Wall-climbing robot-based facade structure feature beacon creation method, apparatus and device, and medium

Abstract

The invention relates to the field of wall-climbing robot control, in particular to a wall-climbing robot-based facade structure feature beacon creation method, device and equipment and a medium. A multi-source detection module is used for collecting a visual image of the surface of the facade and an echo signal in the facade; obtaining an effective target and an initial position of the effective target from the visual image and the echo signal through an edge calculation module; controlling the wall-climbing robot to move to a target alignment position according to the initial position, and calculating to obtain an accurate coordinate of the effective target in a preset local coordinate system based on a current pose of the wall-climbing robot and a visual ranging result; a characteristic beacon pattern with unique coding information is formed on a vertical face position corresponding to the accurate coordinate through spraying by using a spraying and laying module; and the attribute information of the effective target and the unique coding information of the feature beacon are stored in an associated manner, so that the digital identity information is converted into a long-term stable physical graphic mark, and the overall operation efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of wall-climbing robot control, and specifically to a method, apparatus, equipment, and medium for creating facade structural feature beacons based on wall-climbing robots. Background Technology

[0002] Safety inspection of facade structures (such as large concrete buildings, industrial storage tanks, ship hulls, bridge piers, etc.) is a long-standing and critical need in civil engineering, petrochemical industry, municipal infrastructure and other fields. Traditional inspection methods mainly rely on manual scaffolding or the use of suspended baskets for close-range visual inspection or handheld equipment inspection. These methods have inherent defects such as low efficiency, high risk, limited coverage and poor adaptability to complex curved surfaces.

[0003] With the development of robotics technology, wall-climbing robots are gradually being applied to facade inspection tasks to replace high-risk manual inspections.

[0004] Existing wall-climbing robots can identify defect targets in facade inspection operations, but the defect location information relies on the robot's own positioning system, which is susceptible to environmental interference and error accumulation. It is impossible to accurately place corresponding physical markers for each identified defect target on site, making it difficult to quickly and accurately reproduce the target location in subsequent re-inspection and maintenance operations. Overall, the operation efficiency is low and the reliability is insufficient. Summary of the Invention

[0005] The purpose of this invention is to provide a method, apparatus, equipment, and medium for creating facade structural feature beacons based on a wall-climbing robot, which solves the problems in the prior art.

[0006] This invention is achieved through the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a method for creating facade structural feature beacons based on a wall-climbing robot, applied to a wall-climbing robot, the wall-climbing robot including a multi-source detection module, an edge computing module, and a spraying and deployment module, the method comprising:

[0008] The multi-source detection module acquires visual images of the facade surface and echo signals from inside the facade. The multi-source detection module includes at least a visual detection unit and a ground-penetrating radar unit.

[0009] The edge computing module obtains the effective target and the preliminary position of the effective target from the visual image and echo signal. The effective target is a target with a preset defect on the facade surface that requires the creation of a beacon.

[0010] Based on the initial position, the wall-climbing robot is controlled to move to the target alignment position. Based on the current pose of the wall-climbing robot and the visual ranging result, the precise coordinates of the effective target in the preset local coordinate system are calculated.

[0011] The wall-climbing robot is controlled to move to the spraying operation position according to the precise coordinates, and the spraying deployment module is used to spray a feature beacon pattern with unique coded information on the vertical position corresponding to the precise coordinates.

[0012] The attribute information of the effective target is associated with the unique encoding information of the feature beacon and stored together. The attribute information includes at least the target type and the precise coordinates.

[0013] Preferably, the step of acquiring visual images of the facade surface and echo signals from inside the facade through the multi-source detection module includes:

[0014] Based on the facade material, select and activate the corresponding adsorption module to enable the wall-climbing robot to stably adhere to the facade.

[0015] The control module of the wall-climbing robot drives the wall-climbing robot to move along the facade along a preset path, and at the same time activates the multi-source detection module;

[0016] During the movement, the high-definition industrial camera of the multi-source detection module captures images of the facade surface in real time as visual images;

[0017] The ground-penetrating radar unit of the multi-source detection module synchronously transmits electromagnetic waves and receives reflected signals from inside the facade as echo signals.

[0018] The acquired visual images and echo signals are transmitted to the edge computing module in real time for processing.

[0019] Preferably, obtaining the effective target and the preliminary position of the effective target from the visual image and the echo signal includes:

[0020] The edge computing module processes the visual image using an improved deep learning algorithm to obtain the identification results of facade surface defects and their corresponding identification confidence levels.

[0021] By performing wavelet denoising and signal analysis on the echo signal, the difference signal of the internal structure of the facade is obtained;

[0022] Targets whose confidence level is not lower than the first preset threshold, or whose difference between the difference signal and the normal region signal is not lower than the second preset threshold, are determined to be valid targets.

[0023] Based on the location of the effective target, determine the preliminary location of the effective target.

[0024] Preferably, the step of controlling the wall-climbing robot to move to the target alignment position based on the initial position, and calculating the precise coordinates of the effective target in a preset local coordinate system based on the current pose and visual ranging results of the wall-climbing robot, includes:

[0025] The control module in the wall-climbing robot drives the movement and adsorption modules of the wall-climbing robot according to the initial position, so that the wall-climbing robot moves to the front of the effective target area.

[0026] Adjust the posture of the wall-climbing robot so that the visual detection unit in the multi-source detection module is aligned with the center of the effective target;

[0027] The visual detection unit acquires close-up images of the target.

[0028] A binocular vision stereo matching algorithm is used to calculate the relative distance between the center of the effective target and the wall-climbing robot based on the close-up image of the target;

[0029] Based on the current pose data obtained by the wall-climbing robot from the inertial measurement unit and the odometer, the precise coordinates of the effective target in the preset local coordinate system are calculated using a coordinate transformation formula. The preset local coordinate system is established with the starting point of the operation as the origin.

[0030] Preferably, the step of using the spraying deployment module to spray a feature beacon pattern with unique coded information on the facade position corresponding to the precise coordinates includes:

[0031] The coding generation unit in the spraying and application module assigns a unique binary identification code to the current valid target.

[0032] The binary recognition code is converted into a preset visual feature pattern, which is a positioning pattern that includes the coding area and the positioning corner point;

[0033] Based on the precise coordinates, the control module drives the high-pressure spray gun of the spraying and application module to spray the weather-resistant coating onto the facade according to the visual feature pattern.

[0034] Preferably, before the high-pressure spray gun of the driving spraying module, it further includes:

[0035] The control module dynamically adjusts the spraying pressure and paint atomization particle size of the high-pressure spray gun according to the facade material;

[0036] The spray angle of the high-pressure spray gun is adjusted to be perpendicular to the vertical surface by the angle adjustment mechanism of the spraying layout module.

[0037] Preferably, the method further includes:

[0038] The visual detection unit of the wall-climbing robot captures images of the feature beacons after spraying;

[0039] The edge computing module identifies the corner points in the feature beacon image and calculates the beacon center position.

[0040] Verify the alignment error between the beacon center position and the precise coordinates. If the error exceeds the preset tolerance, initiate a secondary spray adjustment.

[0041] Secondly, embodiments of the present invention provide a facade structure feature beacon creation device based on a wall-climbing robot, applied to a wall-climbing robot, the wall-climbing robot including a multi-source detection module, an edge computing module, and a spraying and deployment module, the device including:

[0042] The data acquisition module is used to acquire visual images of the facade surface and echo signals inside the facade through the multi-source detection module. The multi-source detection module includes at least a visual detection unit and a ground-penetrating radar unit.

[0043] The target determination module is used to obtain the effective target and the preliminary position of the effective target from the visual image and echo signal through the edge computing module. The effective target is a target with a preset defect on the facade surface that requires the creation of a beacon.

[0044] The precise coordinate module is used to control the wall-climbing robot to move to the target alignment position according to the preliminary position, and to calculate the precise coordinates of the effective target in the preset local coordinate system based on the current pose and visual ranging results of the wall-climbing robot.

[0045] The spraying beacon module is used to control the wall-climbing robot to move to the spraying operation position according to the precise coordinates, and to use the spraying deployment module to spray a feature beacon pattern with unique coding information on the vertical position corresponding to the precise coordinates.

[0046] An association storage module is used to associate and store the attribute information of the valid target with the unique encoding information of the feature beacon, wherein the attribute information includes at least the target type and the precise coordinates.

[0047] Thirdly, embodiments of the present invention provide an electronic device, including: at least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method of the first aspect described above.

[0048] Fourthly, embodiments of the present invention provide a storage medium storing computer program instructions, which, when executed by a processor, implement the method of the first aspect described above.

[0049] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0050] By integrating visual images and ground-penetrating radar echo signals as dual data sources, and utilizing locally deployed edge computing modules for concurrent processing and fusion judgment, the system can comprehensively identify defect targets by combining surface morphology and internal structure information, reducing the possibility of missed detections or misjudgments by a single sensor. This process is completed automatically, outputting a list of tasks to be processed with spatial orientation, laying the foundation for subsequent precise operations. By controlling the robot to approach the target and using technologies such as binocular vision for close-range stereo ranging, combined with the robot's own real-time high-frequency pose data, the local position of the target relative to the robot is converted into precise coordinates in a global coordinate system with the operation starting point as the origin. This method bypasses the drawbacks of simply relying on the cumulative error of robot navigation mileage, providing an accurate mathematical description for the reliable recording of the target position. By generating a unique binary code for each target and converting it into a machine vision-optimized pattern containing positioning corner points and coded areas, and then controlling a high-pressure spray gun with parameters adapted to the facade material to precisely apply weather-resistant paint according to the pattern, a physical beacon with both identification and visual guidance functions is created next to the defect. This establishes a direct, stable, and machine-readable binding relationship between defect information and physical location. By immediately capturing images of the markings after spraying, identifying the positioning corners, and calculating the actual positions, the system compares these actual positions with the expected coordinates to automatically determine the accuracy of the marking position. When the deviation exceeds the limit, a second round of re-spraying is automatically triggered for adjustment. This mechanism constitutes an execution-verification-correction quality control loop, reducing the risk of errors in a single operation and ensuring the reliability and consistency of the markings. Attached Figure Description

[0051] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0052] Figure 1 This is a flowchart illustrating the method for creating facade structural feature beacons based on a wall-climbing robot provided by the present invention.

[0053] Figure 2 A schematic diagram of the facade structural feature beacon creation device based on a wall-climbing robot provided by the present invention;

[0054] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0056] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0057] It should be noted that all actions involving the acquisition of signals, information, or data in this invention are carried out in compliance with the relevant data protection laws and regulations of the locality and with authorization from the owner of the relevant device.

[0058] Example 1

[0059] This invention provides a wall-climbing robot that integrates a movement and adsorption module, a multi-source detection module, an edge computing module, a spraying and deployment module, a control module, and a data storage and communication unit. These modules work together to achieve the core process of facade target detection → feature beacon spraying and deployment. The core functions of each module are as follows:

[0060] The moving and adsorption module adopts magnetic adsorption wheels or negative pressure adsorption chassis (adaptable to different facade materials such as concrete and metal), and has curvature self-adaptation capability, which can realize safe movement, posture adjustment and stable docking of the facade, providing a basis for inspection and spraying.

[0061] Multi-source detection module: Integrates high-definition industrial cameras (visual inspection), ground penetrating radar (internal defect detection) and other multi-source detection modules to collect facade surface image data and internal structure echo data in real time. The data sampling frequency is not less than 30Hz to ensure complete capture of target information.

[0062] Edge computing module: Equipped with a lightweight processor, it has an improved target detection algorithm and data fusion logic built in, which can process multi-source detection data in real time and complete target recognition and detection effectiveness verification;

[0063] Spraying deployment module: includes a weather-resistant paint storage tank, a high-pressure spray gun, an angle adjustment mechanism and a coding generation unit. The spray gun can achieve 0-90° angle fine adjustment. The paint is a special waterproof spraying material with high contrast, UV resistance and strong adhesion.

[0064] Control module: The core controller is an STM32H7 series microcontroller, which is responsible for coordinating the timing of each module, receiving control commands from the edge computing module, and driving the movement, adsorption and spraying modules to perform actions.

[0065] Data storage and communication unit: The built-in SD card stores target information and beacon code associated data, and the WiFi module enables real-time data transmission with the ground workstation.

[0066] Please see Figure 1 This invention provides a method for creating facade structural feature beacons based on a wall-climbing robot, applicable to a wall-climbing robot. The wall-climbing robot includes a multi-source detection module, an edge computing module, and a spraying and deployment module. The method includes:

[0067] S1. Visual images of the facade surface and echo signals inside the facade are acquired through the multi-source detection module. The multi-source detection module includes at least a visual detection unit and a ground-penetrating radar unit.

[0068] Specifically, a wall-climbing robot is an automated device that can move and perform tasks on vertical or inclined surfaces.

[0069] The multi-source detection module is a data acquisition system integrated into the wall-climbing robot. Its function is to simultaneously acquire physical information of different dimensions of the target area.

[0070] A visual image is a two-dimensional digital image of the shape, color, and texture of a facade surface captured by a visual inspection unit (such as an optical camera). It reflects the visible features of the surface, such as a crack appearing as an area of ​​a specific shape and grayscale in the image.

[0071] Echo signals refer to the waveform data received by the internal structural interface after electromagnetic waves are emitted into the interior of the facade by a ground penetrating radar unit. This data contains information about the discontinuities within the material. For example, voids inside concrete can cause specific changes in the amplitude and time characteristics of the echo signal.

[0072] To achieve comprehensive defect perception, the system first drives a wall-climbing robot to move across the facade, simultaneously activating its onboard multi-source detection module. The visual inspection unit continuously captures surface images of the facade at a certain frame rate, while the ground-penetrating radar unit simultaneously emits electromagnetic pulses of a specific frequency towards the facade and receives the reflected signals. These two types of data are collected and correlated synchronously in time and space, together forming a preliminary characterization of the detection area from the surface to the shallow interior. This step, by fusing sensor data based on different physical principles, provides a richer and more reliable information foundation for subsequent analysis than a single sensing mode, helping to reduce missed detections caused by surface occlusion or the limitations of a single sensor.

[0073] In some embodiments, S1, acquiring visual images of the facade surface and echo signals from inside the facade through the multi-source detection module, includes:

[0074] Based on the facade material, select and activate the corresponding adsorption module to enable the wall-climbing robot to stably adhere to the facade.

[0075] The control module of the wall-climbing robot drives the wall-climbing robot to move along the facade along a preset path, and at the same time activates the multi-source detection module;

[0076] During the movement, the high-definition industrial camera of the multi-source detection module captures images of the facade surface in real time as visual images;

[0077] The ground-penetrating radar unit of the multi-source detection module synchronously transmits electromagnetic waves and receives reflected signals from inside the facade as echo signals.

[0078] The acquired visual images and echo signals are transmitted to the edge computing module in real time for processing.

[0079] Specifically, system initialization and job preparation:

[0080] (1) Select the climbing robot model according to the facade material. For example, magnetic adsorption is used for metal storage tanks and ships, and negative pressure adsorption is used for concrete structures. The operator places the climbing robot at the starting point of the facade operation, issues the operation command through the ground workstation, and the control module starts the system self-check to confirm that the functions of each module (detection, spraying, movement) are normal.

[0081] (2) The movement and adsorption module is activated to ensure that the robot is stably adsorbed on the facade;

[0082] (3) The control module drives the robot to move to the starting position of the operation. The initial environmental image is collected by the camera in the multi-source detection module. Combined with the IMU (Inertial Measurement Unit) and odometer data, a local coordinate system of the operation area is established (with the starting point as the origin, the horizontal direction of the elevation as the X-axis, and the vertical direction as the Y-axis), which provides a reference for subsequent target positioning.

[0083] Facade movement and multi-source data acquisition:

[0084] (1) The control module drives the movement and adsorption module to move the robot along the facade at a constant speed according to the preset operation path. The movement speed is controlled at 0.1-0.3m / s to ensure that no detection data is missed.

[0085] (2) During the movement, the multi-source detection module works synchronously: the high-definition industrial camera captures images of the facade surface in real time (resolution not less than 1920×1080), the ground penetrating radar emits low-frequency electromagnetic waves and receives echo signals from the internal structure, and the two types of data are transmitted to the edge computing module in real time with a transmission delay of no more than 50ms.

[0086] (3) The control module receives odometry and IMU data in real time, updates the robot’s real-time pose (position coordinates and attitude angle) in the local coordinate system, and synchronizes the pose data to the edge computing module to provide a spatial reference for data association.

[0087] S2. The edge computing module obtains the effective target and the preliminary position of the effective target from the visual image and echo signal. The effective target is a target with a preset defect on the facade surface that requires the creation of a beacon.

[0088] Specifically, the edge computing module is a computing unit deployed locally on the robot, responsible for real-time processing and analysis of the collected raw data.

[0089] A valid target refers to a defective object that has been determined by the algorithm to actually exist and needs further processing, such as a crack whose width exceeds a threshold and is identified according to a preset standard.

[0090] Preliminary position refers to the initial estimate of the spatial location of the defect target in a temporary frame of reference established by the robot itself. It is usually expressed as the approximate direction and distance relative to the robot's current position.

[0091] In this step, the edge computing module receives visual images and echo signal streams from the multi-source detection module. It first processes the visual images, using image recognition algorithms to analyze whether there are regions matching preset defect characteristics and estimating their geometric contours and confidence levels. Simultaneously, it analyzes the echo signals, extracting their time-frequency features and comparing them with a baseline signal from a normal area to determine if any anomalies exist. Combining the results of both analyses, the system determines whether the target is a valid target to be marked based on preset decision logic (e.g., the image recognition confidence level reaches a certain standard, or the echo signal contrast difference exceeds a certain threshold). Once confirmed, the system combines the pixel coordinates of the target region in the image with the robot's approximate pose when acquiring the frame data to calculate the target's preliminary position in the temporary coordinate system of the current operation stage. This step transforms the raw data into a task with a clear spatial orientation, achieving a crucial transition from perception to decision-making and providing target guidance for subsequent precise operations.

[0092] In some embodiments, obtaining the effective target and the preliminary position of the effective target from the visual image and echo signal includes:

[0093] The edge computing module processes the visual image using an improved deep learning algorithm to obtain the identification results of facade surface defects and their corresponding identification confidence levels.

[0094] By performing wavelet denoising and signal analysis on the echo signal, the difference signal of the internal structure of the facade is obtained;

[0095] Targets whose confidence level is not lower than the first preset threshold, or whose difference between the difference signal and the normal region signal is not lower than the second preset threshold, are determined to be valid targets.

[0096] Based on the location of the effective target, determine the preliminary location of the effective target.

[0097] Specifically, this step describes the logical flow of how the edge computing module fuses visual and radar data to determine and locate valid targets. First, the edge computing module applies an improved deep learning-based target detection algorithm to the acquired visual images. This algorithm outputs the possible defect categories in the image and their corresponding region bounding boxes, while assigning a recognition confidence score between 0 and 1 to each recognition result. This score represents the algorithm's assessment probability of the recognition result's correctness. Second, the module preprocesses and performs feature analysis on the synchronously acquired echo signals: wavelet denoising filters out environmental noise and random interference, while signal analysis extracts time-domain or frequency-domain features that reflect the internal structural state. The features of the target region are compared with a pre-learned or real-time acquired baseline of normal region features to obtain a quantified difference signal intensity value. The determination process employs a fusion decision rule: if a region's visual recognition confidence reaches a preset first threshold (e.g., 0.9), its surface defect is considered significant; or, if the difference intensity of its internal echo signal reaches a preset second threshold, its internal structure is considered abnormal. Regions meeting either condition are determined to be valid targets requiring labeling. Finally, for each valid target, the system calculates the approximate direction and distance of the target relative to the robot based on the original visual image frame in which it is located and the rough estimate of the robot's pose during image acquisition, combined with the camera imaging model, thereby determining its initial position in the temporary working coordinate system and providing initial guidance for subsequent precise positioning.

[0098] Furthermore, the edge computing module performs target detection and verification:

[0099] (1) Image data processing: The edge computing module uses an improved deep learning algorithm to extract features from the real-time acquired surface images, identify surface targets such as cracks, corrosion, and chipping, and output the preliminary position, outline and size information of the targets (such as crack length and width).

[0100] (2) Internal defect detection: The edge computing module uses wavelet denoising algorithm to filter and analyze the ground penetrating radar echo data to determine whether there are internal cavities, steel corrosion and other defects under the target;

[0101] (3) Target validity confirmation: The edge computing module sets two verification rules - when the confidence level of surface image recognition is ≥90%, or when the ground penetrating radar signal shows that there is a significant difference between the target area and the normal area, it is determined to be a valid target.

[0102] S3. Based on the preliminary position, control the wall-climbing robot to move to the target alignment position. Based on the current pose of the wall-climbing robot and the visual ranging result, calculate the precise coordinates of the effective target in the preset local coordinate system.

[0103] Specifically, the target alignment position refers to the docking point where the robot, after being moved and adjusted, is positioned so that its detection equipment can perform precise measurements of the target with better geometric relationships.

[0104] Current pose refers to the real-time position and orientation of a robot as described in its navigation system (such as a system that integrates odometry and inertial measurement data).

[0105] Visual ranging results refer to the straight-line distance from the robot's visual sensor to a specific point on the target, calculated using visual methods (e.g., utilizing binocular parallax or prior knowledge of the target size).

[0106] A preset local coordinate system is a two-dimensional or three-dimensional coordinate framework established for the convenience of this operation. It usually takes the starting point of the operation or a fixed reference point as the origin.

[0107] Precise coordinates refer to a set of highly accurate coordinate values ​​that describe the position of a target point within a preset local coordinate system.

[0108] Based on the preliminary position obtained in the previous step, the control system drives the robot to move, enabling its onboard vision sensors to align more directly and closely with the target area. Once the robot reaches and stabilizes at the target alignment position, the system records the robot's pose obtained through its own navigation system. Subsequently, a specialized visual ranging program (e.g., stereoscopic imaging analysis of the target area) calculates the accurate distance from the sensor to the target point (e.g., the center of the defect). Finally, combining the robot's current pose (including position and orientation) with the distance and angle information obtained from visual ranging, the precise coordinates of the target point in a preset local coordinate system are calculated using a coordinate transformation formula. This step elevates the initial estimate based on coarse navigation to precise positioning based on real-time direct measurement, providing accurate spatial anchor points for physical marking.

[0109] In some embodiments, the step of controlling the wall-climbing robot to move to the target alignment position based on the initial position, and calculating the precise coordinates of the effective target in a preset local coordinate system based on the current pose and visual ranging results of the wall-climbing robot, includes:

[0110] The control module in the wall-climbing robot drives the movement and adsorption modules of the wall-climbing robot according to the initial position, so that the wall-climbing robot moves to the front of the effective target area.

[0111] Adjust the posture of the wall-climbing robot so that the visual detection unit in the multi-source detection module is aligned with the center of the effective target;

[0112] The visual detection unit acquires close-up images of the target.

[0113] A binocular vision stereo matching algorithm is used to calculate the relative distance between the center of the effective target and the wall-climbing robot based on the close-up image of the target;

[0114] Based on the current pose data obtained by the wall-climbing robot from the inertial measurement unit and the odometer, the precise coordinates of the effective target in the preset local coordinate system are calculated using a coordinate transformation formula. The preset local coordinate system is established with the starting point of the operation as the origin.

[0115] Specifically, this step describes the complete control and calculation process from initial positioning to obtaining precise coordinates. First, based on the initial spatial clue of the preliminary position, the control module generates control commands to drive the robot's movement mechanism (such as wheels or tracks) and adsorption mechanism (such as magnets or vacuum devices) to work together, enabling the robot to move directly in front of the target area, completing the initial approach positioning. Next, the robot adjusts the angle of its joints or gimbal to change the orientation of its onboard vision detection unit (such as a camera), so that the center of the unit's visual axis is aligned as closely as possible with the target to be identified (such as the center point of a crack). This posture adjustment aims to establish the optimal observation geometry for subsequent precise measurements.

[0116] Once the robot stabilizes in the alignment posture, the vision detection unit is triggered to acquire one or more frames of close-up images focused on the target area, i.e., target close-up images. These images contain richer target details compared to the wide-angle images taken during inspection. Subsequently, the system processes the target close-up images using a binocular vision stereo matching algorithm. The principle of this algorithm is based on the disparity information of the same scene obtained by two cameras with known relative positions (or two images taken by a single camera at different positions). By matching the pixel positions of the target point in the two images, combined with camera intrinsic parameters and baseline distance, the three-dimensional spatial coordinates of the target point relative to the camera coordinate system can be calculated using the principle of triangulation. This yields the relative distance and direction between the target center and the vision sensor on the robot.

[0117] After obtaining the precise relative position (i.e., relative distance and azimuth) of the target relative to the robot, it needs to be transformed into a global, stable reference frame. The system reads the robot's current pose data, calculated by fusing the inertial measurement unit (IMU, providing attitude angular velocity and acceleration) and the odometry (providing displacement increments). This data describes the robot's real-time position and orientation in a preset local coordinate system. Finally, using a standard coordinate transformation formula, the local three-dimensional coordinates of the target relative to the robot are superimposed and fused with the robot's global pose, undergoing rotation and translation transformations to calculate the final precise coordinates of the effective target's center point in the preset local coordinate system with the operation start point as the origin. This series of operations achieves the transformation from a coarse area indication to the target point's centimeter-level or millimeter-level absolute coordinates.

[0118] Furthermore, the initial target positioning includes:

[0119] (1) The control module receives the effective target preliminary position information output by the edge computing module, drives the robot to move to the front of the target area, and adjusts the posture so that the multi-source detection module is aligned with the target center;

[0120] (2) A high-definition industrial camera captures a close-up image of the target. The relative distance between the target center and the robot is calculated using a binocular vision stereo matching algorithm. Combined with the robot's current pose data in the local coordinate system, the precise coordinates (X, Y, Z) of the target center are calculated using the following coordinate transformation formula. t ,Y t ):

[0121] X t =Xᵣ+d×cosθ;Y t =Yᵣ+d×sinθ;

[0122] Where (Xᵣ,Yᵣ) are the robot's current coordinates, d is the relative distance between the target and the robot, and θ is the angle between the robot and the target (obtained from IMU data).

[0123] S4. Control the wall-climbing robot to move to the spraying operation position according to the precise coordinates, and use the spraying deployment module to spray a feature beacon pattern with unique coding information on the vertical position corresponding to the precise coordinates.

[0124] Specifically, the spraying operation location refers to the docking point that the robot eventually arrives at in order to perform the spraying action, maintaining a specific working distance from the target point.

[0125] The spraying and application module is the actuator on the robot responsible for storing paint and applying it to the facade according to a predetermined pattern.

[0126] Feature beacon patterns refer to two-dimensional graphics with specific visual structures and coded information that are formed by spraying. They are designed to be easily detected, identified, and located by subsequent vision systems.

[0127] After obtaining the precise coordinates of the target, the control system plans a path to drive the robot to a suitable relative position for the spraying operation, i.e., the spraying work location. Upon arrival, the spraying deployment module is activated. The system first calculates the desired application point of the nozzle based on the precise coordinate information, possibly combined with the geometric characteristics of the facade surface. Then, the spraying deployment module is controlled to spray a specific coating (such as high-contrast, weather-resistant coating) onto the facade in an atomized form. The spraying process is not random smearing, but controlled according to a pre-generated digital pattern template. This template contains unique encoded information (e.g., a binary ID represented by a specific arrangement of graphic elements), thereby forming a characteristic beacon pattern with identification function at the target location on the facade. This step transforms the coordinates and encoded information determined in the digital world into a stable physical marker attached to the defect in the physical world, establishing a reliable and intuitive connection between virtual information and physical location.

[0128] In some embodiments, the step of using the spraying deployment module to spray a feature beacon pattern with unique coded information on the facade position corresponding to the precise coordinates includes:

[0129] The coding generation unit in the spraying and application module assigns a unique binary identification code to the current valid target.

[0130] The binary recognition code is converted into a preset visual feature pattern, which is a positioning pattern that includes the coding area and the positioning corner point;

[0131] Based on the precise coordinates, the control module drives the high-pressure spray gun of the spraying and application module to spray the weather-resistant coating onto the facade according to the visual feature pattern.

[0132] Specifically, this step defines the concrete implementation of physical marking after precise positioning. First, the system invokes the encoding generation unit integrated into the spraying and deployment module. This unit generates a unique identifier for the currently processed valid target based on certain encoding rules (such as sequential incrementing or a specific algorithm). This identifier is typically represented internally by a binary data string. Subsequently, the system maps this binary recognition code onto a pre-designed graphic template, generating a corresponding, machine-readable visual feature pattern. This pattern is a graphic optimized for visual recognition, containing two main functional areas: one is the encoding area, where the arrangement of graphic elements directly represents the binary encoded information; the other is the positioning corner point, composed of graphics of specific shapes and positions, used to assist the vision system in quickly and stably detecting and positioning the entire pattern's position and orientation within the image. Common positioning patterns of this type are adaptive improvements based on open-source visual benchmark markers such as Apriltag.

[0133] After the pattern data is ready, the control module enters the execution phase. Based on the previously calculated precise coordinates of the effective target, combined with parameters such as the robot's current pose, the installation relationship between the spray gun and the robot body, and the working distance, it calculates the absolute spatial position and spray angle that the spray gun nozzle needs to point to. Then, the control module drives the execution mechanism of the spraying deployment module to adjust the pose of the high-pressure spray gun to align it with the target point and start the spraying process. The high-pressure spray gun sprays the weather-resistant paint (a special paint with the ability to resist environmental factors such as ultraviolet rays, temperature changes, and wind and rain erosion) in the storage tank onto the facade in an atomized form. The spraying trajectory is not arbitrary, but precisely controlled by the control module, so that the adhesion mark formed by the sprayed paint at the specified coordinate position on the facade is completely consistent with the previously generated digital visual feature pattern, thereby directly creating a stable physical graphic mark containing unique encoded information next to the defective target.

[0134] In some embodiments, the high-pressure spray gun of the driving spraying module is preceded by:

[0135] The control module dynamically adjusts the spraying pressure and paint atomization particle size of the high-pressure spray gun according to the facade material;

[0136] The spray angle of the high-pressure spray gun is adjusted to be perpendicular to the vertical surface by the angle adjustment mechanism of the spraying layout module.

[0137] Specifically, this step describes the parameter pre-adjustment and attitude calibration steps performed before the spraying action, with the aim of ensuring that qualified markings are formed under different operating conditions. First, the control module dynamically sets the key operating parameters of the high-pressure spray gun based on the judgment of the material of the working surface (e.g., distinguishing between concrete, metal, and tile through preset task parameter input or real-time identification based on sensor data). Specifically, it adjusts two main parameters: one is the spraying pressure, which is the gas or hydraulic pressure value that drives the paint to spray out. Different material surfaces (such as rough concrete or smooth metal) require different pressures to ensure effective paint adhesion without splashing or running; the other is the paint atomization particle size, which is the average size of the paint atomized into droplets. This affects the uniformity of the coating, drying speed, and adhesion, and is usually achieved by adjusting the atomizing air pressure of the spray gun or the nozzle structure.

[0138] Secondly, to ensure the resulting pattern is clear, undistorted, and has optimal adhesion and visual recognition, the system activates the angle adjustment mechanism on the spraying module. This mechanism may consist of a motor-driven joint or a linear module. The control module calculates the theoretical angle at which the central axis of the high-pressure spray gun nozzle should align with the normal direction (i.e., the vertical direction) of the vertical surface, given the current robot pose and target point coordinates. Then, it drives the angle adjustment mechanism to precisely adjust the high-pressure spray gun's posture, aligning its actual spray angle with the calculated theoretical vertical angle. This vertical spraying posture minimizes paint dripping, accumulation, or graphic distortion caused by tilted spraying, ensuring accurate shape and clear edges of the sprayed visual feature pattern, thereby guaranteeing reliable identification and positioning of the beacon pattern by the subsequent vision system.

[0139] For example, the feature beacon is spray-painted and deployed:

[0140] (1) Spraying parameter adaptation: The control module automatically adjusts the spraying module parameters according to the target type and facade material. For rough surfaces such as concrete, the spraying pressure is set to 0.3-0.5MPa; for smooth surfaces such as metal, the spraying pressure is set to 0.2-0.3MPa, and the paint atomization particle size is controlled at 50-100μm to ensure paint adhesion.

[0141] (2) Posture and angle adjustment: The control module drives the robot to move to the spraying position (10-15cm in front of the center of the target) according to the final coordinates of the target. The spraying angle of the spray gun is adjusted by the angle adjustment mechanism to make the spraying direction perpendicular to the vertical surface and avoid paint dripping.

[0142] (3) Feature beacon generation and spraying: Weather-resistant fluorescent acrylic paint is used. The spraying pressure (0.2-0.5MPa) and atomization particle size (50-100μm) are dynamically adjusted according to the facade material (concrete / metal). The beacon pattern is based on Apriltag and is integrated with a unique coding area and positioning corner point (size 5×5cm). The high-pressure spray gun is started and sprayed at a uniform speed according to the generated feature pattern. The robot remains stationary during the spraying process. After the spraying is completed, it stays for 2 seconds to ensure that the paint is initially cured.

[0143] (4) Beacon and target alignment verification: After the spraying is completed, the high-definition industrial camera takes a beacon image, the edge computing module identifies the beacon positioning corner point, and confirms that the alignment error between the beacon center and the target center is ≤ ±3mm. If the error exceeds the standard, a second spraying adjustment is started.

[0144] Data association, storage, and communication:

[0145] (1) The control module associates the key information of the effective target (type, size, precise coordinates, detection time) with the unique code of the corresponding beacon and the spraying parameters and stores it on the local SD card. The data format adopts the JSON standard format and is transmitted to the ground workstation in real time through the WiFi module.

[0146] (2) After receiving the data, the ground workstation generates a “target-beacon” association database to provide basic data support for subsequent UAV scanning modeling (obtaining global coordinates of the beacon) and robot re-inspection navigation. The data storage format adopts the JSON standard format to ensure compatibility.

[0147] The method further includes:

[0148] The visual detection unit of the wall-climbing robot captures images of the feature beacons after spraying;

[0149] The edge computing module identifies the corner points in the feature beacon image and calculates the beacon center position.

[0150] Verify the alignment error between the beacon center position and the precise coordinates. If the error exceeds the preset tolerance, initiate a secondary spray adjustment.

[0151] Specifically, this step describes the process by which the system automatically verifies and performs closed-loop correction of the marking quality after the feature beacon spraying is completed. First, as the input for verification, the wall-climbing robot uses its onboard visual detection unit—either the same camera used for initial target recognition or another dedicated camera—to acquire an image of the facade area where the feature beacon pattern has just been applied after the spraying operation is completed. This image contains the complete sprayed pattern and is called the feature beacon image.

[0152] Next, the edge computing module processes the newly acquired image. It runs a pre-defined image recognition algorithm specifically designed for fast and robust detection of the aforementioned visual feature patterns. The algorithm first searches for corner points in the image—feature points within the pattern that have high contrast, specific shapes, and are easily distinguishable from the background. Once all the required corner points are successfully detected, the algorithm calculates the precise outline of the entire feature beacon pattern in the image's pixel coordinate system based on the known geometric relationships of these corner points in the image (predefined by the pattern design) through perspective transformation or similar calculation methods. It then further determines the position of its geometric center point in the image—the calculated beacon center position.

[0153] Finally, the system performs alignment error verification. It calculates the beacon center position (pixel coordinates) from the image, combines the intrinsic parameters of the vision detection unit with the current robot pose, and performs an inverse coordinate transformation to convert it into a preset local coordinate system identical to the previously stored accurate coordinates of the valid target. Then, the system calculates the spatial distance deviation between this beacon's actual center position calculated from the image and the initially planned and expected accurate coordinates—the alignment error. The system compares this error value with a preset tolerance (e.g., 3 mm). If the calculated alignment error is less than or equal to the preset tolerance, the spraying quality is considered acceptable, and the process continues. If the alignment error exceeds the preset tolerance, it indicates an unacceptable deviation in the spraying position. In this case, the system does not immediately terminate the operation but triggers a correction process: the control module replans a fine-tuned spraying coordinate or path based on the calculated deviation vector and restarts the spraying layout module to perform secondary spraying adjustments on or near the original pattern to correct positional deviations or cover incomplete patterns, ensuring that the final physical marker position meets accuracy requirements.

[0154] S5. The attribute information of the effective target is associated with the unique encoding information of the feature beacon and stored together. The attribute information includes at least the target type and the precise coordinates.

[0155] Specifically, attribute information refers to the set of data used to describe the characteristics of a valid target itself.

[0156] Unique encoding information refers to an identifier assigned to each feature beacon that is unique within its application scope; it is usually a string of numbers or codes. Association storage refers to the process of establishing index relationships between different data items belonging to the same entity and storing them in a storage medium.

[0157] Once the feature beacon spraying is complete, the on-site physical operations for that target are finished. The system then performs a data archiving operation. It binds a series of attribute information about the valid target acquired and confirmed in the previous steps, such as its classification (e.g., cracks, corrosion), size, and the final calculated precise coordinates, with the unique coded information carried by the feature beacon sprayed at the target location. This binding relationship is structured and recorded and stored, either in the robot's local storage device or transmitted to the backend database via a wireless network. This forms a queryable mapping relationship: by scanning and identifying the beacon's code, the complete file information of the corresponding defect can be retrieved, and vice versa. This step completes a closed loop from detection, location, marking to information integration, enabling any subsequent operation to quickly and accurately retrieve all historical detection data for that location by identifying the physical beacon, improving the systematicness and consistency of the entire detection and maintenance process.

[0158] Example 2

[0159] Please see Figure 2 This invention provides a facade structural feature beacon creation device based on a wall-climbing robot, applied to a wall-climbing robot. The wall-climbing robot includes a multi-source detection module, an edge computing module, and a spraying and deployment module. The device includes:

[0160] The data acquisition module is used to acquire visual images of the facade surface and echo signals inside the facade through the multi-source detection module. The multi-source detection module includes at least a visual detection unit and a ground-penetrating radar unit.

[0161] The target determination module is used to obtain the effective target and the preliminary position of the effective target from the visual image and echo signal through the edge computing module. The effective target is a target with a preset defect on the facade surface that requires the creation of a beacon.

[0162] The precise coordinate module is used to control the wall-climbing robot to move to the target alignment position according to the preliminary position, and to calculate the precise coordinates of the effective target in the preset local coordinate system based on the current pose and visual ranging results of the wall-climbing robot.

[0163] The spraying beacon module is used to control the wall-climbing robot to move to the spraying operation position according to the precise coordinates, and to use the spraying deployment module to spray a feature beacon pattern with unique coding information on the vertical position corresponding to the precise coordinates.

[0164] An association storage module is used to associate and store the attribute information of the valid target with the unique encoding information of the feature beacon, wherein the attribute information includes at least the target type and the precise coordinates.

[0165] It should be noted that each module and unit in the facade structure feature beacon creation device based on the wall-climbing robot in this embodiment corresponds one-to-one with each step in the facade structure feature beacon creation method based on the wall-climbing robot in the aforementioned embodiment. Therefore, the specific implementation of this embodiment can refer to the implementation of the facade structure feature beacon creation method based on the aforementioned wall-climbing robot, and will not be repeated here.

[0166] Example 3

[0167] Please see Figure 3 This embodiment provides an electronic device, including at least one processor 301 and a memory 302. Optionally, the device further includes a communication component 303. The processor 301, memory 302, and communication component 303 are connected via a bus 304.

[0168] In a specific implementation, at least one processor 301 executes computer execution instructions stored in memory 302, causing at least one processor 301 to perform the above-described method.

[0169] The specific implementation process of processor 301 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0170] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0171] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0172] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0173] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0174] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0175] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0176] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0177] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0178] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0179] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0180] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0181] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0182] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for creating facade structural feature beacons based on a wall-climbing robot, characterized in that, Applied to a wall-climbing robot, the wall-climbing robot including a multi-source detection module, an edge computing module, and a spraying and deployment module, the method includes: The multi-source detection module acquires visual images of the facade surface and echo signals from inside the facade. The multi-source detection module includes at least a visual detection unit and a ground-penetrating radar unit. The edge computing module obtains the effective target and the preliminary position of the effective target from the visual image and echo signal. The effective target is a target with a preset defect on the facade surface that requires the creation of a beacon. Based on the initial position, the wall-climbing robot is controlled to move to the target alignment position. Based on the current pose of the wall-climbing robot and the visual ranging result, the precise coordinates of the effective target in the preset local coordinate system are calculated. The wall-climbing robot is controlled to move to the spraying operation position according to the precise coordinates, and the spraying deployment module is used to spray a feature beacon pattern with unique coded information on the vertical position corresponding to the precise coordinates. The attribute information of the effective target is associated with the unique encoding information of the feature beacon and stored together. The attribute information includes at least the target type and the precise coordinates.

2. The method according to claim 1, characterized in that, The acquisition of visual images of the facade surface and echo signals from inside the facade through the multi-source detection module includes: Based on the facade material, select and activate the corresponding adsorption module to enable the wall-climbing robot to stably adhere to the facade. The control module of the wall-climbing robot drives the wall-climbing robot to move along the facade along a preset path, and at the same time activates the multi-source detection module; During the movement, the high-definition industrial camera of the multi-source detection module captures images of the facade surface in real time as visual images; The ground-penetrating radar unit of the multi-source detection module synchronously transmits electromagnetic waves and receives reflected signals from inside the facade as echo signals. The acquired visual images and echo signals are transmitted to the edge computing module in real time for processing.

3. The method according to claim 1, characterized in that, Obtaining the effective target and its preliminary location from the visual image and echo signal includes: The edge computing module processes the visual image using an improved deep learning algorithm to obtain the identification results of facade surface defects and their corresponding identification confidence levels. By performing wavelet denoising and signal analysis on the echo signal, the difference signal of the internal structure of the facade is obtained; Targets whose confidence level is not lower than the first preset threshold, or whose difference between the difference signal and the normal region signal is not lower than the second preset threshold, are determined to be valid targets. Based on the location of the effective target, determine the preliminary location of the effective target.

4. The method according to claim 1, characterized in that, The step of controlling the wall-climbing robot to move to the target alignment position based on the initial position, and calculating the precise coordinates of the effective target in a preset local coordinate system based on the current pose and visual ranging results of the wall-climbing robot, includes: The control module in the wall-climbing robot drives the movement and adsorption modules of the wall-climbing robot according to the initial position, so that the wall-climbing robot moves to the front of the effective target area. Adjust the posture of the wall-climbing robot so that the visual detection unit in the multi-source detection module is aligned with the center of the effective target; The visual detection unit acquires close-up images of the target. A binocular vision stereo matching algorithm is used to calculate the relative distance between the center of the effective target and the wall-climbing robot based on the close-up image of the target; Based on the current pose data obtained by the wall-climbing robot from the inertial measurement unit and the odometer, the precise coordinates of the effective target in the preset local coordinate system are calculated using a coordinate transformation formula. The preset local coordinate system is established with the starting point of the operation as the origin.

5. The method according to claim 1, characterized in that, The process of using the spraying deployment module to spray a feature beacon pattern with unique coded information on the facade position corresponding to the precise coordinates includes: The coding generation unit in the spraying and application module assigns a unique binary identification code to the current valid target. The binary recognition code is converted into a preset visual feature pattern, which is a positioning pattern that includes the coding area and the positioning corner point; Based on the precise coordinates, the control module drives the high-pressure spray gun of the spraying and application module to spray the weather-resistant coating onto the facade according to the visual feature pattern.

6. The method according to claim 5, characterized in that, Before the high-pressure spray gun of the driving spraying deployment module, it also includes: The control module dynamically adjusts the spraying pressure and paint atomization particle size of the high-pressure spray gun according to the facade material; The spray angle of the high-pressure spray gun is adjusted to be perpendicular to the vertical surface by the angle adjustment mechanism of the spraying layout module.

7. The method according to claim 1, characterized in that, The method further includes: The visual detection unit of the wall-climbing robot captures images of the feature beacons after spraying; The edge computing module identifies the corner points in the feature beacon image and calculates the beacon center position. Verify the alignment error between the beacon center position and the precise coordinates. If the error exceeds the preset tolerance, initiate a secondary spray adjustment.

8. A facade structural feature beacon creation device based on a wall-climbing robot, characterized in that, Applied to a wall-climbing robot, the wall-climbing robot includes a multi-source detection module, an edge computing module, and a spraying and deployment module; the device includes: The data acquisition module is used to acquire visual images of the facade surface and echo signals inside the facade through the multi-source detection module. The multi-source detection module includes at least a visual detection unit and a ground-penetrating radar unit. The target determination module is used to obtain the effective target and the preliminary position of the effective target from the visual image and echo signal through the edge computing module. The effective target is a target with a preset defect on the facade surface that requires the creation of a beacon. The precise coordinate module is used to control the wall-climbing robot to move to the target alignment position according to the preliminary position, and to calculate the precise coordinates of the effective target in the preset local coordinate system based on the current pose and visual ranging results of the wall-climbing robot. The spraying beacon module is used to control the wall-climbing robot to move to the spraying operation position according to the precise coordinates, and to use the spraying deployment module to spray a feature beacon pattern with unique coding information on the vertical position corresponding to the precise coordinates. An association storage module is used to associate and store the attribute information of the valid target with the unique encoding information of the feature beacon, wherein the attribute information includes at least the target type and the precise coordinates.

9. An electronic device, characterized in that, include: At least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method as described in any one of claims 1-7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, The method as described in any one of claims 1-7 is implemented when the computer program instructions are executed by the processor.

Images