Method and system for detection control of a detection target suspended in a restricted vertical space

Abstract

The present application relates to the technical field of nondestructive testing and intelligent inspection equipment, and provides a detection control method and system for a suspended detection target in a restricted vertical space, comprising: obtaining a first height estimation based on the movement amount of a retracting and releasing mechanism, and a second height estimation based on the atmospheric pressure of the environment where the suspended detection target is located; complementarily calibrating the first height estimation and the second height estimation to obtain a target height; controlling the retracting and releasing mechanism to operate based on the target height, so as to drive the suspended detection target to reach a target detection position, and control the detection load to perform a detection operation at the target detection position. The present application eliminates the cumulative drift and short-term fluctuation of a single sensor through double-source height complementary calibration, realizes high-precision positioning and stable imaging in a restricted vertical space, and improves the coverage rate and data reliability of internal defect detection.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing and intelligent inspection equipment technology, and in particular to a detection control method and system for suspended targets in confined vertical space. Background Technology

[0002] In internal inspection operations within confined vertical spaces such as wind turbine towers, shafts, and enclosed pipelines, winch equipment is typically used to lower the inspection terminal via a flexible traction component for visual data acquisition. Existing positioning technologies primarily rely on single-sensor solutions: on the one hand, relying solely on encoders to measure rope release length is susceptible to factors such as rope stretching deformation, mechanical backlash, and variations in the diameter of multiple layers of winding during long-distance ascents and descents, resulting in cumulative positioning deviations with increasing depth; on the other hand, relying solely on barometric pressure sensors to measure altitude, while providing an absolute height reference, is highly sensitive to ambient temperature, airflow disturbances, and local pressure fluctuations, exhibiting insufficient short-term stability. These limitations of single-sensor positioning methods lead to positional deviations when the inspection terminal hovers at the target location, resulting in blurred image acquisition, missed defects, or duplicate inspections, making it difficult to meet the demands of high-precision automated inspection. Summary of the Invention

[0003] Therefore, it is necessary to provide a detection control method and system for suspended targets in confined vertical spaces to address the problems of low positioning accuracy and poor detection operation stability of existing technologies, so as to achieve high-precision positioning and stable detection based on dual-source height complementary calibration.

[0004] This invention provides a detection and control method for suspended targets in confined vertical space, comprising: Obtain a first altitude estimate based on the motion of the retraction mechanism, and a second altitude estimate based on the ambient air pressure of the suspended detection target. Complementary calibration is performed on the first altitude estimate and the second altitude estimate to obtain the target altitude; The retraction mechanism is controlled based on the target height to drive the suspended detection target to the target detection position and control the detection load to perform detection work at the target detection position.

[0005] In one embodiment, complementary calibration of the first altitude estimate and the second altitude estimate is performed to obtain the target altitude, including: initializing the displacement measurement of the deployment and take-off mechanism at the initial detection position and establishing a reference air pressure; during the lifting and lowering of the suspended detection target, updating the first altitude estimate based on the cumulative value of the displacement measurement, and updating the second altitude estimate based on the difference between the current ambient air pressure and the reference air pressure; using the first altitude estimate to constrain the short-term fluctuations of the second altitude estimate, and using the second altitude estimate to correct the cumulative drift of the first altitude estimate, so as to calculate the target altitude.

[0006] In one embodiment, the operation of the retraction mechanism is controlled based on the target height to drive the suspended detection target to the target detection position and control the detection load to perform detection work at the target detection position, including: determining multiple target detection positions based on detection parameters; determining whether the suspended detection target has reached the current target detection position based on the comparison result between the target height and the current target detection position; controlling the retraction mechanism to enter a fixed-point holding state in response to the suspended detection target reaching the current target detection position; and controlling the detection load to perform image acquisition work in the fixed-point holding state.

[0007] In one embodiment, the first height estimate is used to constrain the short-term fluctuations of the second height estimate, and the second height estimate is used to correct the cumulative drift of the first height estimate to obtain the target height. This includes: calculating the height residual between the first height estimate and the second height estimate; when the height residual exceeds a preset residual threshold, determining that the second height estimate is abnormal and reducing its fusion weight, while maintaining the target height output based on the first height estimate; when the height residual recovers to within the preset residual threshold, dynamically calculating the fusion weight based on the height residual and historical change trends; and using the fusion weight to perform weighted fusion of the first height estimate and the second height estimate to update the target height.

[0008] In one embodiment, based on the comparison between the target height and the current target detection position, it is determined whether the suspended detection target has reached the current target detection position, and the detection load is controlled to perform image acquisition in the fixed-point holding state, including: when the height difference between the target height and the current target detection position is less than a preset pausing distance, the lifting and lowering mechanism is controlled to reduce the lifting speed; when the suspended detection target reaches the current target detection position, the lifting and lowering mechanism is controlled to enter the fixed-point holding state and maintain it for a preset stable duration; after maintaining the preset stable duration, the detection load is triggered to perform image acquisition.

[0009] In one embodiment, determining multiple target detection positions based on detection parameters includes: in calibration mode, determining the detection range based on the start position and the end position, and dividing it into multiple target detection positions according to a preset step size; or, in configuration mode, receiving a configuration instruction containing detection range and step size parameters, and calculating multiple target detection positions based on the detection range and step size parameters.

[0010] In one embodiment, the method further includes: acquiring the wireless channel status between the platform-side control unit and the suspension-side control unit; in response to the wireless channel status meeting the idle condition, transmitting lifting control commands and detection status feedback between the platform-side control unit and the suspension-side control unit respectively using a time-division scheduling mechanism; wherein the lifting control commands and detection status feedback are allocated to different communication opportunities for transmission.

[0011] In one embodiment, in the stationary state, controlling the detection payload to perform image acquisition operations includes: controlling the gimbal of the detection payload to perform at least one of de-rotation or heading lock; after the gimbal attitude is stabilized, triggering the image acquisition device to capture the current detection image; and associating the current detection image with the target height to form a detection record based on the height index.

[0012] Furthermore, the present invention also provides a confined vertical space detection system, the system comprising: A retraction mechanism is used to lift and suspend the detection target via a flexible traction element; a first height detection component is used to output a first height estimate related to the amount of motion of the retraction mechanism; The second altitude detection component is located on the side of the suspended detection target and is used to output ambient air pressure data to obtain a second altitude estimate. The detection payload is used to perform detection and data acquisition operations; The controller is communicatively connected to the take-up and take-down mechanism, the first height detection component, the second height detection component, and the detection load. The controller is configured to perform complementary calibration on the first height estimate and the second height estimate to obtain the target height, and control the operating status of the take-up and take-down mechanism and the detection load based on the target height.

[0013] In one embodiment, the take-up and take-down mechanism includes a winch motor, and the first height detection component includes an encoder mounted on the winch motor shaft or a moving part of the take-up and take-down mechanism; the suspension detection target includes a pod terminal, and the second height detection component includes a barometric pressure sensor mounted on the pod terminal; the detection load includes a gimbal camera; the controller includes a platform-side control unit and a suspension-side control unit, the platform-side control unit communicating with the take-up and take-down mechanism via a wired communication link, and the platform-side control unit and the suspension-side control unit communicating via a wireless communication link; the suspension-side control unit is communicatively connected to the barometric pressure sensor and the gimbal camera, and feeds back status data to the platform-side control unit.

[0014] The aforementioned detection and control method and system for suspended targets in confined vertical spaces obtains the target height by acquiring a first height estimate based on the motion of the deployment and retraction mechanism and a second height estimate based on ambient air pressure, and then performing complementary calibration on both. This leverages the complementary error characteristics between the short-term continuity of rope displacement measurement and the long-term absolute reference value of barometric altimetry. The first height estimate effectively suppresses short-term high-frequency noise from the barometric altimeter caused by ambient airflow and temperature gradients, while the second height estimate corrects for long-term cumulative drift caused by rope tension, creep, and mechanical clearances. This results in a high-precision target height with both dynamic response speed and long-term stability within the confined vertical space. The closed-loop control system built upon this target height ensures that the detection load accurately reaches the predetermined detection position and performs fixed-point operations, eliminating the positioning deviation and imaging blurring problems caused by traditional single-positioning methods, and improving the coverage and data reliability of defect detection within the vertical space.

[0015] Furthermore, the robustness of the system in complex environments is enhanced through residual monitoring and dynamic weight adjustment mechanisms; combined with pausing control and gimbal stabilization strategies, a stable physical benchmark is provided for high-quality image acquisition; and with the wireless time-sharing scheduling mechanism, real-time and reliable interaction of control commands and status data is ensured in shielded environments such as deep wells. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a hardware connection flowchart of the confined vertical space detection system according to an embodiment of the present invention; Figure 2 This is a block diagram of the module architecture of the confined vertical space detection system according to an embodiment of the present invention; Figure 3 This is an interface diagram of the automatic defect identification software according to an embodiment of the present invention.

[0018] Among them, 100-retraction and deployment mechanism, 120-first height detection component, 210-second height detection component, 220-detection load, 310-platform side control unit, 320-suspension side control unit, 400-power system, 500-wireless communication link, and 600-wired communication link. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0020] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items.

[0021] The following is combined with Figures 1-3 The present invention describes a detection and control method and system for suspended targets in confined vertical space.

[0022] like Figure 1 and Figure 2 As shown, in one embodiment, a detection and control method for suspended targets in confined vertical space is provided. This method constructs a high-precision positioning benchmark by fusing heterogeneous sensor data, and uses this benchmark to drive an automated detection process, mainly including the following steps.

[0023] Step S110: Obtain a first altitude estimate based on the motion of the retraction mechanism and a second altitude estimate based on the ambient air pressure of the suspended detection target.

[0024] The first type of height estimation is the relative displacement indirectly calculated by monitoring the mechanical motion of the hoisting mechanism. For example, motion parameters such as the rotation angle of the hoisting motor, the output shaft speed, or the release length of the flexible traction component can be collected and combined with mechanical conversion relationships such as transmission ratio or wheel diameter to calculate the displacement change of the suspended detection target relative to the starting point. The advantage of this method is its high sampling rate and good short-term continuity, allowing it to sensitively reflect the instantaneous motion state of the equipment. However, due to potential elastic stretching, creep, or multi-layer winding diameter changes in the flexible traction component during long-distance lifting and lowering, as well as gaps and wear in the mechanical transmission chain, the measurement error gradually accumulates with increasing operating time and distance. The second type of height estimation is the absolute height reference obtained by directly sensing the physical environmental parameters of the suspended detection target's location. For example, a barometric pressure sensor can be used to collect ambient atmospheric pressure in real time, and the current height value can be calculated based on a model showing the correspondence between air pressure and altitude. This method, which does not rely on mechanical transmission links, can provide absolute position information unaffected by rope deformation, effectively avoiding the problem of accumulated errors. However, its measurement results are easily affected by local airflow disturbances, temperature gradient changes, or air pressure fluctuations caused by the movement of the equipment itself within a confined vertical space, resulting in poor short-term data stability. This step provides data conditions for subsequent high-precision positioning by simultaneously acquiring these two types of data sources with complementary error characteristics.

[0025] Step S120: Perform complementary calibration on the first altitude estimate and the second altitude estimate to obtain the target altitude.

[0026] Complementary calibration is a fusion process based on the error distribution characteristics of two types of data sources at different time scales. In this process, the high-frequency response and continuity characteristics of the first altitude estimate within a short time window are used to smooth and constrain the high-frequency noise generated by environmental disturbances in the second altitude estimate, preventing instantaneous jumps in barometric pressure data from interfering with system judgment. Simultaneously, the absolute reference characteristics of the second altitude estimate over a long time scale are used to continuously correct the low-frequency drift caused by mechanical factors in the first altitude estimate, eliminating the erosion of positioning accuracy by accumulated deviations. The target altitude output after complementary calibration combines the dynamic tracking capability of mechanical measurement with the long-term stability of barometric altimetry, serving as a position reference in the detection and control system for subsequent control stages. The calculation of the target altitude is a real-time, dynamically updated process that spans the entire ascent and descent cycle of the suspended detection target, rather than an offline calibration performed only at a specific moment.

[0027] Step S130: Control the operation of the retraction mechanism based on the target height to drive the suspended detection target to the target detection position, and control the detection load to perform detection work at the target detection position.

[0028] The controller uses the target height output in step S120 as a closed-loop feedback quantity, compares it with the preset target detection position, and generates corresponding drive commands based on the position deviation to adjust the operating state of the take-up and retractor mechanism. For example, when there is a large deviation between the target height and the target detection position, the take-up and retractor mechanism is controlled to run at a higher speed to quickly approach the target; when the deviation is reduced to a predetermined range, the operating strategy is automatically adjusted to achieve smooth deceleration and precise stopping. After confirming that the suspended detection target has accurately reached the target detection position, the controller further coordinates the detection load to perform predetermined detection operations, such as triggering the image acquisition device to capture images, starting the flaw detection sensor to scan, or adjusting the gimbal attitude. This control logic ensures that the detection operation is always carried out under the premise of high-precision positioning, avoiding problems such as missed detection, double detection, or blurred imaging caused by inaccurate positioning. This control process forms a complete control loop from height perception, data fusion to execution drive, enabling the system to perform automatic inspection operations in complex and restricted vertical spaces such as wind turbine towers, shafts, and enclosed pipelines.

[0029] In one embodiment, complementary calibration is performed on the first altitude estimate and the second altitude estimate to obtain the target altitude, specifically including the following steps: Step S121 initializes the displacement measurement of the deployment and take-up mechanism at the initial detection position and establishes the reference air pressure. Specifically, the initial detection position usually refers to the wellhead platform, the top inlet of the tower, or other preset detection starting point. This position is both the starting point of the physical operation and the origin of the system coordinate system. Initializing the displacement measurement of the deployment and take-up mechanism includes zeroing the cumulative pulse count or rotation angle of the encoder, and pre-tightening or compensating the clearance of the mechanical transmission chain to ensure the accuracy of subsequent cumulative displacement values. Establishing the reference air pressure requires sampling when the suspended detection target is stationary and the ambient airflow is relatively stable to avoid interference from dynamic pressure or instantaneous turbulence generated by equipment movement on the air pressure reading. By simultaneously setting the zero point of mechanical displacement and calibrating the atmospheric pressure reference surface at the same spatiotemporal node, the alignment of the mechanical measurement coordinate system and the physical environment coordinate system is achieved, establishing a unified reference benchmark for the subsequent fusion of dual-source data.

[0030] Step S122: During the lifting and lowering process of the suspended detection target, the first altitude estimate is updated based on the cumulative value of displacement measurement, and the second altitude estimate is updated based on the difference between the current ambient air pressure and the reference air pressure. For example... Figure 2As shown, the controller continuously performs data acquisition and processing throughout the entire lifting cycle. On one hand, it acquires encoder signals from the winch motor shaft or guide wheel in real time, calculates the real-time release length of the flexible traction component by integrating the rotation angle and combining it with the wheel diameter parameter, and uses this as the updated value for the first height estimate. On the other hand, it acquires air pressure sensor data from the suspension detection target side in real time, calculates the difference between the current ambient air pressure and the reference air pressure established in step S121, and converts this pressure difference into a relative height change based on the air pressure-height conversion model, using this as the updated value for the second height estimate. These two data streams are generated in parallel on the time axis, reflecting the mechanical motion state of the equipment and the physical environment state, respectively, forming the data basis for complementary calibration.

[0031] Step S123 involves using the first height estimate to constrain the short-term fluctuations of the second height estimate, and using the second height estimate to correct the cumulative drift of the first height estimate, in order to calculate the target height. This step involves the core processing logic of dual-source complementary calibration. In a confined vertical space, local airflow disturbances, temperature gradient changes, or wind pressure effects caused by the movement of the equipment itself often lead to high-frequency noise or instantaneous jumps in the output of the barometric pressure sensor. If used directly for positioning, this would cause control oscillations. Because the mechanical motion of the flexible traction component has physical inertia and continuity, the first height estimate exhibits high stability and reliability within a short time window. Therefore, the system uses the rate of change of the first height estimate as a constraint condition to filter or smooth the abnormal high-frequency components in the second height estimate, thereby suppressing the interference of environmental noise on the positioning results. Conversely, as the lifting operation time and distance increase, the deformation error of the flexible traction component caused by factors such as self-weight stretching, creep, and changes in the diameter of multiple layers of winding will gradually accumulate, causing the first height estimate to deviate from the true position. In principle, barometric altimetry does not rely on the mechanical transmission link, and its measurement results directly reflect the absolute vertical position of the suspended detection target, without any cumulative error as the stroke increases. Therefore, the system utilizes the absolute reference characteristics of the second altitude estimate to identify and correct the long-term drift trend of the first altitude estimate. Through this bidirectional dynamic adjustment mechanism, the calculated target altitude possesses both the high dynamic response capability of mechanical measurement and the long-term stability of barometric altimetry, effectively overcoming the limitations of single sensing methods under complex operating conditions and providing a position reference for subsequent detection and control.

[0032] In one embodiment, to further improve the robustness of dual-source complementary calibration under complex operating conditions, this invention refines the process of using a first height estimate to constrain the short-term fluctuations of the second height estimate, and using the second height estimate to correct the cumulative drift of the first height estimate in order to calculate the target height. Specifically, this process dynamically adjusts the fusion strategy by monitoring the consistency between data sources, rather than using a static mathematical average, thereby ensuring the continuity of positioning while making full use of effective information.

[0033] First, the height residual between the first height estimate and the second height estimate is calculated. The height residual is a direct indicator of the consistency between the two types of heterogeneous sensor data at the current moment. In this embodiment, the height residual... It can be calculated using the following formula: , in, Indicates the current sampling time. This is a first height estimate determined based on the motion of the deployment and take-off mechanism. This is a second altitude estimate based on ambient air pressure. The residual value reflects the instantaneous deviation between the mechanical measurement path and the barometric altimeter path. Ideally, they should converge; however, in actual confined vertical space operations, rope tension deformation can lead to… Slow drift occurs, while localized airflow disturbances or temperature gradients can cause... The high-frequency jump, therefore It is a variable that changes dynamically over time.

[0034] Secondly, when the height residual exceeds a preset residual threshold, the second height estimate is deemed abnormal and its fusion weight is reduced, while the target height output is maintained based on the first height estimate during this period. (Preset residual threshold) The selection of the sensor is not arbitrary, but determined based on statistical analysis of the sensor's nominal accuracy, the environmental noise level of the installation location, and historical normal operation data. For example, in the scenario of wind turbine tower inspection, the potential thermal pressure effect inside the tower and the piston wind effect caused by equipment lifting and lowering are taken into account. It can be set to a value between 0.3 meters and 0.8 meters. When At this point, the system determines that the current air pressure data is significantly affected by environmental interference, and its reliability as an absolute altitude reference has fallen below the safety boundary. The controller then automatically triggers the anomaly handling mechanism, adjusting the fusion weights of the second altitude estimate. Reduced to extremely low levels (e.g.) Even temporarily setting it to zero, so that the target height output depends primarily or entirely on the first height estimate. Because mechanical motion possesses physical inertia, Since sudden changes are unlikely to occur in a short period of time, this weighting strategy can effectively isolate the impact of abnormal air pressure on the positioning results, avoid false actions or oscillations caused by position feedback jumps in the detection load, and ensure the stability and safety of the control system.

[0035] Secondly, when the height residual recovers to within the preset residual threshold, the fusion weight is dynamically calculated based on the height residual and its historical trend. When environmental interference weakens or disappears, This indicates that the two data streams have returned to a consistent range, at which point the system regains confidence in the barometric pressure data. However, the level of confidence is not restored abruptly, but rather finely adjusted based on data quality. This embodiment employs a dynamic weighting calculation method based on residual amplitude and historical volatility. (Fusing weights) The calculation formula is shown in the following example: , in, The fusion weights are the second height estimates at the current time. This is the balance coefficient, and its value range is... This is used to adjust the contribution ratio of instantaneous residuals and historical stability to the weights; For the past The variance of the height residual within each sampling period characterizes the stability of the recent environment; This is the attenuation factor, used to control the sensitivity of the impact of historical fluctuations on the current weights. The physical meaning of this formula is: the smaller the instantaneous residual and the more stable the historical fluctuations, the better. As the pressure approaches a larger value, the system utilizes air pressure data more extensively to correct for accumulated mechanical errors; conversely, even if the residuals do not exceed the limit, if recent data fluctuations are significant, the system will conservatively reduce the air pressure weight. Correspondingly, the fusion weight of the first altitude estimation... Satisfy normalization constraints This dynamic calculation mechanism enables the fusion algorithm to adapt to different environments, automatically finding the optimal data combination under varying conditions.

[0036] Finally, a fusion weight is applied to weight and fuse the first and second height estimates to update the target height. The update formula is: , Through the aforementioned weighted fusion, the output target altitude retains the smoothness and continuity of mechanical measurement while introducing an absolute benchmark correction for barometric altimetry when environmental conditions permit. Compared to traditional fixed-weight fusion methods (e.g., always using...), this approach offers a more comprehensive and efficient solution. The dynamic weighting strategy in this embodiment has the following technical advantages: Fixed weights directly introduce noise into the positioning results when the barometric pressure sensor is disturbed, leading to detection position deviation; while when the barometric pressure data is accurate, its correction capability cannot be fully utilized, resulting in slow elimination of accumulated mechanical errors. This invention, through residual monitoring and adaptive weight adjustment, enables the positioning system to automatically switch dependent paths when a single sensor fails or its performance degrades, effectively solving the problem of unreliable positioning caused by the failure or performance degradation of a single sensor in a confined vertical space.

[0037] In one embodiment, the process of controlling the operation of the retraction mechanism based on the target height in step S130 to drive the suspended detection target to the target detection position and perform the detection operation is specifically defined, forming a complete segmented detection control logic. This control logic converts positioning data into automated operation instructions, enabling the detection process to be executed in a standardized manner according to a preset procedure.

[0038] Step S131: Determine multiple target detection positions based on detection parameters. Specifically, detection parameters are a set of features describing the spatial distribution of the detection task, typically including the detection start height, detection end height, and step interval between adjacent detection points. After receiving these parameters, the controller automatically generates an ordered sequence of target detection positions using a built-in path planning algorithm. For example, if the detection range is 0 meters to 100 meters and the step size is set to 2 meters, the controller will automatically calculate 51 discrete target detection position coordinates from 0 meters, 2 meters, 4 meters up to 100 meters. This process is entirely completed internally by the controller, eliminating the need for manual teaching or input of each point, thus improving job preparation efficiency and avoiding human error. Detection parameters can be obtained in various ways, including from pre-stored calibration data or externally input configuration files, as long as the necessary information for generating the position sequence can be extracted.

[0039] Step S132: Based on the comparison between the target height and the current target detection position, determine whether the suspended detection target has reached the current target detection position. Specifically, during the lifting and lowering process of the suspended detection target, the controller continuously reads the real-time target height obtained through complementary calibration and compares it numerically with the coordinates of the currently executing target detection position. When the absolute value of the difference between the two is less than a preset arrival determination threshold, the system determines that the suspended detection target has accurately reached the predetermined position. The arrival determination threshold here is a tolerance range set according to the system control accuracy and mechanical response characteristics, for example, it can be set between 5 mm and 20 mm. This determination process is based on automatic logical judgment of fused positioning data, without relying on the operator's visual observation or experience estimation, avoiding the influence of subjective factors on the consistency of detection points, and ensuring the high degree of overlap of sampling positions in different detection tasks.

[0040] Step S133: In response to the suspended detection target reaching the current target detection position, the take-up and retraction mechanism is controlled to enter a fixed-point holding state. Specifically, once the determination in step S132 is complete, the controller immediately sends a fixed-point holding command to the drive unit of the take-up and retraction mechanism. In this state, the hoisting motor switches to torque holding mode or activates the mechanical braking device to actively resist gravity and external force disturbances, keeping the suspended detection target stationary at its current vertical position. The physical significance of this step is to cut off the motion inertia and suppress the residual vibration of the flexible traction component, creating a mechanically stable platform environment for subsequent optical acquisition. Without this step, the suspended detection target often experiences a low-frequency oscillation decay period of several seconds or even tens of seconds after a sudden stop, directly leading to blurred imaging or stitching misalignment.

[0041] Step S134: In the stationary holding state, the detection load is controlled to perform image acquisition. Specifically, the triggering and timing of the image acquisition operation are strictly controlled by the stationary holding state. The controller sends the shooting command to the detection load only after confirming that the retraction mechanism has entered the stationary holding state and the attitude is stable. This strong timing constraint ensures that each frame of detection image is acquired under the condition that the suspended detection target is relatively stationary, guaranteeing the image clarity and geometric consistency. At the same time, since the detection position is determined by the high-precision target height, the acquired image naturally possesses accurate spatial index attributes, facilitating subsequent precise defect localization and panoramic stitching. The control process from position determination and steady-state holding to image acquisition is all automatically executed by the controller, enabling the detection operation in the confined vertical space to be completed automatically according to the preset process.

[0042] In one embodiment, to further improve the quality of detection imaging and ensure the smoothness of motion control, the process of determining whether the suspended detection target has reached the current target detection position based on the comparison result between the target height and the current target detection position, and controlling the detection load to perform image acquisition operations in a fixed-point holding state, has been refined. This refinement process transforms the abstract position determination into specific motion control strategies and timing constraints, mainly including the following steps.

[0043] Specifically, when the height difference between the target height and the current target detection position is less than the preset deceleration distance, the lifting and lowering mechanism is controlled to reduce its lifting speed. During the process of the suspended detection target approaching the target point, if high-speed operation is maintained until the moment of arrival before a sudden stop, the huge inertial impact will exacerbate mechanical wear and induce significant low-frequency longitudinal and lateral vibrations in the flexible traction components, resulting in a significantly prolonged subsequent stabilization time or even failure to converge within a short period. By setting a preset deceleration distance, the system can intervene in speed adjustment before reaching the target point, allowing the suspended detection target to approach the target position with a smooth deceleration curve. For example, in the wind turbine tower detection scenario, considering the elastic modulus of the wire rope and the mass of the pod, the preset deceleration distance can be set to a value between 0.15 meters and 0.3 meters, preferably 0.2 meters. When the difference between the real-time calculated target height and the current target detection position falls within this range, the controller automatically reduces the speed of the winch motor from cruising speed linearly or along an S-curve to a low-speed crawling state. This graded deceleration strategy effectively absorbs the system's kinetic energy, causing the residual velocity of the suspended detection target to approach zero when it reaches the target point. This reduces the mechanical impact and initial disturbance amplitude caused by sudden stopping, creating favorable mechanical boundary conditions for rapid entry into a stable state.

[0044] Furthermore, when the suspended target reaches the current target detection position, the deployment and retrieval mechanism enters a stationary holding state and maintains a preset stabilization time. Despite the easing process, the flexible traction component inevitably experiences slight elastic rebound and attitude adjustment in the initial static state, and the gyroscope stabilization circuit inside the gimbal camera also requires a certain amount of time to converge. The preset stabilization time is a time window reserved for this physical transient process; essentially, it is the necessary period to wait for the system's mechanical vibration energy to dissipate below the allowable threshold for optical imaging. For example, for tower inspection operations at depths of hundreds of meters, due to the low stiffness characteristics of long ropes, this preset stabilization time is typically set between 30 and 60 seconds, with a typical value of 40 seconds. During this period, the deployment and retrieval mechanism maintains a torque-locked or braking state, does not respond to any displacement commands, and relies solely on the system's own damping characteristics to gradually bring the pod attitude to a standstill. If this step is omitted or the duration is set too short, even if the position feedback indicates that the pod has reached its destination, the actual pod will still be in a state of slight swaying. Direct shooting will result in blurry or ghosting images, severely affecting the recognition rate of defects such as micro-cracks. Therefore, the preset stabilization time is a technical parameter determined based on the dynamic response characteristics of the flexible suspension system within a confined vertical space.

[0045] Finally, after maintaining a preset stabilization period, the detection payload is triggered to perform image acquisition. This step establishes strict timing constraints: image acquisition can only begin after the preset stabilization period has completely ended, rather than being synchronized with the stationary position or triggered solely based on the position arrival signal. The controller maintains an independent state timer, which only sends a shutter trigger command to the detection payload after confirming that the stationary position duration has reached the preset stabilization period threshold. This rigid timing interlocking mechanism eliminates the possibility of missampling in an unstable platform state from the control logic level, ensuring that each frame of the detection image is acquired under steady-state conditions where the suspended detection target is relatively stationary. Compared to the traditional method of taking a picture solely based on the position sensor's determination of position, this embodiment introduces time-dimensional stability verification, linking motion control parameters with imaging quality indicators. This solves the problem of asynchronous positioning and imaging caused by flexible connections in a confined vertical space, providing data conditions for subsequent defect identification and panoramic stitching. The specific values ​​of the above-mentioned stopping distance and stabilization time are only illustrative examples. In practical applications, they can be adjusted according to factors such as the material and length of the traction component, the load weight, and the ambient wind speed. As long as the timing logic of first decelerating and buffering, then stabilizing the image, and finally triggering the acquisition is followed, it falls within the protection scope of this invention.

[0046] In one embodiment, the process of determining multiple target detection positions based on detection parameters is specifically defined, providing two parallel parameter acquisition and position generation paths to adapt to different operational preparation conditions. Specifically, the system supports two working modes: calibration mode and configuration mode. Although the input sources differ, both ultimately output a target detection position sequence in a uniform format for subsequent lifting control and detection operations.

[0047] In calibration mode, the detection range is determined based on the starting and ending positions, and multiple target detection positions are obtained by dividing the area into multiple positions according to a preset step size. This mode is suitable for scenarios where the detection space boundary is unknown, the environment changes, or high-precision on-site confirmation is required. In practice, the suspended detection target is first placed at the starting position, such as the entrance platform of the detection space. At this moment, the controller performs initialization operations on the displacement measurement unit of the take-up and take-down mechanism, such as resetting the cumulative pulse count or angle value of the encoder to zero, and marking the current position as the detection start point. Subsequently, the take-up and take-down mechanism is controlled to drive the suspended detection target deeper into the detection space at a safe speed until it reaches the preset detection endpoint or the termination position specified by the operator. Upon reaching the termination position, the system records the current displacement measurement value as the physical boundary of the detection range. Based on the measured height difference between the starting and ending positions, combined with the preset step size parameters, the controller automatically calculates the number of detection segments and the precise coordinates of each target detection position. This method directly senses the spatial boundary through physical motion, effectively eliminating positioning reference deviations caused by missing drawings or changes in the on-site structure, ensuring that the detection range strictly matches the actual physical space.

[0048] Alternatively, in configuration mode, the system receives configuration instructions containing detection range and step size parameters, and calculates multiple target detection positions based on these parameters. This mode is suitable for batch operation scenarios where the detection objects are standardized, historical data is complete, or rapid deployment is required. In practice, operators can directly input the start and end height values ​​of the detection range and the step distance between adjacent detection points through a handheld terminal or the human-machine interface of the host computer software. After receiving these configuration instructions, the controller first verifies the rationality of the parameters, such as checking whether the step size is positive and whether the detection range is within the maximum allowable travel of the equipment. After the verification is successful, the controller does not need to drive the extension and retraction mechanism to perform a full physical traversal, but directly generates a complete sequence of target detection positions in real time based on the input numerical parameters and the built-in path planning algorithm. This method eliminates the lifting and retraction time required for on-site calibration, significantly shortens the operation preparation cycle, and is particularly suitable for continuous multi-unit inspections of facilities with fixed specifications, such as wind turbine towers.

[0049] Whether using the calibration mode to acquire boundaries through on-site movement or the configuration mode to acquire boundaries through parameter input, the final generated target detection position sequence is completely consistent in terms of data structure and control interface. This means that subsequent processes such as lifting control, point holding, and image acquisition do not need to distinguish the source of the position data and can all be executed according to a unified logic. This design enables the system to cope with different working conditions while maintaining the stability and reusability of the detection control logic, avoiding the risk of control strategy reconstruction due to mode switching.

[0050] In one embodiment, to ensure the reliability of data transmission within a confined vertical space, the method further includes: acquiring the wireless channel status between the platform-side control unit and the suspension-side control unit; in response to the wireless channel status meeting the idle condition, employing a time-division scheduling mechanism to transmit lifting control commands and detection status feedback between the platform-side control unit and the suspension-side control unit respectively; wherein the lifting control commands and detection status feedback are allocated to different communication opportunities for transmission.

[0051] Specifically, in confined vertical spaces such as wind turbine towers and shafts, metal walls exhibit shielding and multipath reflection effects on radio electromagnetic waves. This makes traditional full-duplex or random contention-based wireless communication prone to packet collisions, loss, or delays, thus affecting the real-time performance of lifting control and the integrity of detection data. This embodiment addresses this physical environment-related transmission bottleneck by introducing a channel-aware time-division scheduling mechanism at the communication protocol level. The process of acquiring the wireless channel status involves real-time evaluation of the physical layer quality of the current wireless communication link. For example, the platform-side control unit or the suspension-side control unit can continuously monitor the Received Signal Strength Indication (RSSI) or Signal-to-Noise Ratio (SNR) through a built-in RF front-end, or detect the channel energy level using a Carrier Sense Multiple Access (CSMA) mechanism. When the monitored channel energy is below a preset threshold and remains stable over multiple consecutive sampling periods, the system determines that the current wireless channel status meets the idle condition, indicating that no other strong interference sources occupy spectrum resources during this period, making it suitable for critical data transmission and reception operations.

[0052] After confirming that the channel meets the idle condition, the system does not immediately send all pending data in an unordered manner, but strictly executes a time-sharing scheduling strategy. Ascent and descent control commands are typically generated by the platform-side control unit and include key control parameters such as speed settings, target position, and start / stop commands. These commands have high requirements for transmission latency and determinism. Detection status feedback is generated by the suspension-side control unit and includes telemetry information such as real-time altitude, air pressure, gimbal attitude, and battery level. This data volume is relatively large, but allows for minor transmission jitter. In this embodiment, these two heterogeneous data streams are orthogonally isolated on the time axis, i.e., allocated to non-overlapping communication time slots for transmission. For example, within a predefined communication frame period, the first time slot is specifically reserved for the platform-side control unit to send descent and descent control commands downwards, while the second time slot is specifically allocated to the suspension-side control unit to send back detection status feedback upwards. A guard interval can also be set between the two time slots to eliminate inter-symbol interference caused by propagation delay. This deterministic time slot allocation method avoids bidirectional data streams competing for channel resources at the same time, avoids the accumulation of communication delays caused by retransmission mechanisms, and ensures that control commands arrive on time and status data is transmitted completely in a metal-sealed environment. This provides communication assurance for the aforementioned dual-source highly complementary calibration and segmented detection control. The aforementioned time-division scheduling mechanism runs automatically as a low-level communication assurance sub-process running in parallel with the detection operation process. Its execution frequency is much higher than the step sequence switching frequency of the detection operation. Therefore, it will not occupy or interfere with the normal lifting and image acquisition timing. Instead, it indirectly ensures the continuity and security of the detection operation by improving the robustness of the link.

[0053] In one embodiment, to further improve the availability and spatial positioning accuracy of the detection data, the process of controlling the detection payload to perform image acquisition operations under the aforementioned fixed-point holding state is specifically refined. This refined process covers the entire chain from optical platform image stabilization to structured data storage, ensuring that each frame of the detection image has clear imaging quality and accurate spatial attributes.

[0054] Specifically, the gimbal controlling the detection load performs at least one of despinning or heading lock. Within a confined vertical space, after the suspended detection target undergoes a long-distance ascent and descent and a fixed-point hovering transition phase, although the deployment and retraction mechanism has entered a macroscopically fixed-point holding state, the detection load itself may still retain residual torsional motion or slight oscillation around the axis of the flexible traction component. Directly shooting in this state would not only cause motion blur in the image but also result in inconsistent azimuth angles at different altitudes, increasing the difficulty of subsequent panoramic stitching. Therefore, before triggering shooting, the controller first sends an active image stabilization command to the gimbal. The despinning mode utilizes the gimbal's built-in inertial measurement unit to sense the rotational angular velocity around the axis in real time, and uses reverse compensation from the motor to counteract the residual torsion of the pod, maintaining the camera's optical axis inertial stability in the horizontal direction. The heading lock mode controls the gimbal's rotation and maintains it at a preset absolute azimuth angle (such as true north or the direction of the tower joint), ensuring a uniform field of view for all detection points. These two control methods can be used individually or in combination depending on the needs of the detection task. For example, first perform derotation to eliminate dynamic disturbances, and then switch to heading lock to align with the shooting reference, thereby establishing a stable optical platform for high-quality imaging.

[0055] Furthermore, after the gimbal's attitude stabilizes, the image acquisition unit is triggered to capture the currently detected image. This attitude stabilization does not rely solely on a fixed delay, but rather on an active determination process based on the gimbal's attitude feedback signal. The controller continuously reads the angular velocity and acceleration data output by the gimbal's inertial measurement unit. Only when the three-axis angular velocities are continuously below a preset static threshold (e.g., 0.5° / s) for a duration exceeding a preset confirmation window is the gimbal considered to have reached optical-grade stabilization. This sensor-feedback-based triggering mechanism effectively avoids the uncertainty in stabilization time caused by differences in mechanical damping or environmental disturbances, ensuring that the camera is in an optimal static state at the moment of each shutter release, thus guaranteeing the clarity of images of defects such as tiny cracks and corrosion.

[0056] Finally, the current detected image is associated with the target height and stored to form a detection record based on a height index. At the moment image acquisition is complete, the controller uses the target height value obtained through complementary calibration at the current moment as core metadata, embedding it into the header information of the image file or writing it to a separate index database. This associated storage establishes a strong binding relationship between visual information and spatial location, ensuring that each detected image carries a precise vertical coordinate label. For example, in a wind turbine tower inspection scenario, when the subsequent software system identifies a weld crack defect in an image, it can directly read the associated height label of that image to accurately locate the actual position of the defect within the tower, without the need for manual backtracking or estimation. Simultaneously, the height-indexed detection record also provides a reliable longitudinal alignment benchmark for multi-image panoramic stitching, enabling the system to automatically stitch discrete single-point images into a continuous unfolded view of the tower's inner wall according to height order. This data association mechanism transforms the advantages of high-precision front-end positioning into the value of back-end data analysis, completing the entire processing flow from physical positioning to data association.

[0057] like Figure 2 As shown, the present invention provides a confined vertical space detection system. The system includes a retraction mechanism 100, a first height detection component 120, a second height detection component 210, a detection load 220, and a controller.

[0058] The deployment and retraction mechanism 100 is used to raise and lower the suspended detection target via a flexible traction element. Specifically, the deployment and retraction mechanism 100 is typically deployed at the top entrance of the wellhead platform or the detection space. It contains a power drive unit and transmission components, and moves the suspended detection target vertically by winding or releasing the flexible traction element. The flexible traction element can be a steel wire rope, synthetic fiber rope, or armored cable, etc., serving both as a mechanical traction function and as a carrier for power or signal transmission. The motion accuracy and response characteristics of the deployment and retraction mechanism 100 directly affect the raising and lowering control effect of the suspended detection target.

[0059] The first height detection component 120 is used to output a first height estimate related to the motion of the take-up and release mechanism 100. Specifically, the first height detection component 120 can be integrated into moving parts such as the power output shaft, guide wheel, or drum of the take-up and release mechanism 100. By monitoring mechanical parameters such as rotation angle, rotation speed, or linear displacement in real time, it indirectly calculates the release length or take-up length of the flexible traction member. The first height estimate output by this component reflects the change in mechanical displacement of the suspended detection target relative to the starting point, and has the characteristics of high sampling rate and good short-term continuity, providing a high-frequency response relative position reference for the system.

[0060] The second altitude detection component 210 is installed on the side of the suspended detection target and is used to output ambient air pressure data to obtain a second altitude estimate. Specifically, the second altitude detection component 210 enters the confined vertical space along with the suspended detection target to directly sense the atmospheric pressure at its location. Since there is a definite physical correspondence between air pressure and altitude, a second altitude estimate reflecting the absolute vertical position of the suspended detection target can be obtained by converting the ambient air pressure data. This measurement path is independent of the mechanical transmission link and is not affected by rope tension, creep, or entanglement errors, effectively compensating for the cumulative deviation of the first altitude estimate during long-distance operation.

[0061] The detection payload 220 is used to perform detection and data acquisition operations. Specifically, the detection payload 220 is mounted on a suspended detection target and may include one or more sensing devices such as a visible light camera, an infrared thermal imager, a laser scanner, or an ultrasonic probe. After reaching the predetermined detection position, the detection payload 220 collects data on the inner wall structure, equipment status, or defect characteristics of the confined vertical space according to a preset strategy, providing initial evidence for subsequent defect identification and safety assessment.

[0062] The controller is communicatively connected to the retraction mechanism 100, the first height detection component 120, the second height detection component 210, and the detection load 220. Specifically, the controller constitutes the information processing and command scheduling center of the entire system. In terms of physical connection, the controller can establish high-speed data interaction with the retraction mechanism 100 and the first height detection component 120 located on the platform side via a wired link, and at the same time maintain real-time communication with the second height detection component 210 and the detection load 220 located on the suspension side via a wireless link, ensuring the synchronous acquisition of multi-source data and the accurate issuance of control commands.

[0063] The controller is configured to perform complementary calibration on the first and second height estimates to obtain the target height, and control the operation of the take-up and retraction mechanism 100 and the detection load 220 based on the target height. This configuration makes the controller a processing unit with embedded dual-source complementary calibration logic, performing the following functions: The controller uses the short-term stability of the first height estimate to constrain the environmental noise of the second height estimate, while using the absolute reference of the second height estimate to correct the cumulative drift of the first height estimate, thereby calculating a target height that combines dynamic response and long-term accuracy. Based on this, the controller further uses the target height as a closed-loop feedback quantity to automatically adjust the lifting speed and stopping position of the take-up and retraction mechanism 100, and triggers the detection load 220 to perform data acquisition after confirming the correct positioning. This integrated hardware and software architecture ensures that the detection method can be fully and reliably implemented at the product level, solving the accuracy loss and response lag problems caused by the separation of the positioning and control modules in traditional systems. Although Figure 2An exemplary connection topology for the components is shown. However, in practical applications, the controller's functions can be implemented by a single processor or by multiple processing units distributed on the platform side and the suspension side in collaboration. As long as the above-mentioned dual-source calibration and closed-loop control functions can be achieved, they are all within the scope of protection of this invention.

[0064] In one embodiment, the hardware composition of the aforementioned restricted vertical space detection system is specifically defined, providing an implementable dual-end collaborative hardware architecture.

[0065] Specifically, the take-up and release mechanism 100 includes a winch motor, and the first height detection component 120 includes an encoder mounted on the winch motor shaft or on a moving part of the take-up and release mechanism 100. (See reference...) Figure 1 The winch motor, as the power core of the system, drives the flexible traction component to raise and lower via forward and reverse rotation, thereby raising and lowering the suspended detection target. The encoder is preferably a non-contact magnetic encoder, directly coaxially mounted on the output shaft of the winch motor, or mounted on components that move synchronously with the rope, such as guide wheels. This mounting method directly acquires the raw signals of motor rotation angle or rope displacement, avoiding additional backlash errors introduced by intermediate transmission links. The pulse signal or angle value output by the encoder, after conversion, becomes the first height estimate. Due to the continuity of mechanical motion, this data has good stability over a short time scale, providing a high-frequency response relative position reference for the system.

[0066] The suspension detection target includes the pod terminal, and the second height detection component 210 includes a pressure sensor mounted on the pod terminal. (See reference...) Figure 2 The pod terminal is a mobile platform that enters confined vertical spaces, integrating sensing, computing, and communication units. A barometric pressure sensor is encapsulated within the pod terminal's housing and includes a vent to sense the external atmospheric pressure. Because the barometric pressure sensor moves with the pod terminal, its measurement directly reflects the device's current absolute vertical position, unaffected by rope tension, creep, or changes in winding diameter. This design, placing the barometric altimeter unit within the moving terminal, allows the second altitude estimate to exist independently of the mechanical transmission link, thus providing a physical reference for correcting the cumulative drift of the first altitude estimate.

[0067] The detection payload 220 includes a gimbal camera. The gimbal camera is fixedly mounted below or to the side of the pod terminal, serving as the core execution component for visual inspection. The gimbal integrates an inertial measurement unit and a servo motor, enabling automatic de-rotation or heading lock operations during the stationary position holding phase to counteract the impact of residual pod sway on image quality. The gimbal camera's trigger command is automatically generated by the suspension-side control unit 320 based on image stabilization criteria, ensuring the sharpness and geometric consistency of the acquired images.

[0068] The controller includes a platform-side control unit 310 and a suspension-side control unit 320. The platform-side control unit 310 communicates with the retraction mechanism 100 via a wired communication link 600, and the platform-side control unit 310 and the suspension-side control unit 320 communicate via a wireless communication link 500. (Refer to...) Figure 1 and Figure 2 This distributed dual-end control architecture is adapted to the special operating environment of confined vertical space. The platform-side control unit 310 is typically integrated into a handheld terminal or wellhead main control box, responsible for human-machine interaction, task planning, dual-source highly fusion algorithm calculation, and generation of macroscopic lifting commands. The wired communication link 600 can use CAN bus or industrial Ethernet to ensure low-latency and high-reliability control of the winch motor and encoder by the platform-side control unit 310, avoiding the instability of wireless signals in metal-shielded environments. The suspension-side control unit 320 is embedded inside the pod terminal, responsible for air pressure data acquisition, gimbal attitude control, image preprocessing, and status monitoring.

[0069] The suspended-side control unit 320 is communicatively connected to the air pressure sensor and the gimbal camera, and feeds back status data to the platform-side control unit 310. The wireless communication link 500 carries two-way data interaction across vertical space, including downlink transmission of lifting control commands and shooting trigger commands, and uplink transmission of real-time air pressure values, gimbal attitude angles, battery power, and other status data. To address the shielding and multipath effects of wireless signals from the metal walls inside deep wells or towers, the wireless communication link 500 operates in conjunction with the aforementioned time-sharing scheduling mechanism, allocating control commands and status feedback to different time slots for transmission, effectively reducing data conflicts and packet loss risks. Although this embodiment describes a specific hardware configuration including a winch motor, magnetic encoder, air pressure sensor, gimbal camera, and two-way controller, this is only a preferred embodiment of the present invention. In other embodiments, the deployment and recovery mechanism can also be a hydraulic winch or a linear module, the first height detection component can be a laser rangefinder or a draw-wire encoder, and the detection load can be replaced with an infrared thermal imager or an ultrasonic probe. As long as the function of dual-source height complementary calibration and closed-loop detection control can be achieved, it falls within the protection scope of the present invention.

[0070] In one embodiment, to verify the effectiveness and integrity of the aforementioned detection and control method and system under real and complex working conditions, this invention uses the detection of internal defects in wind turbine towers as a typical application scenario. Wind turbine towers are typical confined vertical spaces, characterized by their great depth, severe electromagnetic shielding of metal walls, significant temperature gradients, and the presence of piston-wind effects. This places stringent requirements on the positioning accuracy, imaging stability, and communication reliability of the suspended detection target. In this scenario, the specific process for the system to perform the detection operation is as follows.

[0071] First, during the preparation phase, the operator deploys the deployment and retrieval mechanism 100 on the wellhead platform at the top of the tower and places the suspended detection target (i.e., the pod terminal) in the starting position. The platform-side control unit 310 performs a zeroing operation on the encoder 120 on the winch motor shaft, completing the initialization of displacement measurement; simultaneously, the suspension-side control unit 320 collects the current atmospheric pressure value of the wellhead platform as the reference pressure, establishing the physical zero point for altitude calculation. This step achieves spatiotemporal alignment between the mechanical coordinate system and the ambient pressure coordinate system, providing a unified benchmark for subsequent dual-source complementary calibration.

[0072] Subsequently, the system initiates a descent inspection task. During the descent of the pod terminal along the inner wall of the tower, encoder 120 outputs a first height estimate reflecting the rope release length in real time, while barometric pressure sensor 210 outputs a second height estimate reflecting changes in ambient air pressure in real time. Since wind turbine towers typically exceed 100 meters in height, the wire rope experiences significant elastic tension and creep under its own weight and load, causing the first height estimate to gradually decrease in height with increasing depth. Simultaneously, the thermal pressure effect caused by temperature differences due to sunlight inside the tower and the piston-like wind generated by the pod's movement can lead to high-frequency fluctuations or even instantaneous jumps in the barometric pressure sensor readings. To address this characteristic, the controller continuously calculates the height residual between the two height estimates and performs real-time calibration based on the aforementioned dynamic weight fusion strategy. When the residual in the air pressure data exceeds a preset threshold due to airflow disturbance, the system automatically reduces the fusion weight of the second height estimate, relying on the first height estimate to maintain positioning continuity. When the environment is relatively stable, the absolute reference value of the barometric altimeter is used to correct the cumulative drift caused by rope tension. Through this adaptive and complementary mechanism, the target height output by the system can maintain centimeter-level accuracy throughout the entire stroke range, effectively overcoming the physical limitations of a single sensing method in a tower environment.

[0073] When the pod terminal reaches the preset target detection position based on the target height (e.g., a detection point is set every 2 meters), the controller executes segmented closed-loop control. When the difference between the target height and the current target detection position decreases to 0.2 meters, the system automatically controls the winch motor to enter the deceleration and slow-down phase, eliminating the inertial impact of high-speed movement. After reaching the target point, the deployment and take-up mechanism 100 enters a stationary holding state and maintains a preset stable duration of 40 seconds, waiting for the elastic vibration of the wire rope and the pod attitude to fully converge. During this period, the wireless communication link 500 adopts a time-division scheduling mechanism to distribute the control commands issued by the platform side and the status data returned by the suspension side to different time slots for transmission, avoiding conflicts and packet loss in the bidirectional data stream under the shielded environment of the metal tower, and ensuring the real-time performance and determinism of the control link.

[0074] After the preset stabilization period ends and the gimbal camera 220 completes de-rotation or heading lock, the system triggers image acquisition. The captured current detection image is instantly embedded with the current target height metadata, forming a height-indexed detection record, which is stored locally or transmitted back to the host computer. In the detection data visualization software interface, the system can automatically stitch discrete images into a panoramic view of the tower's inner wall based on the height index, and through methods such as... Figure 3 The automatic defect identification interface clearly marks defect features such as cracks, water seepage marks, and damage, along with their corresponding precise height locations. The damage feature marking interface displays defect statistics for each height segment and supports filtering views by height and generating annotation reports. This structured data-driven inspection record provides visual evidence for subsequent defect identification and assessment, enabling maintenance personnel to quickly locate the actual physical location of defects within the tower without manual estimation or reviewing video streams.

[0075] In summary, in the specific application scenario of wind turbine tower inspection, this invention solves the problem of accurate positioning under conditions of long cable tension and complex airflow interference through dual-source height complementary calibration; ensures clear imaging of minute defects through slow-stop stabilization and temporal constraints; adapts to the communication environment of enclosed metal spaces through wireless time-division scheduling; and achieves spatially structured expression of inspection results through height-indexed data association. This embodiment illustrates the applicability of the aforementioned technical solution in practical engineering scenarios. Besides wind turbine towers, this system can also be applied to automated inspection operations in other confined vertical spaces such as shafts, chimneys, and enclosed pipelines.

[0076] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0077] The above-described embodiments are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A detection and control method for a suspended target in a confined vertical space, characterized in that, The method includes: Obtain a first altitude estimate based on the motion of the retraction mechanism, and a second altitude estimate based on the ambient air pressure of the suspended detection target. The first altitude estimate and the second altitude estimate are complementaryly calibrated to obtain the target altitude. The retraction mechanism is controlled based on the target height to drive the suspended detection target to the target detection position and control the detection load to perform detection work at the target detection position.

2. The detection and control method for suspended targets in confined vertical space according to claim 1, characterized in that, The step of performing complementary calibration on the first altitude estimate and the second altitude estimate to obtain the target altitude includes: The displacement measurement of the take-up and release mechanism is initialized at the initial detection position, and a reference air pressure is established. During the lifting and lowering of the suspended detection target, the first height estimate is updated based on the cumulative value of the displacement measurement, and the second height estimate is updated based on the difference between the current ambient air pressure and the reference air pressure. The target height is calculated by using the first height estimate to constrain the short-term fluctuations of the second height estimate, and by using the second height estimate to correct the cumulative drift of the first height estimate.

3. The detection and control method for suspended targets in confined vertical space according to claim 1, characterized in that, The method of controlling the operation of the retraction mechanism based on the target height to drive the suspended detection target to the target detection position and controlling the detection load to perform detection operations at the target detection position includes: Multiple target detection locations are determined based on detection parameters; Based on the comparison between the target height and the current target detection position, it is determined whether the suspended detection target has reached the current target detection position; In response to the suspended detection target reaching the current target detection position, the retraction mechanism is controlled to enter a fixed-point holding state; In the fixed-point holding state, the detection payload is controlled to perform image acquisition operations.

4. The detection and control method for suspended targets in confined vertical space according to claim 2, characterized in that, The step of using the first altitude estimate to constrain the short-term fluctuations of the second altitude estimate, and using the second altitude estimate to correct the cumulative drift of the first altitude estimate, in order to calculate the target altitude, includes: Calculate the height residual between the first height estimate and the second height estimate; When the height residual exceeds a preset residual threshold, the second height estimation is determined to be abnormal and its fusion weight is reduced. During this period, the output of the target height is maintained based on the first height estimation. When the height residual recovers to within the preset residual threshold, the fusion weight is dynamically calculated based on the height residual and historical change trend; The first height estimate and the second height estimate are weighted and fused using the fusion weights to update the target height.

5. The detection and control method for suspended targets in confined vertical space according to claim 3, characterized in that, The step of determining whether the suspended detection target has reached the current target detection position based on the comparison result between the target height and the current target detection position, and controlling the detection load to perform image acquisition in the fixed-point holding state, includes: When the height difference between the target height and the current target detection position is less than the preset slow-stop distance, the retraction mechanism is controlled to reduce its lifting speed. When the suspended detection target reaches the current target detection position, the retraction mechanism is controlled to enter the fixed-point holding state and maintain it for a preset stable duration; After maintaining the preset stable duration, the detection payload is triggered to perform the image acquisition operation.

6. The detection and control method for suspended targets in confined vertical space according to claim 3, characterized in that, The process of determining multiple target detection locations based on detection parameters includes: In calibration mode, the detection range is determined based on the start and end positions, and the multiple target detection positions are obtained by dividing them according to a preset step size; Alternatively, in configuration mode, a configuration instruction containing detection range and step size parameters is received, and the multiple target detection positions are calculated based on the detection range and step size parameters.

7. The detection and control method for suspended targets in confined vertical space according to claim 1, characterized in that, The method further includes: Obtain the wireless channel status between the platform-side control unit and the suspension-side control unit; In response to the wireless channel state meeting the idle condition, a time-division scheduling mechanism is used to transmit lifting control commands and detection status feedback between the platform-side control unit and the suspension-side control unit, respectively. The lifting control command and the detection status feedback are transmitted at different communication times.

8. The detection and control method for a suspended target in a confined vertical space according to claim 5, characterized in that, In the fixed-point holding state, controlling the detection payload to perform image acquisition operations includes: The gimbal controlling the detection load performs at least one of despinning or heading lock; After the gimbal's attitude stabilizes, the image acquisition device is triggered to capture the currently detected image; The current detected image is associated with the target height and stored to form a detection record based on a height index.

9. A confined vertical space detection system, characterized in that, The system includes: A retraction and deployment mechanism is used to lift and suspend the detection target via a flexible traction component; A first height detection component is used to output a first height estimate related to the amount of motion of the retraction mechanism; The second altitude detection component is disposed on the side of the suspended detection target and is used to output ambient air pressure data to obtain a second altitude estimate; The detection payload is used to perform detection and data acquisition operations; The controller is communicatively connected to the take-up and take-down mechanism, the first height detection component, the second height detection component, and the detection load, respectively. The controller is configured to: perform complementary calibration on the first height estimate and the second height estimate to obtain a target height, and control the operating status of the take-up and take-down mechanism and the detection load based on the target height.

10. The confined vertical space detection system according to claim 9, characterized in that, The take-up and release mechanism includes a winch motor, and the first height detection component includes an encoder disposed on the shaft of the winch motor or on the moving part of the take-up and release mechanism. The suspension detection target includes the pod terminal, and the second height detection component includes a pressure sensor disposed on the pod terminal; The detection payload includes a gimbal camera; The controller includes a platform-side control unit and a suspension-side control unit. The platform-side control unit communicates with the take-up and take-down mechanism via a wired communication link, and the platform-side control unit and the suspension-side control unit communicate via a wireless communication link. The suspension-side control unit is communicatively connected to the air pressure sensor and the gimbal camera, and feeds back the status data to the platform-side control unit.

Images