Radar and vision system based hanging basket stability monitoring system
Patent Information
- Application Number
- CN202611050555.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-15
- Publication Date
- 2026-09-29
AI Technical Summary
本方案通过同步采集雷达点云与视觉图像并绑定统一高精度时间戳，实现多源数据在时间维度的严格对齐，消除不同传感器采集时差导致的空间坐标偏差；基于挂篮设计参数与现场标定数据构建统一物理基准与约束库，为所有特征校验与参数计算提供一致的物理参照标准，避免不同环节采用不同基准产生的系统误差；对提取的部件特征执行瞬时阈值与历史趋势双重校验，同时结合空间几何约束与时间运动连续性过滤异常特征，有效区分传感器瞬时干扰与结构真实形变，提升特征数据的可靠性；通过参数间物理耦合关系对计算得到的状态参数进行交叉校验与异常修正，解决单一参数独立计算导致的结果矛盾问题，保证状态参数的物理一致性；根据异常参数对应的可疑基准点标识，动态紧缩关联约束阈值，强化对高风险区域的监测力度，实现异常情况的主动追踪与精准定位。
Smart Images
Figure CN122836721A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of image detection technology, specifically to a hanging basket stability monitoring system based on radar and vision systems. Background Technology
[0002] The field of image detection technology includes target recognition and status monitoring based on optical imaging, feature extraction and deformation sensing based on radar echoes, and engineering safety monitoring technology based on multi-source sensor data fusion. The core of this technology lies in collecting the physical characteristics and operational data of target objects through various sensing devices, converting the collected raw information into quantifiable digital information, and continuously tracking and recording the target object's position, attitude, deformation, vibration, and other state parameters. This technology is widely used in safety monitoring in engineering fields such as bridge and tunnel construction, providing fundamental data support for assessing the operational status of engineering structures.
[0003] The radar and vision-based hanging basket stability monitoring system refers to a hanging basket status monitoring device applied to the cantilever construction of bridges. The patent addresses the following technical aspects during bridge cantilever construction: monitoring the tilt of the hanging basket, monitoring its horizontal and vertical displacement, monitoring its vibration amplitude and frequency, and monitoring the surface deformation of its load-bearing components. Generally, millimeter-wave radar is used to acquire the three-dimensional spatial coordinates and vibration time-series data of the hanging basket, while industrial cameras are used to acquire surface images and component contour data. The acquired radar and vision data are compared frame-by-frame and verified point-by-point, continuously acquiring and synchronously recording various status parameters of the hanging basket during construction. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a hanging basket stability monitoring system based on radar and vision systems, which has the functions of multi-source data synchronization, unified physical benchmark verification, dual feature filtering, parameter coupling correction and threshold dynamic enhancement, thus solving the problems mentioned in the background technology.
[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A hanging basket stability monitoring system based on radar and vision systems includes the following modules: The sensor synchronous acquisition module is used to synchronously acquire radar point clouds and visual images, generate spatiotemporally aligned sensing data streams, and output them. The physical benchmark library construction module is used to set benchmark geometric relationships and constraint thresholds based on the hanging basket design parameters and on-site calibration data, generate and output the hanging basket physical benchmark and constraint library; The feature timing verification module is used to receive the spatiotemporal alignment sensing data stream and the hanging basket physical benchmark and constraint library, perform component feature extraction and dual verification, generate a spatiotemporally consistent reliable coordinate set and output it. The parameter coupling correction module is used to receive the trusted coordinate set, calculate the state parameters by calling it, verify and correct abnormal parameters based on the physical coupling relationship, generate a stable state parameter set, and output suspicious reference point identifiers. The verification rule enhancement module is used to receive the suspicious benchmark point identifier, proportionally tighten the associated constraint threshold in subsequent monitoring cycles, and generate a rule dynamic enhancement status code.
[0006] Preferably, in the sensor synchronous acquisition module, the specific process of generating a spatiotemporally aligned sensing data stream is as follows: a built-in high-precision hardware clock generates a unified timestamp; the main controller generates trigger pulses according to a preset timing sequence and sends them simultaneously to the radar device and the visual acquisition device; after receiving the trigger pulse, the radar device completes structural scanning to generate a radar point cloud and binds it to the current unified timestamp; after receiving the same trigger pulse, the visual acquisition device completes image exposure to generate a visual image and binds it to the unified timestamp of the same moment; after binding, the data is sorted and integrated according to the timestamp, and the output is a spatiotemporally aligned sensing data stream.
[0007] Preferably, in the physical benchmark library construction module, the specific process for generating the basic part of the hanging basket physical benchmark and constraint library is as follows: Multiple measurements of key feature points of the hanging basket are taken using a total station, and the average value is obtained to acquire effective calibration coordinates; the Euclidean distance between any two feature points is calculated based on the effective calibration coordinates as the benchmark distance; the effective calibration coordinates are projected onto the horizontal plane, and the azimuth angle is calculated as the benchmark azimuth angle, forming a benchmark geometric relationship; the inherent stiffness parameters of the structure are extracted from the hanging basket design parameters, and matched with a preset safety factor to calculate and generate the distance variation tolerance and azimuth variation tolerance for each pair of feature points as constraint thresholds; the benchmark geometric relationship and the constraint thresholds are integrated to generate the basic part of the hanging basket physical benchmark and constraint library.
[0008] Preferably, in the physical benchmark library construction module, the specific process of forming a complete physical benchmark and constraint library for the hanging basket is as follows: extract the vertical distance between the main truss rotation hinge point and the front end point of the bottom basket from the hanging basket design drawings as the design hinge point height, calculate the tilt angle-displacement conversion coefficient with the design hinge point height as the only variable, and supplement the conversion coefficient into the foundation part to form a complete physical benchmark and constraint library for the hanging basket.
[0009] Preferably, in the feature time-series verification module, the specific process of performing component feature extraction and dual verification is as follows: extract feature point coordinates from the spatiotemporal alignment sensing data stream, calculate the measured distance and measured azimuth angle; perform instantaneous threshold verification: call the hanging basket physical reference and constraint library, subtract the measured value from the corresponding reference value to obtain the deviation, and compare it with the corresponding tolerance; perform historical trend verification on feature points with deviation exceeding the limit: retrieve its previous continuous historical measured sequence to calculate the historical average rate of change, calculate the current instantaneous rate of change based on the deviation between the current period and the previous period, and compare the difference between the two with a preset reasonable acceleration threshold; feature points with deviation not exceeding the limit or with deviation exceeding the limit but the difference in rate of change not exceeding the threshold are determined to pass the verification.
[0010] Preferably, in the feature time-series verification module, the specific process of generating a spatiotemporally consistent reliable coordinate set is as follows: obtain radar and visual measured coordinates for the same feature point respectively and perform the dual verification respectively; only when both sets of data pass the verification, take the average of the two types of coordinates as the final coordinates of the feature point; summarize the final coordinates of all valid feature points to generate the reliable coordinate set.
[0011] Preferably, in the parameter coupling correction module, the specific process of calculating and generating the state parameters in the stable state parameter set is as follows: extract all feature point coordinates from the reliable coordinate set, calculate the average value of each axial coordinate component to obtain the X, Y, and Z axis displacement components of the hanging basket; extract the feature point coordinates of the front and rear ends of the main truss, and calculate the tilt angle around the X axis; use the stable state parameter set output in the previous cycle as a reference, and automatically iterate and update it in each cycle.
[0012] Preferably, in the parameter coupling correction module, the specific process for verifying abnormal parameters is as follows: The tilt angle-displacement conversion coefficient is retrieved from the hanging basket physical reference and constraint library; the difference between the measured tilt angle around the X-axis in the current cycle and the tilt angle around the X-axis in the previous cycle is calculated to obtain the tilt angle change; the tilt angle change is multiplied by the tilt angle-displacement conversion coefficient to obtain the theoretical displacement change; the measured displacement change is calculated; the displacement coupling residual between the theoretical value and the measured value is calculated and compared with a preset residual allowable threshold; if the threshold is exceeded, the corresponding parameter is determined to be an abnormal parameter.
[0013] Preferably, in the parameter coupling correction module, the specific process of correcting abnormal parameters and marking suspicious reference points is as follows: the historical verification pass rates of feature points corresponding to the tilt angle around the X-axis and the horizontal displacement are counted respectively, and the one with the higher pass rate is selected as the anchor point reference parameter. The tilt angle-displacement conversion coefficient is used to correct another parameter. The corrected parameter and the uncorrected parameter are merged to generate a stable state parameter set, and the reference points associated with the abnormal parameters are marked as suspicious reference points.
[0014] Preferably, in the verification rule enhancement module, the specific process of generating the rule dynamic enhancement status code is as follows: determine the corresponding benchmark point based on the received suspicious benchmark point identifier, retrieve all the associated constraint thresholds from the hanging basket physical benchmark and constraint library, multiply by a preset first tightening coefficient to generate a temporary tightening tolerance; replace the original constraint threshold with the temporary tightening tolerance within the subsequent first preset monitoring period; if the same suspicious identifier is repeatedly triggered within the period, extend the period and use a second tightening coefficient; record the tightening coefficient, enhancement period and associated benchmark point information to generate the rule dynamic enhancement status code.
[0015] Compared with the prior art, the present invention provides a hanging basket stability monitoring system based on radar and vision systems, which has the following beneficial effects: This solution achieves strict temporal alignment of multi-source data by simultaneously acquiring radar point clouds and visual images and binding them with a unified high-precision timestamp, eliminating spatial coordinate deviations caused by time differences in data acquisition from different sensors. A unified physical benchmark and constraint library is constructed based on the hanging basket design parameters and on-site calibration data, providing a consistent physical reference standard for all feature verification and parameter calculations, avoiding systematic errors caused by using different benchmarks in different stages. Extracted component features undergo dual verification using instantaneous thresholds and historical trends, while spatial geometric constraints and temporal motion continuity are combined to filter anomalous features, effectively distinguishing between instantaneous sensor interference and genuine structural deformation, thus improving the reliability of feature data. The calculated state parameters are cross-verified and anomaly corrected through the physical coupling relationship between parameters, resolving contradictory results caused by independent calculation of a single parameter and ensuring the physical consistency of state parameters. Based on the suspicious benchmark point identifiers corresponding to anomalous parameters, the associated constraint thresholds are dynamically tightened, strengthening the monitoring of high-risk areas and enabling proactive tracking and precise location of anomalies. Attached Figure Description
[0016] Figure 1 This is a system flowchart of the hanging basket stability monitoring system of the present invention. Detailed Implementation
[0017] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.
[0018] Please see Figure 1 The system provided by this invention includes a sensor synchronous acquisition module, a physical benchmark library construction module, a feature timing verification module, a parameter coupling correction module, and a verification rule enhancement module. The working process of the system is described in detail below through two specific embodiments.
[0019] Example 1: Stability monitoring under normal construction conditions This embodiment demonstrates the complete monitoring process of the system during normal concrete pouring construction using a hanging basket.
[0020] Before monitoring begins, the physical benchmark library construction module performs initialization. Staff use a total station to precisely measure key feature points such as the front end of the main truss and the four corners of the bottom basket, obtaining on-site calibration coordinates. The system reads the hanging basket design drawings and obtains parameters such as the design hinge point height. The module calculates the benchmark distances between each feature point (e.g., the benchmark distance L12 between the front end point P1 of the main truss and the left front corner point P2 of the bottom basket is 8.00 meters) and benchmark azimuth angles, and sets the tolerance for distance variation between each pair of points (e.g., ΔL_max12 is 0.02 meters) and azimuth angle variation tolerances based on the design stiffness. These data, along with the calculated tilt-displacement conversion coefficients (which depend on the hinge point height), are stored in the hanging basket physical benchmark and constraint library. Once established, this library provides the system with a unique and unchanging physical benchmark throughout the entire monitoring cycle.
[0021] During construction, the sensor synchronous acquisition module triggers the millimeter-wave radar and industrial camera every 100 milliseconds (one monitoring cycle). The radar scan generates a point cloud of the current moment, and the camera captures an image. Both are bound to the same microsecond-precision timestamp and output as a spatiotemporally aligned sensing data stream.
[0022] At monitoring time t=10.0 seconds, the feature timing verification module begins operation. It identifies and calculates the 3D coordinates of point P1 as C_radar from the radar point cloud, and identifies and back-projects the 3D coordinates of point P1 as C_visual from the synchronized image. Next, the module calls the reference data for points P1 and P2 from the physical reference library. First, an instantaneous threshold verification is performed: the distance between P1 and P2 in the current radar data is calculated as L_radar(10.0s) = 8.018 meters, which is 0.018 meters different from the reference distance of 8.00 meters, less than the tolerance of 0.02 meters; this step passes. The azimuth deviation is also calculated and does not exceed the limit. Therefore, the radar observation of point P1 is marked as "preliminarily reliable." The same calculation is performed on the visual data; the distance deviation is 0.015 meters, also within the limit, and the visual observation is also marked as "preliminarily reliable."
[0023] Since all instantaneous checks passed, historical trend checks were not triggered. Ultimately, because both radar and visual observations of point P1 were considered "reliable," the module fused their coordinates (e.g., by averaging) to obtain the final reliable coordinates of point P1. All similar reliable point coordinates constitute a spatiotemporally consistent reliable coordinate set. Subsequently, the parameter coupling correction module called this reliable coordinate set. It calculated the current tilt angle of the basket around the X-axis as θ_m = 0.50 degrees, and the vertical displacement of the front end of the basket as dz_m = 0.008 meters. The module then called the tilt angle reference value θ_c = 0.48 degrees and the displacement reference value dz_c = 0.005 meters output from the previous cycle (t = 9.9 seconds).
[0024] The module performs physical coupling verification: Based on the tilt angle change Δθ = 0.02 degrees and the hinge height H in the physical reference library, a theoretical vertical displacement change can be calculated. Comparing the theoretical change with the measured change (0.003 meters), the calculated coupling error is less than the preset residual allowable threshold. This indicates that the changes in both tilt angle and displacement are self-consistent and conform to physical laws. Therefore, the module directly outputs the currently calculated tilt angle of 0.50 degrees, vertical displacement of 0.008 meters, and other state parameters as the stable state parameter set for this cycle. The verification rule enhancement module detects that the parameters for this cycle are normal and no suspicious point identifiers are recorded; therefore, the enhancement rules are not activated, and the system uses the default tolerance to enter the next cycle.
[0025] Example 2: Addressing Transient Sensor Interference and Active Augmentation Monitoring This embodiment demonstrates how the system can correct errors and proactively enhance the monitoring of potential risk points when the visual sensor is momentarily obstructed.
[0026] During a monitoring period of t=10.5 seconds, a bird briefly flew by, obstructing the industrial camera's view of point P1 at the front end of the main truss. The feature timing verification module, while processing the synchronization data, discovered that the distance between points P1 and P2, obtained from the radar point cloud computing, was 8.022 meters, a deviation of 0.022 meters, slightly exceeding the tolerance by 0.02 meters. The module then initiated historical trend verification: it found that the change in this distance value over the previous three periods was relatively smooth, and the calculated difference between the current instantaneous rate of change and the historical average rate of change was very small, within the normal acceleration range of the hanging basket. Therefore, it was determined that the radar data exceeding the limit was caused by reasonable movement of the hanging basket, and the radar observation was corrected to "reliable".
[0027] However, the visual coordinates of point P1 calculated from the occluded image were severely distorted, with a huge distance deviation from point P2. Not only did instantaneous verification fail, but historical trend verification also showed an abnormal rate of change, far exceeding the reasonable range. Therefore, the visual observation was deemed "unreliable." Due to the inconsistency between radar and visual observations of point P1, point P1 failed to enter the spatiotemporally consistent reliable coordinate set generated in this cycle. The parameter coupling correction module discovered that the reliable coordinate set lacked data for the key point P1, leading to errors in the calculated vertical displacement dz_m. When the module used this displacement value for physical coupling verification with the theoretical value calculated based on tilt angle changes, the calculated coupling error exceeded the allowable threshold. The system determined that there was a physical inconsistency between parameters and initiated the correction procedure.
[0028] The module compares the reliability records of the feature points on which the calculated tilt angle and displacement depend in historical periods and finds that the historical reliability of "tilt angle-related features" is significantly higher than that of "displacement-related features". Therefore, the module selects the currently measured tilt angle as a reliable benchmark (anchor point parameter), recalculates a theoretical vertical displacement value based on the physical relationship (tilt angle-displacement conversion formula), and replaces the erroneous measured displacement value with this theoretical value, outputting the corrected set of steady-state parameters. Simultaneously, the module records the number of feature point P1, on which the displacement calculation depends but is questionable in the current period's data, as an identifier for the questionable benchmark point.
[0029] The verification rule enhancement module activates immediately upon receiving a suspicious flag for point P1. Over the next five consecutive monitoring cycles (10.6 to 11.0 seconds), it initiates an active defense rule: dynamically tightening the associated constraint threshold. Specifically, it retrieves all distance and azimuth variation tolerances related to point P1 from the hanging basket physical benchmark and constraint library, and multiplies their values by a uniform enhancement coefficient of 0.6. For example, the distance tolerance between points P1 and P2 is temporarily tightened from 0.02 meters to 0.012 meters. These temporary tightened tolerances are then sent to the feature timing verification module.
[0030] At t=10.6 seconds, the bird interference disappeared. However, due to the unusually stringent verification criteria (tolerance only 60% of the original), the visual data for point P1 must achieve higher precision to be accepted. This is equivalent to the system conducting a period of "intensive monitoring" and "enhanced check-up" on this point. If point P1 does not cause parameter anomalies again within 5 enhancement cycles, the tolerance will automatically revert to its default value. If point P1 is marked as suspicious again during enhancement, the enhancement cycle will be extended, and the tolerance tightening factor will become smaller (e.g., 0.5), making monitoring more stringent. The module generates rules to dynamically enhance status codes and record this process.
[0031] As can be seen from Examples 1 and 2, the system of the present invention can not only process normal data with high precision, but also achieve self-verification and correction through dual-layer physical constraints (feature layer geometric constraints and parameter layer coupling constraints) when some sensors are interfered with. It also intelligently locks potential problem points for proactive and enhanced monitoring, thereby significantly improving reliability and early warning capabilities in complex engineering environments. The above describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made without departing from the spirit and scope of the invention, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.
Claims
1. A hanging basket stability monitoring system based on radar and vision systems, characterized in that, Includes the following modules: The sensor synchronous acquisition module is used to synchronously acquire radar point clouds and visual images, generate spatiotemporally aligned sensing data streams, and output them. The physical benchmark library construction module is used to set benchmark geometric relationships and constraint thresholds based on the hanging basket design parameters and on-site calibration data, generate and output the hanging basket physical benchmark and constraint library; The feature timing verification module is used to receive the spatiotemporal alignment sensing data stream and the hanging basket physical benchmark and constraint library, perform component feature extraction and dual verification, generate a spatiotemporally consistent reliable coordinate set and output it. The parameter coupling correction module is used to receive the trusted coordinate set, calculate the state parameters by calling it, verify and correct abnormal parameters based on the physical coupling relationship, generate a stable state parameter set, and output suspicious reference point identifiers. The verification rule enhancement module is used to receive the suspicious benchmark point identifier, proportionally tighten the associated constraint threshold in subsequent monitoring cycles, and generate a rule dynamic enhancement status code.
2. The hanging basket stability monitoring system based on radar and vision system according to claim 1, characterized in that: In the sensor synchronous acquisition module, the specific process of generating spatiotemporally aligned sensing data stream is as follows: a built-in high-precision hardware clock generates a unified timestamp, and the main controller generates trigger pulses according to a preset timing sequence and sends them to the radar device and the vision acquisition device at the same time. After receiving the trigger pulse, the radar equipment completes structural scanning, generates radar point cloud, and binds it to the current unified timestamp. After receiving the same trigger pulse, the visual acquisition equipment completes image exposure, generates visual image, and binds it to the same unified timestamp. After binding, the data is sorted and integrated by timestamp, and the output is a spatiotemporally aligned sensor data stream.
3. The hanging basket stability monitoring system based on radar and vision system according to claim 1, characterized in that: In the physical benchmark library construction module, the specific process of generating the physical benchmark and constraint library foundation of the hanging basket is as follows: Multiple measurements of key feature points of the hanging basket are taken using a total station, and the average value is obtained to acquire effective calibration coordinates; the Euclidean distance between any two feature points is calculated based on the effective calibration coordinates as the benchmark distance; the effective calibration coordinates are projected onto the horizontal plane, and the azimuth angle is calculated as the benchmark azimuth angle, forming a benchmark geometric relationship; the inherent stiffness parameters of the structure are extracted from the hanging basket design parameters and matched with a preset safety factor to calculate and generate the distance variation tolerance and azimuth angle variation tolerance for each pair of feature points as constraint thresholds. Integrate the baseline geometric relationship with the constraint threshold to generate the basic part of the hanging basket physical baseline and constraint library.
4. The hanging basket stability monitoring system based on radar and vision system according to claim 3, characterized in that: The specific process of forming a complete physical benchmark and constraint library for the hanging basket in the physical benchmark library construction module is as follows: extract the vertical distance between the main truss rotation hinge point and the front end point of the bottom basket from the hanging basket design drawings as the design hinge point height, calculate the tilt angle-displacement conversion coefficient with the design hinge point height as the only variable, and add the conversion coefficient to the foundation part to form a complete physical benchmark and constraint library for the hanging basket.
5. The hanging basket stability monitoring system based on radar and vision system according to claim 1, characterized in that: In the feature time-series verification module, the specific process of performing component feature extraction and dual verification is as follows: extract feature point coordinates from the spatiotemporal alignment sensing data stream, calculate the measured distance and measured azimuth angle; perform instantaneous threshold verification: call the hanging basket physical reference and constraint library, subtract the measured value from the corresponding reference value to obtain the deviation, and compare it with the corresponding tolerance; perform historical trend verification on feature points with deviation exceeding the limit: retrieve its previous continuous historical measured sequence to calculate the historical average rate of change, calculate the current instantaneous rate of change based on the deviation between the current period and the previous period, and compare the difference between the two with the preset reasonable acceleration threshold; feature points with deviation not exceeding the limit or with deviation exceeding the limit but with the difference in rate of change not exceeding the threshold are determined to pass the verification.
6. The hanging basket stability monitoring system based on radar and vision system according to claim 5, characterized in that: In the feature time-series verification module, the specific process of generating a spatiotemporally consistent reliable coordinate set is as follows: obtain radar and visual measured coordinates for the same feature point respectively and perform the dual verification respectively; only when both sets of data pass the verification, take the average of the two types of coordinates as the final coordinates of the feature point; summarize the final coordinates of all valid feature points to generate the reliable coordinate set.
7. The hanging basket stability monitoring system based on radar and vision system according to claim 1, characterized in that: In the parameter coupling correction module, the specific process of calculating and generating the state parameters in the stable state parameter set is as follows: extract all feature point coordinates from the reliable coordinate set, calculate the average value of each axial coordinate component to obtain the X, Y, and Z axis displacement components of the hanging basket; extract the feature point coordinates of the front and rear ends of the main truss, and calculate the tilt angle around the X axis; use the stable state parameter set output in the previous cycle as a reference, and automatically iterate and update it in each cycle.
8. The hanging basket stability monitoring system based on radar and vision system according to claim 4, characterized in that: In the parameter coupling correction module, the specific process for verifying abnormal parameters is as follows: call the tilt angle-displacement conversion coefficient from the physical reference and constraint library of the hanging basket; calculate the difference between the measured tilt angle around the X-axis in this cycle and the tilt angle around the X-axis in the previous cycle to obtain the tilt angle change; multiply the tilt angle change by the tilt angle-displacement conversion coefficient to obtain the theoretical displacement change. Calculate the measured displacement change; calculate the displacement coupling residual between the theoretical value and the measured value, compare it with the preset residual allowable threshold, and determine the corresponding parameter as an abnormal parameter when the limit is exceeded.
9. The hanging basket stability monitoring system based on radar and vision system according to claim 8, characterized in that: In the parameter coupling correction module, the specific process of correcting abnormal parameters and marking suspicious reference points is as follows: statistically analyze the historical verification pass rates of feature points corresponding to the tilt angle and horizontal displacement around the X-axis, select the one with the higher pass rate as the anchor point reference parameter, and use the tilt angle-displacement conversion coefficient to correct the other parameter. After merging the corrected parameters with the uncorrected parameters, a stable state parameter set is generated, and the benchmark points associated with the abnormal parameters are marked as suspicious benchmark points.
10. The hanging basket stability monitoring system based on radar and vision system according to claim 1, characterized in that: In the verification rule enhancement module, the specific process of generating the rule dynamic enhancement status code is as follows: determine the corresponding benchmark point according to the received suspicious benchmark point identifier, retrieve all the associated constraint thresholds from the hanging basket physical benchmark and constraint library, multiply by the preset first tightening coefficient to generate a temporary tightening tolerance; in the subsequent first preset monitoring period, the temporary tightening tolerance replaces the original constraint threshold; if the same suspicious identifier is repeatedly triggered in the period, the period is extended and the second tightening coefficient is used. The compression coefficient, enhancement cycle, and associated benchmark point information are recorded to generate a rule-based dynamic enhancement status code.