmenting the track path.

[0014] Step S5: Set the starting point of the track in the track path scheme, and plan the location of the replenishment node according to the endurance of the inspection robot and the operation inspection strategy. The starting point of the track and the replenishment node are preferably set in locations that are convenient for manual inspection and equipment maintenance. The track layout scheme is composed of the track spacing scheme, the track path scheme and the location of the replenishment node.

[0015] Step S6: Import the track layout scheme into the 3D model of the target inspection area. Perform spatial interference verification based on the posture extension capability and obstacle avoidance range of the inspection robot during the inspection process. If the spatial interference verification fails, return to step S3 or step S4 to modify the track spacing scheme or track path scheme and re-execute. If the verification passes, proceed to step S7.

[0016] Step S7: Perform a full-scene coverage integrity check on the verified track layout scheme. If it is confirmed that there are no blind spots in the target inspection area, the final track full-scene coverage design scheme will be output.

[0017] Preferably, in step S1, the target inspection area also preferentially includes areas that are inaccessible to humans or pose a risk of high-altitude operations, and areas accessible to humans are used as auxiliary inspection coverage areas, which are staggered from the manual inspection trails.

[0018] Preferably, in step S2, the establishment of the detection coverage radius model also incorporates the attenuation factor of image acquisition accuracy as distance increases. The attenuation factor is the change in imaging resolution of the detection module at different detection distances, to obtain the maximum effective detection distance; the maximum coverage radius is the maximum effective detection distance that meets the requirements for disease identification accuracy.

[0019] Preferably, in step S3, the method for correcting the spacing between adjacent tracks is to reduce the spacing between adjacent tracks in areas with dense structural components or severe obstruction to eliminate the detection blind spot caused by obstruction; and to increase the spacing between adjacent tracks in areas with sparse structural components and open space, provided that the maximum coverage radius is not exceeded.

[0020] The method for determining the spacing between adjacent tracks also includes discretizing the surface of the target inspection component in the target inspection area into multiple sampling points. Based on the imaging resolution of the detection module under different combinations of detection distance and zoom magnification, a functional relationship is established between the number of pixels per unit area of ​​the target surface and the detection distance and zoom magnification. The minimum number of pixels corresponding to the minimum identifiable size of the target defect is used as a threshold to obtain a set of feasible detection distances that meet the defect identification accuracy requirements. The set of feasible detection distances is then mapped to a detectable spatial region surrounding each target inspection component.

[0021] For each detection posture combination, the visibility of each sampling point is determined by the ray casting algorithm. Based on the condition that the sampling point is in the detectable space area and is visible, the effective coverage number of each target inspection component is statistically obtained. When the effective coverage number of a target inspection component is lower than the preset lower limit, the distance between adjacent tracks in the area where the target inspection component is located is automatically reduced. When the effective coverage number of a target inspection component is higher than the preset upper limit and the corresponding imaging resolution margin is greater than the preset margin threshold, the distance between adjacent tracks in the area where the target inspection component is located is automatically enlarged.

[0022] Preferably, in step S4, the determination of the maximum length of a single track also takes into account the degree of thermal expansion and contraction deformation of the track structure, the power consumption constraints of the inspection robot, inspection efficiency, and the convenience of segmented maintenance.

[0023] Determining the maximum length of a single track group and segmenting the track path also includes:

[0024] Based on the typical temperature conditions of the area where the building to be inspected is located, an overall finite element analysis model of the steel structure is established. The temperature deformation of the overall finite element analysis model of the steel structure is calculated to obtain the displacement of the track support nodes under each temperature condition and mapped to the track centerline to obtain the spatial deformation distribution of the track under each temperature condition. The spatial deformation distribution includes longitudinal expansion, vertical deflection and planar curvature distribution.

[0025] The throughput parameters of the inspection robot when it runs on the track are obtained. The throughput parameters include at least the robot's allowable track gradient change rate, allowable track local curvature, and allowable track gauge change range.

[0026] By comparing the spatial deformation distribution of the track under various temperature conditions with the throughput parameters, the track positions where any condition is not met under any temperature condition are marked as potential segmentation positions, forming a preliminary set of segmentation positions.

[0027] In the initial segmentation location set, expansion compensation nodes or support adjustment nodes are set so that the spatial deformation distribution of each single track under all temperature conditions does not exceed the range of the inspection robot's throughput parameters, and the maximum length of a single track is determined.

[0028] Preferably, in step S5, the track starting point and the recharge node are planned in a location that facilitates manual inspection and equipment maintenance; the distance between adjacent recharge nodes is not greater than the maximum continuous running distance of the inspection robot in a fully charged state.

[0029] Preferably, in step S6, the posture expansion of the inspection robot includes the extension, rotation, or pitching motion of the detection module. The spatial interference verification verifies whether the detection range of the inspection robot under different postures can be effectively covered without structural collision by simulating the detection range of the inspection robot in the three-dimensional model of the target inspection area.

[0030] Preferably, in step S7, the full-scene coverage integrity verification is completed through three-dimensional model simulation or digital twin method, and the target inspection area is verified one by one to confirm that each key inspection target is within at least one maximum coverage radius.

[0031] Based on the same inventive concept, the present invention also provides a full-scene coverage design device for inspection robot tracks, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to call the executable instructions stored in the memory to execute a full-scene coverage design method for inspection robot tracks.

[0032] Based on the same inventive concept, the present invention also provides a readable storage medium storing a computer program, which is executed by a processor as a method for designing a full-scene coverage of an inspection robot track.

[0033] (3) Beneficial effects

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] 1. With full coverage of inspection targets as the core of the design, the system fundamentally solves the problem of coverage omissions caused by the reliance on experience in track layout in existing technologies through a complete design process, including 3D modeling, coverage radius model establishment, track spacing system planning, spatial interference verification, and full-scene coverage integrity verification, and significantly improves the coverage integrity of the inspection system.

[0036] 2. By using detection capabilities to infer track layout parameters, a precise match between coverage capacity and track density can be achieved, avoiding the waste of resources caused by insufficient coverage or excessive track density due to experience-based estimations in existing technologies.

[0037] 3. The robot's posture extension capabilities are integrated into both the spacing design and interference verification stages. The extension capabilities of the detection module, such as extension, rotation, and pitch, serve as compensation factors for track spacing correction. For areas where occlusion blind spots can be eliminated through posture extension, the robot's detection range under different postures is simulated in a 3D model of the target inspection area to verify whether effective detection coverage of occluded areas can be achieved without structural collisions. Targeted feedback corrections are triggered when the verification fails. This two-stage collaborative mechanism fully unleashes the robot's posture extension capabilities, significantly improving the detection coverage and design adaptability in complex steel structure occlusion scenarios.

[0038] 4. Incorporate the coordinated planning of the track starting point and the energy replenishment node into the track layout scheme to ensure the continuity of inspection with quantitative constraints and to take into account the convenience of later operation and maintenance with the requirement of manual accessibility.

[0039] 5. Applicable to various steel structure concealed space scenarios, with strong engineering versatility and easy to promote and apply. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating a method for designing a full-scene coverage track for an inspection robot according to the present invention.

[0041] Figure 2 This is a structural diagram of a full-scene coverage design device for an inspection robot track according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Example 1: As Figure 1 As shown in the figure, this embodiment provides a method for designing a full-scene coverage track for an inspection robot, the method including the following steps:

[0044] Step S1: Based on the structural drawings and on-site data of the building to be inspected, a three-dimensional model is created for the steel structure space area. The key inspection targets are then identified in the three-dimensional model to obtain a three-dimensional model of the target inspection area.

[0045] The target inspection area also prioritizes areas that are inaccessible by humans or pose a risk of working at height, and areas accessible by humans are used as auxiliary inspection coverage areas, which are staggered from the manual inspection paths.

[0046] For example, this embodiment takes the roof steel structure space frame of a large railway station building as the application object (hereinafter referred to as "this station building"). The roof of this station building adopts a large-span steel structure space frame system, with a roof plan dimension of approximately several hundred meters and a space area of ​​tens of thousands of square meters. The interior contains various types of inspection targets, such as ball joints, rods, ceiling plate joints, roof floor plates, purlin joints, and various ancillary facilities. It is a typical large-span, highly repetitive three-dimensional steel structure concealed space.

[0047] Data collection is conducted on the steel structure space of the building to be inspected. Data sources include, but are not limited to: original architectural design drawings, structural construction drawings and as-built drawings, BIM model files, and on-site 3D laser scanning or total station re-measurement data. When there are discrepancies between the on-site measured data and the original drawings, the on-site measured data shall prevail to ensure that the 3D model accurately reflects the as-built status of the structure.

[0048] Based on the above data, a high-fidelity 3D model of the steel structure space is established using 3D modeling software. The 3D model should fully include the following elements: the spatial location and geometric dimensions of load-bearing components (main truss chords, web members, ball joints); the spatial distribution of secondary components (purlins, support members); the location of key connection nodes and welds; and the spatial location of existing walkways, maintenance access, and other manual passage facilities.

[0049] After the three-dimensional model is established, the inspection targets in the model are systematically identified and classified to form a three-dimensional model of the target inspection area. The specific marking rules are as follows: (1) Priority target area (Class A): Areas that cannot be reached by manpower or where there is a risk of high-altitude operation, including all ball nodes, rods and connecting nodes inside the grid structure whose height exceeds the safety operation limit, as well as high-altitude cantilever areas and structural mezzanine spaces. This type of area is the primary coverage object of the track inspection system. (2) Auxiliary coverage area (Class B): Areas that can be reached by manpower via the walkway. Although this type of area can be inspected by manpower, the track inspection system can also cover it in order to improve the overall inspection efficiency. When laying the track, the track and the manual inspection walkway should be staggered as much as possible to avoid functional overlap and waste of resources, and to prevent the track installation from interfering with the passage of the walkway. (3) Exclusion area (Class C): Dead corners that cannot be reached by any inspection means due to structural closure or physical obstruction. These areas are marked in the coverage design and are treated separately during subsequent coverage verification. Through the above classification and labeling, a three-dimensional model of the target inspection area is formed, which includes the spatial coordinates of the target inspection components, component type, priority level, and spatial orientation information.

[0050] Step S2: Based on the light sensitivity, zoom level, field of view, and effective detection distance of the detection module carried by the inspection robot, establish a detection coverage radius model and determine the maximum coverage radius of a single effective detection.

[0051] The detection coverage radius model is also combined with the attenuation factor of image acquisition accuracy as distance increases. The attenuation factor is the change in imaging resolution of the detection module at different detection distances, to obtain the maximum effective detection distance; the maximum coverage radius is the maximum effective detection distance that meets the requirements of disease identification accuracy.

[0052] For example, the establishment of the detection coverage radius model depends on the core technical parameters of the detection module, including but not limited to: (1) Photosensitive capability: The equivalent photosensitiveness (ISO) and dynamic range of the image sensor of the detection module determine the upper limit of the available imaging distance under different lighting conditions; (2) Zoom magnification: The maximum magnification of the optical zoom or digital zoom of the detection module directly affects the pixel resolution at long distances; (3) Field of view: The horizontal field of view and the vertical field of view determine the spatial coverage sector of a single detection; (4) Effective detection distance: The maximum working distance at which the target surface state can be clearly collected under the nominal detection accuracy.

[0053] The imaging quality of the detection module degrades with increasing detection distance, and this degradation pattern needs to be incorporated into the establishment of the coverage radius model. Specifically, the degradation factor is the change in imaging resolution of the detection module at different detection distances, which manifests as follows: as the detection distance increases, the number of pixels per unit area of ​​the target surface decreases. When the number of pixels is lower than the minimum threshold required for disease identification accuracy, this detection distance becomes the upper limit of effective detection.

[0054] In this station embodiment, taking the identification of surface cracks on spherical nodes as an example, the maximum effective detection distance is calculated using the accuracy attenuation model, provided that the identification accuracy requirements are met. The maximum effective detection distance is then mapped to an equivalent coverage radius in three-dimensional space, establishing a detection coverage radius model.

[0055] It should be noted that when the inspection robot is equipped with multiple types of detection modules (such as visible light cameras, infrared thermal imagers, and 3D laser scanners), a corresponding detection coverage radius model should be established for each type of detection module, and the intersection of the coverage radii of each module should be taken as the upper limit of the constraint for the track spacing design to ensure that all types of detection can meet the accuracy requirements.

[0056] Step S3: Based on the detection coverage radius model, determine the distance between adjacent tracks by combining the spatial scale of the target inspection area and the distribution density of structural components. The distance between adjacent tracks is determined with the maximum coverage radius as the upper limit. The distance between adjacent tracks is then corrected based on the occlusion between structural components and the posture extension capability of the inspection robot to generate a track spacing scheme.

[0057] Methods for correcting the spacing between adjacent tracks include:

[0058] In areas with dense structural components or severe obstruction, reduce the spacing between adjacent tracks to eliminate detection blind spots caused by obstruction; in areas with sparse structural components and open space, increase the spacing between adjacent tracks without exceeding the maximum coverage radius.

[0059] For example, using the maximum coverage radius given by the detection coverage radius model as the upper limit, and combining the spatial scale of the target inspection area and the distribution density of structural components, the reference spacing between adjacent tracks is determined. The principle for determining the reference spacing is: under the normal operating posture of the tracks (the detection module is in the middle position, without posture extension movement), the coverage area of ​​two adjacent tracks can just completely cover all the target inspection components between the two tracks, without any coverage gaps; in this station embodiment, based on the calculation of the detection coverage radius model, the initial value of the reference spacing between adjacent tracks is set to 2 × the maximum coverage radius to ensure that the coverage areas of adjacent tracks are seamlessly connected.

[0060] In steel structure spaces, there are varying degrees of occlusion between components, which can lead to local blind spots in the direct-view detection mode. To eliminate blind spots caused by occlusion, the reference spacing needs to be corrected for occlusion. The specific correction rules are as follows: (1) In areas with dense structural components or severe occlusion, reduce the track spacing so that the detection module can bypass the occluding components from a closer distance and different perspectives to achieve effective coverage of the occluded target. The reduction range is determined by simulation calculation in the three-dimensional model based on the occlusion depth and the area of ​​the occluded target. (2) In areas with sparse structural components and open space, the track spacing can be appropriately increased without exceeding the maximum coverage radius to reduce the total number of tracks and lower the engineering cost.

[0061] The inspection robot is equipped with a detection module that has a certain attitude expansion capability (including extension, rotation, pitch, etc.), which can expand the actual detection coverage area without changing the track position. In the process of determining the track spacing scheme, the attitude expansion capability is included in the spacing correction calculation: (1) When the detection module is in the maximum expanded attitude, its effective detection range is larger than that in the normal attitude. Under the condition of attitude expansion capability compensation, the upper limit of track spacing in local areas can be appropriately widened to reduce the total number of tracks. (2) The use of attitude expansion compensation is based on the premise of not affecting the coverage integrity and not causing structural collisions. Through the above three steps of determining the reference spacing, occlusion correction and attitude expansion compensation, a track spacing scheme covering the entire range of the target inspection area is formed, and the specific track spacing values ​​of each area are recorded.

[0062] The method for determining the spacing between adjacent tracks also includes discretizing the surface of the target inspection component in the target inspection area into multiple sampling points. Based on the imaging resolution of the detection module under different combinations of detection distance and zoom magnification, a functional relationship is established between the number of pixels per unit area of ​​the target surface and the detection distance and zoom magnification. The minimum number of pixels corresponding to the minimum identifiable size of the target defect is used as a threshold to obtain a set of feasible detection distances that meet the defect identification accuracy requirements. The set of feasible detection distances is then mapped to a detectable spatial region surrounding each target inspection component.

[0063] For each detection posture combination, the visibility of each sampling point is determined by the ray casting algorithm. Based on the condition that the sampling point is in the detectable space area and is visible, the effective coverage number of each target inspection component is statistically obtained. When the effective coverage number of a target inspection component is lower than the preset lower limit, the distance between adjacent tracks in the area where the target inspection component is located is automatically reduced. When the effective coverage number of a target inspection component is higher than the preset upper limit and the corresponding imaging resolution margin is greater than the preset margin threshold, the distance between adjacent tracks in the area where the target inspection component is located is automatically enlarged.

[0064] Specifically, after establishing the detection coverage radius model, to integrate the occlusion effect, imaging accuracy attenuation law, and robot posture extension capability into the quantitative determination process of adjacent track spacing, it is necessary to first discretize the surface of each target inspection component within the target inspection area. Specifically, the outer surface of each target component is divided into several surface elements, and at least one sampling point is selected on each surface element. The spatial coordinates of this sampling point in the 3D model of the target inspection area are recorded. The spatial density of the sampling points should match the minimum identifiable size of the target defect. In principle, the sampling interval should not exceed half of the minimum identifiable size of the target defect to ensure that the discretization process does not miss any potential defect areas.

[0065] After discretizing the sampling points, it is necessary to establish a quantitative functional relationship between the imaging resolution of the detection module and the detection distance and zoom level. The basic principle is that when the detection module captures an image of a target component, the actual physical size corresponding to each pixel in the image increases with increasing detection distance and decreases with increasing zoom level. In other words, the greater the detection distance or the lower the zoom level, the fewer pixels can be allocated per unit area on the surface of the target component, and the weaker the ability to identify minute defects.

[0066] Based on the above patterns, the imaging resolution of the detection module under different combinations of detection distance and zoom level is systematically calibrated, establishing a functional relationship between the number of pixels per unit area of ​​the target surface and the detection distance and zoom level. On this basis, the minimum number of pixels corresponding to the smallest identifiable size of the target lesion is used as the judgment threshold. All distance and zoom combinations are judged one by one: if the number of pixels corresponding to the target lesion is not lower than the threshold under a certain distance and zoom combination, the detection distance is considered to meet the lesion identification accuracy requirements at the corresponding zoom level, and it is included in the feasible detection distance set; otherwise, it is excluded. By traversing all discrete distance points and typical zoom levels within the preset detection distance range, a feasible detection distance set that meets the lesion identification accuracy requirements is finally formed, and the maximum distance value in the set is taken as the maximum effective detection distance of the detection module.

[0067] After determining the feasible detection distance set, a corresponding detectable spatial region is constructed in the 3D model around each target inspection component, with the surface of each component as the inner boundary and the maximum effective detection distance as the outer boundary. The physical meaning of this region is: only when the detection module is located within this spatial region can it have the geometric conditions to effectively detect the target component while meeting the imaging accuracy requirements; detection positions outside this region, regardless of occlusion, cannot be counted as effective detection due to insufficient imaging resolution caused by excessive distance.

[0068] After establishing the detectable spatial region, it is necessary to further determine whether the detection module can effectively observe each sampling point under unobstructed conditions. This is because even if the detection position meets the imaging accuracy requirements, effective detection cannot be completed if there are other structural components obstructing the sampling point and the detection module. To this end, the track path is discretized along the track centerline with a preset step size to obtain a series of discrete detection position points. For each detection position point, the attitude space is discretized according to the range of allowable extension, horizontal rotation angle, and pitch angle of the detection module, forming a finite number of detection attitude combinations.

[0069] For each "detection position-attitude combination" and each sampling point, a ray casting algorithm is used to determine visibility: a detection ray is emitted from the optical center position of the detection module in the current attitude towards the target sampling point, and it is determined whether the ray intersects with other structural surfaces in the 3D model before reaching the sampling point. If the ray is not blocked by any structural surface during its journey to the sampling point, the sampling point is determined to be visible under the current "detection position-attitude combination"; if the ray intersects with other structural surfaces along its path, the sampling point is determined to be occluded, and this detection is invalid.

[0070] Based on the combined visibility assessment results and the constraints of the detectable spatial area, a detection is recorded as valid only if both of the following conditions are met simultaneously: First, the current detection distance falls within the set of feasible detection distances, meaning the imaging resolution meets the accuracy requirements for disease identification; second, the sampling point is visible under the current "detection position-attitude combination," meaning the detection ray path is unobstructed. Both conditions are indispensable and together constitute the complete criteria for valid detection.

[0071] By traversing all discrete detection positions and attitude combinations, the total number of times each sampling point is effectively detected is counted. The minimum number of effective detections for all sampling points on the same target inspection component is taken as the effective coverage index for that component. The reason for selecting the minimum value instead of the average value is that if any sampling point on the component fails to be effectively detected, it means that the component has a local detection blind spot. The minimum number of effective coverages is a direct measure of this weakest link, which can objectively reflect the sufficiency of the component's coverage under the current track layout scheme.

[0072] After calculating the effective coverage count of each target inspection component, the spacing between adjacent tracks is adaptively adjusted according to the following rules. When the effective coverage count of a target inspection component is lower than the preset lower limit, it indicates that the target inspection component has insufficient coverage under the existing track layout density. The area where it is located is marked as a weak coverage area, and the spacing between adjacent tracks in this area is automatically reduced. After the track spacing is reduced, the detection module can approach the target component from a closer distance and from more perspectives. This helps to bypass obstructing components and complete detection from different directions, and also allows more detection positions to fall within the detectable space. The reduced track spacing must be re-triggered to complete the entire calculation process of discretization, ray projection, and effective coverage count statistics, iterating repeatedly until the effective coverage count of the target inspection component reaches the preset lower limit requirement, or the track spacing has been reduced to the preset minimum value allowed by the project.

[0073] When the effective coverage count of a target inspection component exceeds a preset upper limit, and the imaging resolution margin under most effective detection posture combinations is greater than a preset margin threshold, it indicates that there is redundancy in the track layout of that area, and the existing track density exceeds the minimum requirement for coverage. There is room to appropriately thin out the tracks while ensuring coverage integrity. The imaging resolution margin is introduced as an additional criterion to avoid hastily increasing the spacing based solely on a high coverage count. If the current effective detection imaging resolution is already close to the lower limit of recognition accuracy, even with sufficient coverage count, it is not advisable to further increase the spacing, to prevent the imaging accuracy of some detection positions from falling below the threshold after spacing enlargement. Only when both the coverage count and imaging resolution margin meet the conditions is the spacing between adjacent tracks in that area automatically enlarged, and the coverage verification is re-performed to ensure that the effective coverage count of each component after enlargement is still not lower than the preset lower limit, and the track spacing does not exceed the upper limit of the spacing defined by the maximum effective detection distance.

[0074] Step S4: Plan the track path according to the overall spatial orientation of the target inspection area, determine the maximum length of a single track, and generate a track path scheme after segmenting the track path.

[0075] The determination of the maximum length of a single track also takes into account the degree of thermal expansion and contraction deformation of the track structure, the power consumption constraints of the inspection robot, inspection efficiency, and the convenience of segmented maintenance.

[0076] For example, the track path is planned in the three-dimensional model according to the overall spatial orientation of the target inspection area. The track path planning should follow the following principles: (1) The orientation of the track path should be consistent with the main axis direction of the target inspection area to achieve the maximum spatial coverage efficiency with the shortest total track length; (2) The spacing between adjacent tracks should conform to the track spacing scheme; (3) The track path should avoid dense areas with existing walkways, pipelines and other obstacles as much as possible, and reserve space for installation operations; (4) The design of the track path should consider the feasibility of installation and construction, and prioritize the selection of structural nodes with installation conditions as track hanging points.

[0077] In this embodiment of the station building, the main axis of the roof steel structure space frame is longitudinal (parallel to the building's long axis), and the track path is laid out parallel to the longitudinal direction, with the lateral spacing implemented according to the track spacing scheme. For the roof ends, variable cross-section areas, and local irregular spaces, the track path is adjusted accordingly to ensure full coverage.

[0078] After completing the overall path planning, the maximum length of a single track group needs to be determined according to the comprehensive constraints, and the track path needs to be segmented. The determination of the maximum length L of a single track group takes into account the following factors: (1) Thermal expansion and contraction deformation constraints: The track structure will undergo thermal expansion and contraction deformation under temperature change conditions. When a single track group is too long, the cumulative amount of deformation may exceed the allowable deviation range of the clamping mechanism of the inspection robot, resulting in the robot's operation being hindered; (2) Power consumption constraints: The length of a single track group shall not exceed the maximum continuous running distance of the inspection robot under full power conditions, so as to ensure that the robot can safely complete the inspection task of the entire single track group without power failure and shutdown under full power conditions; (3) Inspection efficiency constraints: When the length of a single track group is too large, the time taken for a single inspection task is too long, affecting the overall inspection frequency and efficiency. A reasonable inspection cycle should be determined according to the actual needs of the project, and the reasonable upper limit of the length of a single track group should be determined accordingly; (4) Convenience of segmented maintenance: Shorter single tracks are easier to disassemble and maintain locally, reducing the difficulty of later operation and maintenance.

[0079] Taking into account the above four constraints, the maximum length of a single track is the minimum value among the upper limits of each constraint, that is: maximum length of a single track = min (upper limit of thermal expansion and contraction constraint, upper limit of power consumption constraint, upper limit of inspection efficiency constraint, and upper limit of maintenance convenience suggestion).

[0080] Based on the maximum length of a single track group, the overall track path is divided into several independent track segments, each segment's length not exceeding the maximum length of a single track group. Dock transition nodes are set between adjacent track segments, allowing the inspection robot to complete cross-segment transfers at these nodes, enabling continuous inspection of the entire track network. After segmentation, a track path scheme is generated, containing the start and end coordinates of each track segment, segment number, and track length.

[0081] Determining the maximum length of a single track group and segmenting the track path also includes:

[0082] Based on the typical temperature conditions of the area where the building to be inspected is located, an overall finite element analysis model of the steel structure is established. The temperature deformation of the overall finite element analysis model of the steel structure is calculated to obtain the displacement of the track support nodes under each temperature condition and mapped to the track centerline to obtain the spatial deformation distribution of the track under each temperature condition. The spatial deformation distribution includes longitudinal expansion, vertical deflection and planar curvature distribution.

[0083] The throughput parameters of the inspection robot when it runs on the track are obtained. The throughput parameters include at least the robot's allowable track gradient change rate, allowable track local curvature, and allowable track gauge change range.

[0084] By comparing the spatial deformation distribution of the track under various temperature conditions with the throughput parameters, the track positions where any condition is not met under any temperature condition are marked as potential segmentation positions, forming a preliminary set of segmentation positions.

[0085] In the initial segmentation location set, expansion compensation nodes or support adjustment nodes are set so that the spatial deformation distribution of each single track under all temperature conditions does not exceed the range of the inspection robot's throughput parameters, and the maximum length of a single track is determined.

[0086] Specifically, the roof steel structure space frame is a large-span structural system, with its supporting components spanning hundreds of meters. Under different seasonal temperature conditions, the steel structure as a whole will undergo significant thermal expansion and contraction deformation. As an auxiliary facility attached to the steel structure, the track system's supporting nodes shift synchronously with the steel structure, causing changes in the spatial shape of the track centerline. If the length of a single track set is set too large, the cumulative deformation of the track under extreme temperature conditions may exceed the inspection robot's throughput capacity, causing the robot to get stuck, derail, or even be damaged in localized areas of the track.

[0087] Based on meteorological data of the area where the building to be inspected is located, typical temperature conditions covering the extreme temperature variations throughout the year are selected as the calculation boundary conditions. Typical conditions generally include: extreme high-temperature conditions (summer's highest temperature, considering the additional warming effect of roof solar radiation), extreme low-temperature conditions (winter's lowest temperature), construction and installation closure conditions (reference temperature during track installation), and normal-temperature conditions in spring and autumn. These conditions together constitute a complete envelope set for temperature load analysis, ensuring that track deformation calculations cover the most unfavorable conditions throughout the year.

[0088] Based on meteorological data of the area where the building to be inspected is located, typical temperature conditions covering the extreme temperature variations throughout the year are selected as the calculation boundary conditions. Typical temperature conditions generally include: extreme high-temperature conditions (summer's highest temperature, considering the additional warming effect of roof solar radiation), extreme low-temperature conditions (winter's lowest temperature), construction and installation closure conditions (reference temperature during track installation), and normal-temperature conditions in spring and autumn. These typical temperature conditions collectively constitute the complete envelope set for temperature load analysis, ensuring that track deformation calculations cover the most unfavorable conditions throughout the year.

[0089] Based on the 3D model of the target inspection area, a finite element analysis model of the entire steel structure is established. Temperature difference values ​​relative to the installation reference temperature are assigned to each component of the steel structure under corresponding temperature conditions. ,in, For the first Component temperature under various working conditions To establish a reference temperature for installation, and considering the non-uniform temperature distribution effect caused by differences in solar shading between the upper and lower chord members of the roof, temperature deformation calculations were performed on the overall finite element analysis model of the steel structure after applying temperature loads. The spatial displacement vectors of the track support nodes relative to the installation reference state were extracted under various temperature conditions. subscript For the support node number, superscript The temperature condition is numbered, and the three components correspond to the longitudinal displacement component, lateral displacement component, and vertical displacement component of the track, respectively.

[0090] Using the displacement calculation results of each support node as input, an interpolation method is employed to map the displacement results of discrete support points to the continuous track centerline, thereby obtaining the spatial deformation distribution of the track centerline under various temperature conditions. The spatial deformation distribution includes longitudinal expansion, vertical deflection, and planar curvature distribution.

[0091] The longitudinal expansion / contraction is the cumulative elongation or shortening of the track centerline along the track's direction. It reflects the change in track length caused by thermal expansion and contraction and is a core indicator for determining whether expansion compensation nodes are needed. For a length of... The longitudinal linear expansion of a single orbit can be approximated as: ;in, The coefficient of linear expansion of steel (taken as...) , For the first The temperature difference between the operating temperature and the installation reference temperature. This is the length of a single track group. When... Exceeding the maximum compensation amount of the expansion compensation node When that happens, segments need to be set at the corresponding positions.

[0092] The vertical deflection is the offset of the track centerline in the vertical direction, reflecting the undulating deformation of the track caused by uneven vertical displacement of the support nodes. (Adjacent support nodes...) and The rate of change of track gradient between them can be expressed as: ;in, Support node In the Vertical displacement components relative to the installation reference state under various temperature conditions. Support node In the Vertical displacement components relative to the installation reference state under various temperature conditions. This is the arc length of the track centerline between adjacent support nodes. Exceeding the robot's allowed slope change rate limit When the time is right, the corresponding position is the potential segmentation position.

[0093] The planar curvature distribution refers to the degree of curvature of the track centerline in the horizontal plane, reflecting the lateral deflection of the track caused by the difference in lateral displacement between adjacent support points. The track's curvature in the horizontal plane... The local curvature at the segment is approximately: ; Support node In the The lateral displacement component relative to the installation reference state under various temperature conditions. Support node In the The lateral displacement component relative to the installation reference state under certain temperature conditions, when Exceeding the robot's allowed local curvature limit When this happens, the corresponding position is also recorded as the potential segmentation position.

[0094] Obtain the inspection robot's track clearance parameters from the robot manufacturer or calibrate them through bench testing. These clearance parameters should at least cover the following three constraints: allowable track gradient variation rate. This refers to the maximum permissible variation in the vertical slope of the track centerline per unit length. Exceeding this value will prevent the robot's drive mechanism from smoothly traversing the track due to excessive pitch angle; the permissible local curvature of the track. This refers to the maximum permissible curvature of the track centerline in the horizontal plane. Exceeding this value will cause motion interference or derailment of the robot's steering mechanism due to excessive curvature; the permissible range of track gauge variation. This refers to the distance between the centerlines of the two rails after temperature deformation, relative to the design gauge. The maximum permissible deviation is defined; if this range is exceeded, the robot's wheels will not be able to properly engage with the track. The constraint can be expressed as: ;in, For the first The actual track gauge under various temperature conditions.

[0095] Using discrete points along the track centerline as units, the longitudinal expansion, vertical deflection, and planar curvature distributions under various temperature conditions are compared and verified segment by segment with the aforementioned throughput parameters to form a unified set of over-limit judgment conditions:

[0096] .

[0097] The above formula must be applied to all All temperature conditions must be met simultaneously. Any track position where any condition is not met under any operating condition is recorded as a potential segment position. Finally, the union of all out-of-limit positions under all temperature conditions is taken to form a preliminary set of segment positions.

[0098] The initial set of segment locations is processed for engineering rationality: when the distance between two adjacent potential segment locations is too close, the location with better structural conditions and better human accessibility is selected as the actual segment location to reduce the number of unnecessary segments; at the same time, the power consumption constraints and inspection efficiency of the inspection robot are comprehensively considered to avoid excessively short single track lengths leading to excessively high recharging frequency or too many inspection interruptions. At the determined segment locations, corresponding compensation measures are set according to the different types of exceedances.

[0099] For segmentation caused by excessive longitudinal expansion, expansion compensation nodes are installed. These nodes allow relative sliding between adjacent track segments in the longitudinal direction to release axial thermal stress caused by temperature. For segmentation caused by excessive vertical deflection or planar curvature, support adjustment nodes are installed. During installation, these nodes can precisely correct the local track alignment by adjusting the support height and lateral position, eliminating alignment exceedances caused by uneven support point displacement. Temperature deformation calculations and throughput capacity comparisons are performed on each individual track group until the deformation distribution of each track segment meets the above-mentioned set of exceedance criteria under all temperature conditions. This ensures that the geometric deformation of each individual track group remains within the allowable range for safe operation of the inspection robot throughout the year and under all temperature conditions, ultimately determining the maximum allowable length of each track group and the complete track path segmentation scheme.

[0100] Step S5: Set the starting point of the track in the track path scheme, and plan the location of the replenishment node according to the endurance of the inspection robot and the operation inspection strategy. The starting point of the track and the replenishment node are preferably set in locations that are convenient for manual inspection and equipment maintenance. The track layout scheme is composed of the track spacing scheme, the track path scheme and the location of the replenishment node.

[0101] The starting point of the track and the recharge nodes are planned in locations that facilitate manual inspection and equipment maintenance; the distance between adjacent recharge nodes is no greater than the maximum continuous running distance of the inspection robot when it is fully charged.

[0102] For example, the starting point of the track is the starting position where the inspection robot begins to perform its inspection task, and it is also the main work point for robot parking, charging, and manual maintenance. The setting of the starting point of the track should meet the following requirements: (1) It should be located in a position that is accessible to humans and facilitates the loading and unloading of the equipment and maintenance; (2) It should be located at the end of each set of tracks or at a node with conditions for manual passage; (3) It should be coordinated with the segment node positions of the track path plan to avoid conflicts between the starting point and the segment docking node positions.

[0103] The power replenishment node is an intermediate node that provides charging or power replacement for the inspection robot. The spacing of its arrangement is directly related to the robot's endurance. The planning rules for power replenishment nodes are as follows: (1) The spacing between adjacent power replenishment nodes shall not be greater than the maximum continuous running distance of the inspection robot when fully charged, so as to ensure that the robot can reach the next power replenishment node before the power is exhausted and that there is no power outage or shutdown; (2) Power replenishment nodes should be set at the segment docking nodes of the track path to reduce the number of dedicated docking structures and reduce engineering costs; (3) Power replenishment nodes should be located in a manually accessible location to facilitate the installation, maintenance and power replacement of charging equipment; (4) In areas with long track paths, if the spacing between segment docking nodes exceeds the maximum continuous running distance of the inspection robot when fully charged, an additional power replenishment node should be added between the two segment nodes. The location of the additional node should be selected at the structural node to facilitate installation and fixation.

[0104] After completing the track starting point setting and energy replenishment node planning, the following three components are integrated to form a complete preliminary track layout scheme: (1) track spacing scheme; (2) track path scheme; (3) energy replenishment node location: including the spatial coordinates and installation method of each track starting point and intermediate energy replenishment node.

[0105] Step S6: Import the track layout scheme into the 3D model of the target inspection area. Perform spatial interference verification based on the posture extension capability and obstacle avoidance range of the inspection robot during the inspection process. If the spatial interference verification fails, return to step S3 or step S4 to modify the track spacing scheme or track path scheme and re-execute. If the verification passes, proceed to step S7.

[0106] The inspection robot's posture extension includes the extension, rotation, or pitching movements of the detection module. The spatial interference verification simulates the detection range of the inspection robot in different postures in a three-dimensional model of the target inspection area to verify whether it can effectively detect and cover the occluded area without structural collision.

[0107] For example, after importing the track layout scheme into the 3D model of the target inspection area, a multi-pose dynamic collision envelope model is first established for the inspection robot. Specifically, independent rigid body geometric models are established for the robot's body, track clamping mechanism, detection module support, and detection module head, with each rigid body represented by a bounding box. The detection module has three degrees of freedom: extension, horizontal rotation, and pitch. The extension action allows the detection module to extend or retract along a direction perpendicular to the track, used to adjust the detection distance between the module and the target inspection component without moving the track position. The horizontal rotation action allows the detection module to rotate around the vertical axis, used to cover the target inspection components that are blocked in the horizontal direction on both sides of the track. The pitch action allows the detection module to shoot upwards or downwards around the horizontal axis, used to alternately cover the target inspection components above and below the track. The parameters of the above three degrees of freedom are uniformly discretized within their respective motion ranges, and their respective discrete step lengths are taken to form a complete attitude configuration space. The total number N of discrete attitude nodes in the attitude configuration space is the product of the number of discrete nodes for each of the three degrees of freedom. For each discrete attitude node in the attitude configuration space, the spatial pose of each rigid body is detected by the forward kinematics calculation module, and the bounding boxes of each rigid body are superimposed under that attitude to obtain the complete collision envelope of the inspection robot under that attitude. After the collision envelopes corresponding to all discrete attitude nodes are pre-calculated offline, they are cached as a collision envelope database with the attitude node number as the index.

[0108] After obtaining the collision envelope database, spatial interference verification is performed on the complete process of the inspection robot running along each track segment. The track is discretized with a fixed step size along its direction, resulting in a series of discrete detection positions. For each discrete detection position, structural components near that position are selected as candidate interference objects from the spatial octree index structure to reduce the scope of subsequent precise calculations. For each candidate interference object, the collision envelope corresponding to each discrete attitude node is retrieved sequentially from the collision envelope database, and the minimum distance between the collision envelope and the geometry of the candidate interference object is precisely calculated using the GJK distance calculation algorithm. When the minimum distance < 0, it is determined to be hard interference, i.e., the collision envelope and the structural component overlap; when 0 ≤ minimum distance < preset safety clearance threshold, it is determined to be critical interference, recorded but not forcibly triggered for correction; when the minimum distance ≥ preset safety clearance threshold, it is determined to be no interference, and the discrete attitude node is safe and usable at that discrete detection position. The aforementioned interference detection covers two scenarios: first, when the inspection robot is moving along the track in its normal operating posture, the interference between the robot body and the surrounding steel structure components, walkway structure and pipeline equipment; second, when the inspection robot takes posture extension actions at each target inspection component, the interference between the detection module and its support and the surrounding structural components.

[0109] After determining the interference state, for combinations of discrete detection positions and discrete attitude nodes that are determined to be interference-free, the visibility of the detection module on the target inspection component under that position and attitude combination is further quantitatively evaluated using a ray casting method. The surface of the target inspection component is discretized into several sampling points. A ray is emitted from the current position of the detection module to each sampling point, and it is determined whether the ray is truncated by other structural components in the 3D model before reaching the sampling point: if the ray is not truncated, the sampling point is visible under the current position and attitude combination; if the ray is truncated, the sampling point is not visible.

[0110] During the spatial interference verification process, for locations with localized occlusion blind spots in the track layout scheme, an attempt is made to eliminate the blind spots using a combination of the aforementioned attitude extension actions, followed by a verification of collision scenarios under the extended attitude state. If the attitude extension effectively eliminates the blind spots without causing collisions, the occlusion problem at that location is considered resolved, and no modification to the track spacing is required. If the attitude extension still cannot completely eliminate the blind spots, or if there is a risk of collision under the extended attitude state, a feedback correction mechanism must be triggered.

[0111] If the spatial interference check fails, return to the following steps for targeted correction based on the specific reason for the failure: (1) If the interference is caused by a collision between the track position and an obstacle, return to step S4, adjust the local path of the affected track segment in the track path scheme, bypass the obstacle, and then re-execute S5 and S6; (2) If the interference is caused by severe occlusion in a local area and the attitude extension cannot eliminate the detection blind spot, return to step S3 and reduce the track spacing in that area; (3) If the interference involves both path collision and coverage blind spot problems, return to step S3 and step S4 for coordinated correction.

[0112] Step S7: Perform a full-scene coverage integrity check on the verified track layout scheme. If it is confirmed that there are no blind spots in the target inspection area, the final track full-scene coverage design scheme will be output.

[0113] The full-scene coverage integrity verification is completed through 3D model simulation or digital twin method, and the target inspection area is checked one by one to confirm that each key inspection target is within the maximum coverage radius at least once.

[0114] For example, full-scene coverage integrity verification can be performed through 3D model simulation or digital twin methods. 3D model simulation verification: In the 3D model of the target inspection area, an independent coverage verification is performed on each target inspection component (structural component, connection node) to determine whether it is within the maximum coverage radius of at least one detection position on at least one track, and whether the detection position is not obstructed by other components or whether obstruction can be effectively avoided through posture extension actions. For complex occlusion scenarios that are difficult to verify intuitively through 3D model simulation, a digital twin platform can be used for higher-fidelity virtual inspection simulation. By rendering the robot's actual field of view at each detection position in real time, the coverage integrity can be visually confirmed.

[0115] The final track full-scene coverage design scheme marks the number of times each key inspection target is covered and the corresponding inspection track number. Key inspection targets that are not covered are highlighted to form a distribution map of inspection blind spots.

[0116] If the full-scene coverage integrity check finds a detection blind spot, then targeted feedback correction is made according to the cause of the blind spot: (1) If the blind spot is caused by the track spacing being too large, resulting in the coverage radius not being able to connect, then return to step S3 and reduce the track spacing in the corresponding area; (2) If the blind spot is caused by the unreasonable track path direction, resulting in no track coverage in a local area, then return to step S4 and add or adjust the track path in the blind spot area; (3) If the blind spot is caused by local occlusion and the attitude extension capability was not fully utilized during the check in step S6, then return to step S6, re-check the attitude extension for that location, confirm whether the blind spot can be eliminated through attitude extension, and then re-execute S7; (4) If the blind spot is the exclusion area marked in step S1 (Class C, which cannot be reached by any physical means), then the blind spot is described in writing in the final solution and is not considered a design defect. The above feedback correction process may also require multiple iterations until the full-scene coverage integrity check confirms that there are no accessibility blind spots in the target inspection area.

[0117] After the full scene coverage integrity verification is passed, the final track full scene coverage design scheme is output. The final track full scene coverage design scheme includes the following file contents: (1) Track layout general diagram: mark the spatial position, number and direction of all tracks in the three-dimensional model of the target inspection area; (2) Track spacing scheme table: record the track spacing value and correction basis of each area; (3) Single track segment table: record the number, start and end coordinates, length and maximum allowable length basis of each track segment; (4) Starting point and energy replenishment node layout diagram: mark the spatial position and installation method of each track starting point and energy replenishment node; (5) Coverage integrity verification report: include the coverage status statistics of each target inspection component and the verification pass certificate.

[0118] Example 2: Based on the same inventive concept, this example provides a full-scene coverage design device for inspection robot tracks, including: a processor; a memory for storing executable instructions of the processor; wherein the processor is configured to call the executable instructions stored in the memory to execute a full-scene coverage design method for inspection robot tracks.

[0119] like Figure 2 As shown, the device consists of a memory 11, a program and operating system pre-stored in the memory 11, a processor 10, a communication bus 13, and a communication interface 12. The operating system provides the basic software environment for program execution, the communication bus 13 is the core data transmission link within the hardware architecture, enabling bidirectional data interaction between the processor 10, the memory 11, and the communication interface 12, and the communication interface 12 is the information interaction port between the hardware architecture and external devices. The above hardware components perform their respective functions and cooperate to provide complete hardware support for the retrieval, execution, and data interaction of computer programs in the storage medium.

[0120] In specific implementations, the processor can be a general-purpose computer processor (CPU), a graphics processing unit (GPU), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), or any combination of the above processors, without being limited to a specific processor type. The memory can be random access memory (RAM), read-only memory (ROM), flash memory, hard disk drive (HDD), solid-state drive (SSD), or any combination of the above storage media.

[0121] The device may further include an input interface and an output interface. The input interface is used to receive building structure drawing data, on-site 3D scan data, detection module technical parameters, and design constraints, and can be connected to a Building Information Modeling (BIM) database, an engineering drawing management system, or on-site measurement equipment. The output interface is used to output track layout schemes, spatial interference verification reports, and coverage integrity verification reports, and can be connected to an engineering design platform, a construction management system, or an inspection robot control system.

[0122] Example 3: Based on the same inventive concept, this example provides a readable storage medium storing a computer program, which, when executed by a processor, implements a method for designing a full-scene coverage of the inspection robot's track.

[0123] In this embodiment, the scope of protection for readable storage media is not limited to any specific physical form, including but not limited to: tangible storage media such as random access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD), hard disk drive (HDD), USB flash drive, portable hard drive, optical disc, etc., as well as media that are transmitted over a network in the form of program signals and temporarily stored in the hardware architecture memory for the processor to retrieve and execute.

[0124] Finally, it should be noted that although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for designing a full-scene coverage track for an inspection robot, characterized in that, The method includes the following steps: Step S1: Based on the structural drawings and on-site data of the building to be inspected, a three-dimensional model is created for the steel structure space area. The key inspection targets are then identified in the three-dimensional model to obtain a three-dimensional model of the target inspection area. Step S2: Based on the light sensitivity, zoom level, field of view, and effective detection distance of the detection module carried by the inspection robot, establish a detection coverage radius model and determine the maximum coverage radius of a single effective detection. Step S3: Based on the detection coverage radius model, determine the distance between adjacent tracks by combining the spatial scale of the target inspection area and the distribution density of structural components. The distance between adjacent tracks is determined with the maximum coverage radius as the upper limit. The distance between adjacent tracks is then corrected based on the occlusion between structural components and the posture extension capability of the inspection robot to generate a track spacing scheme. Step S4: Plan the track path according to the overall spatial orientation of the target inspection area, determine the maximum length of a single track, and generate a track path scheme after segmenting the track path. Step S5: Set the starting point of the track in the track path scheme, and plan the location of the replenishment node according to the endurance of the inspection robot and the operation inspection strategy. The starting point of the track and the replenishment node are preferably set in locations that are convenient for manual inspection and equipment maintenance. The track layout scheme is composed of the track spacing scheme, the track path scheme and the location of the replenishment node. Step S6: Import the track layout scheme into the three-dimensional model of the target inspection area, and perform spatial interference verification based on the posture extension capability and obstacle avoidance range of the inspection robot during the inspection process; If the spatial interference check fails, return to step S3 or step S4 to correct the track spacing scheme or track path scheme and then re-execute; If the spatial interference check passes, proceed to step S7; Step S7: Perform a full-scene coverage integrity check on the verified track layout scheme. If it is confirmed that there are no blind spots in the target inspection area, the final track full-scene coverage design scheme will be output.

2. The method for designing a full-scene coverage track for an inspection robot according to claim 1, characterized in that, In step S1, the target inspection area also preferentially includes areas that are inaccessible to humans or pose a risk of high-altitude operations, and areas accessible to humans are used as auxiliary inspection coverage areas, which are staggered from the manual inspection paths.

3. The method for designing a full-scene coverage track for an inspection robot according to claim 1, characterized in that, In step S2, the establishment of the detection coverage radius model also incorporates the attenuation factor of image acquisition accuracy as distance increases. The attenuation factor is the change in imaging resolution of the detection module at different detection distances, to obtain the maximum effective detection distance. The maximum coverage radius is the maximum effective detection distance that meets the requirements for disease identification accuracy.

4. The method for designing a full-scene coverage track for an inspection robot according to claim 1, characterized in that, In step S3, the method for correcting the spacing between adjacent tracks is to reduce the spacing between adjacent tracks in areas with dense structural components or severe occlusion to eliminate the detection blind spot caused by occlusion; and to increase the spacing between adjacent tracks in areas with sparse structural components and open space, provided that the maximum coverage radius is not exceeded. The method for determining the spacing between adjacent tracks also includes discretizing the surface of the target inspection component in the target inspection area into multiple sampling points. Based on the imaging resolution of the detection module under different combinations of detection distance and zoom magnification, a functional relationship is established between the number of pixels per unit area of ​​the target surface and the detection distance and zoom magnification. The minimum number of pixels corresponding to the minimum identifiable size of the target defect is used as a threshold to obtain a set of feasible detection distances that meet the defect identification accuracy requirements. The set of feasible detection distances is then mapped to a detectable spatial region surrounding each target inspection component. For each detection posture combination, the visibility of each sampling point is determined by the ray casting algorithm. Based on the condition that the sampling point is in the detectable space area and is visible, the effective coverage number of each target inspection component is statistically obtained. When the effective coverage number of a target inspection component is lower than the preset lower limit, the distance between adjacent tracks in the area where the target inspection component is located is automatically reduced. When the effective coverage number of a target inspection component is higher than the preset upper limit and the corresponding imaging resolution margin is greater than the preset margin threshold, the distance between adjacent tracks in the area where the target inspection component is located is automatically enlarged.

5. The method for designing a full-scene coverage track for an inspection robot according to claim 1, characterized in that, In step S4, the determination of the maximum length of a single track also takes into account the degree of thermal expansion and contraction deformation of the track structure, the power consumption constraints of the inspection robot, inspection efficiency, and the convenience of segmented maintenance. Determining the maximum length of a single track group and segmenting the track path also includes: Based on the typical temperature conditions of the area where the building to be inspected is located, an overall finite element analysis model of the steel structure is established. The temperature deformation of the overall finite element analysis model of the steel structure is calculated to obtain the displacement of the track support nodes under each temperature condition and mapped to the track centerline to obtain the spatial deformation distribution of the track under each temperature condition. The spatial deformation distribution includes longitudinal expansion, vertical deflection and planar curvature distribution. The throughput parameters of the inspection robot when it runs on the track are obtained. The throughput parameters include at least the robot's allowable track gradient change rate, allowable track local curvature, and allowable track gauge change range. By comparing the spatial deformation distribution of the track under various temperature conditions with the throughput parameters, the track positions where any condition is not met under any temperature condition are marked as potential segmentation positions, forming a preliminary set of segmentation positions. In the initial segmentation location set, expansion compensation nodes or support adjustment nodes are set so that the spatial deformation distribution of each single track under all temperature conditions does not exceed the range of the inspection robot's throughput parameters, and the maximum length of a single track is determined.

6. The method for designing a full-scene coverage track for an inspection robot according to claim 1, characterized in that, In step S5, the starting point of the track and the recharge node are planned in a location that facilitates manual inspection and equipment maintenance; the distance between adjacent recharge nodes is not greater than the maximum continuous running distance of the inspection robot when it is fully charged.

7. The method for designing a full-scene coverage track for an inspection robot according to claim 1, characterized in that, In step S6, the posture expansion of the inspection robot includes the extension, rotation, or pitching motion of the detection module. The spatial interference verification verifies whether the detection range of the inspection robot under different postures can be effectively covered without structural collision by simulating the detection range of the inspection robot in the three-dimensional model of the target inspection area.

8. The method for designing a full-scene coverage track for an inspection robot according to claim 1, characterized in that, In step S7, the full-scene coverage integrity verification is completed through three-dimensional model simulation or digital twin method, and the target inspection area is checked one by one to confirm that each key inspection target is within at least one maximum coverage radius.

9. A design device for full-scene coverage of inspection robot tracks, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke executable instructions stored in the memory to execute the inspection robot track full-scene coverage design method according to any one of claims 1 to 8.

10. A readable storage medium, characterized in that, The readable storage medium stores a computer program, which, when executed by a processor, implements a full-scene coverage design method for an inspection robot track as described in any one of claims 1 to 8.

Images