A high-precision bridge pier inspection system and method thereof

Abstract

The present application relates to the technical fields of automation inspection and bridge health monitoring, and particularly relates to a high-precision inspection system for bridge piers and a method thereof, which comprises: a pose perception module, which collects multi-axis inclination data and lifting displacement data of a work platform, generates an inclination deviation value based on the multi-axis inclination data, generates a current height coordinate value based on the lifting displacement data, and outputs sensor connection state information; a leveling calculation module, which receives the inclination deviation value, calculates a posture correction amount when the inclination deviation value is greater than a preset dead zone threshold, and generates a motor duty cycle correction parameter. In the present application, the inclination of the work platform is collected in real time, and the posture correction amount is calculated in combination with the dead zone threshold and the dynamic change rate, and the smoothed correction duty cycle is distributed to each lifting motor, which effectively alleviates the static shaking phenomenon of the flexible suspension platform in the air, and improves the dynamic correction ability of the system against external environmental wind load.

Description

Technical Field

[0001] This invention relates to the field of automated inspection and bridge health monitoring technology, and in particular to a high-precision bridge pier inspection system and method. Background Technology

[0002] In the daily maintenance and long-term safety monitoring of large bridges, the detection of appearance defects (such as cracks and concrete spalling) in the towering facade structures such as piers and towers is a key task in assessing the overall structural health of the bridge. Because bridge piers are typically massive and extremely tall, and often located in complex outdoor environments such as crossing rivers or seas, traditional manual suspended inspection methods are highly dangerous and inefficient. Automated visual inspection systems are gradually becoming the mainstream solution. These inspection systems usually utilize a multi-motor driven work platform based on flexible rope suspension. This platform integrates multiple camera arrays and various pose sensing devices. By controlling lifting motors distributed at various corners of the platform, the flexible work platform can perform large-span vertical lifting and lowering operations, either close to or along the pier facade, thereby achieving efficient and comprehensive visual scanning and defect image acquisition of the vast surface area of ​​the pier.

[0003] Existing flexible suspended inspection platforms often lack adaptive closed-loop adjustment mechanisms for attitude control. When hovering in the air, they are prone to continuous mechanical vibration due to motor drive errors or the flexibility of the ropes. Furthermore, when faced with external interference such as sudden gusts of wind on the river surface, the system's correction response is relatively slow, which can easily cause the work platform to tilt and sway. This results in motion blurring of the inspection images captured by the onboard equipment, affecting the quality of subsequent defect identification. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides a high-precision bridge pier inspection system and method, aiming to improve the problem that existing flexible suspension inspection platforms often lack adaptive closed-loop adjustment means in attitude control.

[0005] In a first aspect, the present invention provides the following technical solution: a high-precision bridge pier inspection system, comprising: The pose perception module collects multi-axis tilt angle data and lifting displacement data of the work platform, generates tilt angle deviation value based on the multi-axis tilt angle data, generates current height coordinate value based on the lifting displacement data, and outputs sensor connection status information. The leveling calculation module receives the tilt angle deviation value, calculates the attitude correction amount when the tilt angle deviation value is greater than the preset dead zone threshold, and generates motor duty cycle correction parameters. The safety verification module receives the motor duty cycle correction parameter and the sensor connection status information. When the sensor connection status information meets the preset conditions and the motor duty cycle correction parameter does not exceed the preset limit threshold, it generates a leveling control command. The lifting control module receives the leveling control command, adjusts the speed of multiple lifting motors on the work platform according to the leveling control command, and outputs a leveling completion signal. The image acquisition module receives the leveling completion signal, triggers multiple cameras to acquire images in response to the leveling completion signal, and generates a cached image with an acquisition timestamp. The filtering and synthesis module acquires the cached image and receives the current height coordinate value. It extracts the sharpness and exposure parameters of the cached image. When the sharpness and exposure parameters meet a preset image threshold, it generates a valid inspection image. It associates the valid inspection image with the current height coordinate value and outputs the inspection record.

[0006] Preferably, in the pose sensing module, the step of outputting sensor connection status information specifically includes: Send data read requests to the tilt sensor according to the preset polling cycle; Determine whether a valid response frame is received from the tilt sensor within a preset timeout window; If the valid response frame is received, the multi-axis tilt data is parsed and output, and the sensor connection status information is marked as normal. If no valid response frame is received and the number of consecutive timeouts reaches the preset disconnection threshold, the sensor connection status information will be marked as abnormal.

[0007] Preferably, in the leveling calculation module, the step of calculating the attitude correction amount and generating the motor duty cycle correction parameters specifically includes: Obtain the absolute value of the tilt angle deviation and compare it with a preset dead zone threshold; When the absolute value is less than or equal to the preset dead zone threshold, the attitude correction amount is set to zero; When the absolute value is greater than the preset dead zone threshold, the tilt angle deviation value is multiplied by a preset proportional coefficient to obtain the attitude correction amount; The attitude correction amount is superimposed on the preset base operating duty cycle to generate motor duty cycle correction parameters.

[0008] Preferably, the step of multiplying the tilt angle deviation value by a preset scaling factor to obtain the attitude correction amount specifically includes: Calculate the difference in tilt angle deviation values ​​within two consecutive acquisition cycles to obtain the deviation change rate; The deviation change rate is compared with a preset wind load sudden change threshold; When the deviation change rate is less than or equal to the preset wind load sudden change threshold, the basic proportional coefficient is used as the preset proportional coefficient multiplied by the tilt angle deviation value; When the deviation change rate is greater than the preset wind load sudden change threshold, the basic proportional coefficient is multiplied by the dynamic gain compensation value that is positively correlated with the deviation change rate to generate an updated preset proportional coefficient, and the updated preset proportional coefficient is multiplied by the tilt deviation value to obtain the attitude correction amount.

[0009] Preferably, in the safety verification module, the step of generating a leveling control command when the sensor connection status information meets preset conditions and the motor duty cycle correction parameter does not exceed a preset limit threshold specifically includes: Determine whether the sensor connection status information is normal. If it is abnormal, intercept the command and forcibly generate a safety braking command with a duty cycle of zero. If it is normal, then determine whether the motor duty cycle correction parameter is greater than the preset limit threshold. If the value is greater than the preset limit threshold, the motor duty cycle correction parameter is forcibly clamped to the preset limit threshold, and a leveling control command is generated. If the value is not greater than the preset limit threshold, a leveling control command is generated directly based on the motor duty cycle correction parameter.

[0010] Preferably, in the lifting control module, the step of adjusting the speed of multiple lifting motors on the work platform according to the leveling control command and outputting a leveling completion signal specifically includes: The leveling control command is analyzed to extract the target duty cycle of each of the lifting motors distributed on the work platform; Based on a preset smooth speed regulation curve, the current duty cycle of each of the lifting motors is controlled to gradually approach the target duty cycle; When the current duty cycle reaches the target duty cycle and the platform tilt angle remains within the preset horizontal range for a set time, a leveling completion signal is output.

[0011] Preferably, in the image acquisition module, the step of triggering multiple cameras to acquire images and generating a buffered image with an acquisition timestamp in response to the leveling completion signal specifically includes: During system standby, the multi-camera system continuously pulls preview video streams and cyclically overwrites the current preview frame and its corresponding timestamp in memory. When the leveling completion signal is received, the physical hardware shooting action of the multi-channel camera is suspended. The preview frame with the largest frame timestamp is directly extracted from the memory as the captured image, and a cached image with the capture timestamp is generated by combining the frame timestamp with the capture timestamp.

[0012] Preferably, the step of directly extracting the preview frame with the largest frame timestamp from the memory as the captured image specifically includes: Obtain the reference timestamp when the leveling completion signal is triggered; Extract the maximum frame timestamp of each of the multiple cameras in the memory, and calculate the absolute value of the time difference between each maximum frame timestamp and the reference timestamp; Determine whether the absolute value of all the time differences is less than a preset concurrent synchronization tolerance threshold; If so, then simultaneously extract the preview frame with the largest image timestamp from the multiple cameras as the captured image; If any of the time differences is greater than or equal to the concurrent synchronization tolerance threshold, the current extraction operation is discarded and the process waits for the next polling.

[0013] Preferably, in the filtering and synthesis module, the step of extracting the sharpness and exposure parameters of the cached image, and generating a valid inspection image when the sharpness and exposure parameters meet a preset image threshold, specifically includes: The gray-level variance of the cached image is calculated as a sharpness parameter, and the mean gray-level histogram of the cached image is calculated as an exposure parameter. Compare the difference between the collected timestamp and the current system time to determine whether the difference is within a preset validity period; When the difference is within the preset validity period, and the sharpness parameter is greater than the preset sharpness threshold and the exposure parameter is within the preset exposure range, the cached image is extracted and marked as a valid inspection image.

[0014] Secondly, the present invention provides the following technical solution: a high-precision inspection method for bridge piers, the method comprising the following steps: S1. Collect multi-axis tilt angle data and lifting displacement data of the work platform, generate tilt angle deviation value based on the multi-axis tilt angle data, generate current height coordinate value based on the lifting displacement data, and output sensor connection status information; S2. Receive the tilt angle deviation value. When the tilt angle deviation value is greater than the preset dead zone threshold, calculate the attitude correction amount and generate the motor duty cycle correction parameter. S3. Receive the motor duty cycle correction parameter and the sensor connection status information. When the sensor connection status information meets the preset conditions and the motor duty cycle correction parameter does not exceed the preset limit threshold, generate a leveling control command. S4. Receive the leveling control command, adjust the speed of multiple lifting motors on the work platform according to the leveling control command, and output a leveling completion signal; S5. Receive the leveling completion signal, and in response to the leveling completion signal, trigger multiple cameras to acquire images and generate a buffered image with an acquisition timestamp. S6. Obtain the cached image and receive the current height coordinate value, extract the sharpness parameter and exposure parameter of the cached image, generate a valid inspection image when the sharpness parameter and the exposure parameter meet the preset image threshold, associate the valid inspection image with the current height coordinate value, and output the inspection record.

[0015] The present invention has the following beneficial effects: 1. In this invention, the tilt angle of the work platform is collected in real time and the attitude correction amount is calculated by combining the dead zone threshold and the dynamic change rate. The smoothed correction duty cycle is then distributed to each lifting motor, which effectively alleviates the static shaking phenomenon of the flexible suspension platform in the air and improves the system's dynamic correction ability against external wind loads, providing a stable physical basis for high-altitude visual inspection.

[0016] 2. In this invention, by continuously overwriting the preview images and corresponding timestamps of multiple cameras in memory, the system cache images that meet the synchronization tolerance are directly compared and extracted after receiving the leveling command. This avoids the defects of warm-up delay and shutter asynchrony that exist in traditional multi-camera hardware triggering, and achieves accurate alignment of multiple visual images in the time dimension, thereby improving the execution efficiency of concurrent acquisition of multi-channel images.

[0017] 3. In this invention, the exposure and sharpness of the image are quantitatively evaluated by using the grayscale mean and variance to automatically remove blurry and unusable images. The valid inspection images are bound to the current height coordinates generated by the displacement sensor in a structured file, so that the final output inspection record directly includes reliable physical spatial positioning information, which reduces the processing cost of manually screening bridge pier surface defect data in the later stage. Attached Figure Description

[0018] Figure 1 This is an architecture diagram of a high-precision bridge pier inspection system proposed in this invention; Figure 2 This is a flowchart of a high-precision bridge pier inspection method proposed in this invention. Detailed Implementation

[0019] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1: In a first embodiment of the present invention, a high-precision bridge pier inspection system is provided, such as... Figure 1 As shown, it includes: The pose perception module collects multi-axis tilt angle data and lifting displacement data of the work platform, generates tilt angle deviation value based on multi-axis tilt angle data, generates current height coordinate value based on lifting displacement data, and outputs sensor connection status information. Furthermore, the step of outputting sensor connection status information in the pose perception module specifically includes: Send data read requests to the tilt sensor according to the preset polling cycle; Determine whether a valid response frame from the tilt sensor has been received within the preset timeout window; If a valid response frame is received, the multi-axis tilt data is parsed and output, and the sensor connection status information is marked as normal. If no valid response frame is received and the number of consecutive timeouts reaches the preset disconnection threshold, the sensor connection status information will be marked as abnormal.

[0021] Specifically, during the initial and operational phases of pier inspection, the pose perception module is responsible for acquiring the basic physical state data of the work platform in real time. The main control unit sends data read requests to the tilt sensors installed on the work platform via a serial communication bus according to a preset polling cycle. In one feasible implementation, the preset polling cycle is configured to be 50 milliseconds. After sending the request, the system determines in real time whether a valid response frame is received from the tilt sensor within a preset timeout window, which is set to 10 milliseconds. The logic for determining a valid response frame involves extracting the checksum of the received data packet and performing a cyclic redundancy check (CRC) operation. If the check passes, it is confirmed as a valid response frame. The system then parses the multi-axis tilt data and marks the sensor connection status information as normal. If no data packet is received or the CRC fails, the system begins to accumulate timeout counts. When the number of consecutive timeouts reaches a preset disconnection threshold, such as three consecutive times, the sensor connection status information is marked as abnormal. This status information is directly input to the subsequent safety verification module to trigger underlying protection actions.

[0022] After successfully resolving the multi-axis tilt data, the pose sensing module generates tilt deviation values ​​based on the multi-axis tilt data. Let the tilt angles resolved by the tilt sensor along the X-axis and Y-axis directions be... and The preset target leveling angles of the work platform along the X and Y axes are respectively and Since the ideal working environment is a perfectly horizontal plane, and All are set to 0 degrees. The formula for calculating the tilt angle deviation is: as well as In the formula and These represent the tilt angle deviations of the working platform in the two orthogonal directions, X and Y, respectively, with positive and negative directional attributes, in degrees. These deviation values ​​objectively characterize the platform's spatial physical attitude deviating from the horizontal plane and will be used as direct input parameters for subsequent closed-loop leveling calculation modules.

[0023] While acquiring attitude data, the pose perception module collects the lifting displacement data of the work platform through a wire-type absolute encoder, and generates the current height coordinate value based on this lifting displacement data. The specific height conversion formula is as follows: In the formula This represents the calculated current height coordinates, in millimeters. The resolution conversion factor representing the encoder is configured to be 0.1 mm per pulse in the implementation. The absolute displacement pulse value represents the value that is linearly proportional to the actual tensile length of the wire rope in the draw encoder. The offset height constant of the working platform at the zero point of the bridge pier lifting is set to 500 mm in practice.

[0024] By fully extracting the three-dimensional spatial attitude and elevation status of the working platform and clarifying the effectiveness of the underlying data transmission link, accurate numerical basis was provided for the subsequent generation of duty cycle correction parameters for driving motors and image elevation correlation.

[0025] The leveling calculation module receives the tilt angle deviation value. When the tilt angle deviation value is greater than the preset dead zone threshold, it calculates the attitude correction amount and generates the motor duty cycle correction parameters. Furthermore, the steps in the leveling calculation module to calculate the attitude correction amount and generate the motor duty cycle correction parameters specifically include: Obtain the absolute value of the tilt angle deviation and compare it with the preset dead zone threshold; When the absolute value is less than or equal to the preset dead zone threshold, the attitude correction amount is set to zero. When the absolute value is greater than the preset dead zone threshold, the tilt deviation value is multiplied by the preset proportional coefficient to obtain the attitude correction amount; The attitude correction amount is superimposed on the preset base operating duty cycle to generate the motor duty cycle correction parameter.

[0026] Furthermore, the step of multiplying the tilt angle deviation value by a preset scaling factor to obtain the attitude correction amount specifically includes: Calculate the difference in tilt angle deviation values ​​within two consecutive acquisition cycles to obtain the deviation change rate; The rate of change of deviation is compared with the preset wind load mutation threshold; When the rate of change of deviation is less than or equal to the preset wind load sudden change threshold, the basic proportional coefficient is used as the preset proportional coefficient multiplied by the tilt angle deviation value. When the rate of change of deviation exceeds the preset wind load change threshold, the basic proportional coefficient is multiplied by the dynamic gain compensation value that is positively correlated with the rate of change of deviation to generate the updated preset proportional coefficient. The updated preset proportional coefficient is then multiplied by the tilt deviation value to obtain the attitude correction amount.

[0027] Specifically, the leveling calculation module receives the tilt angle deviation value output by the pose perception module from the preceding step. This is illustrated using a single-axis control channel as an example; the calculation logic for the other axis is identical. The system first extracts the absolute value of the tilt angle deviation and compares it with a preset dead zone threshold set internally. In the flexible rope suspension environment of bridge pier inspection operations, the preset dead zone threshold is set to 0.2 degrees. When the absolute value is less than or equal to the preset dead zone threshold, the system determines that the current working platform is within the allowable physical level tolerance range, and at this point, the attitude correction amount is directly set to 0.

[0028] When the absolute value exceeds the preset dead zone threshold, the system initiates the calculation of the attitude correction. The system reads the tilt deviation value of the current acquisition cycle and the tilt deviation value of the previous adjacent acquisition cycle, calculates the absolute value of the difference between the two to obtain the deviation change rate, and the specific calculation formula is as follows: In the formula Represents the rate of change of deviation, in degrees; The tilt deviation value for the current cycle, in degrees; This represents the tilt angle deviation value of the previous cycle, in degrees. Since the sampling time period of the underlying hardware of the system is constant, this difference is physically equivalent to representing the absolute angular velocity of the tilt of the working platform when it is disturbed by wind load.

[0029] The system compares the deviation change rate with a preset wind load abrupt change threshold, which is set to 0.5 degrees in practice. When the deviation change rate is less than or equal to the preset wind load abrupt change threshold, it indicates that the environmental wind load is moderate, and the system uses a basic proportional coefficient as the preset proportional coefficient, which is set to 5.0 percentage points per degree in practice. The formula for calculating the attitude correction is as follows: In the formula This represents the calculated attitude correction amount, expressed as a percentage. Represents the basic proportional coefficient; It carries positive and negative polarities that characterize the current physical tilt direction of the platform.

[0030] When the rate of change of deviation exceeds the preset wind load abrupt change threshold, it indicates that the platform is experiencing a sudden strong wind load. The system first calculates the dynamic gain compensation value, which is positively correlated with the rate of change of deviation. The calculation formula is as follows: In the formula Represents the dimensionless dynamic gain compensation value; This represents the gain sensitivity coefficient, measured in units of degree, and is set to 2.0 per degree in practice. This represents the preset wind load abrupt change threshold. The system multiplies the base proportional coefficient by the dynamic gain compensation value to generate the updated preset proportional coefficient. In the formula The unit is a percentage per degree, and the amplified attitude correction is calculated by multiplying the updated preset scaling factor by the tilt deviation value of the current cycle. .

[0031] After obtaining the final attitude correction, the system superimposes it onto the preset base operating duty cycle to generate the motor duty cycle correction parameters. The specific synthesis formula is as follows: In the formula This represents the corrected motor duty cycle parameter after synthesis, expressed as a percentage. The system represents the baseline operating duty cycle, expressed as a percentage, based on the current manual lever position; in practice, it is set to 50%. This motor duty cycle correction parameter will be sent to the control bus after it is generated.

[0032] By introducing a fixed dead zone and a dynamic gain calculation method based on the rate of change of deviation, the system's attitude correction capability against strong sudden wind loads is improved while eliminating the small static jitter of the flexible platform, providing precise duty cycle control data for downstream actions.

[0033] The safety verification module receives motor duty cycle correction parameters and sensor connection status information. When the sensor connection status information meets the preset conditions and the motor duty cycle correction parameters do not exceed the preset limit threshold, it generates a leveling control command. Furthermore, in the safety verification module, the step of generating a leveling control command when the sensor connection status information meets preset conditions and the motor duty cycle correction parameter does not exceed a preset limit threshold specifically includes: Determine if the sensor connection status information is normal. If it is abnormal, intercept the command and forcibly generate a safety braking command with a duty cycle of zero. If it is normal, then determine whether the motor duty cycle correction parameter is greater than the preset limit threshold. If the value exceeds the preset limit threshold, the motor duty cycle correction parameter will be forcibly clamped to the preset limit threshold before a leveling control command is generated. If the value is not greater than the preset limit threshold, the leveling control command is generated directly based on the motor duty cycle correction parameters.

[0034] Specifically, the safety verification module receives sensor connection status information output by the front-end pose perception module, and simultaneously receives motor duty cycle correction parameters output by the front-end leveling calculation module. The system first performs a safety check on the underlying physical feedback link, determining whether the sensor connection status information is normal. If the sensor connection status information is abnormal, it indicates that the current main control unit has lost the actual physical attitude feedback data of the bridge pier facade. Continuing to send closed-loop control data would lead to a risk of the work platform tilting or overturning on one side. At this point, the system intercepts all correction commands at the software logic layer and forcibly generates a safety braking command with a zero duty cycle. After this safety braking command is sent to the driver, it will cut off the drive energy output of the corresponding lifting motor and trigger the mechanical brake, forcing the work platform to hover at the current elevation position.

[0035] If the sensor connection status information is normal, it indicates that the system's attitude feedback link is effective, and the system then enters the parameter limiting determination process. First, the system extracts the absolute value of the motor duty cycle correction parameter and determines whether this absolute value is greater than the preset limiting threshold. In the mechanical characteristics of a flexible rope hoisting system, if a duty cycle drive command close to the limit is applied to the lifting motor instantaneously, it will cause a sudden increase in the tension of the load-bearing rope and an overcurrent trip of the drive circuit. Therefore, the output value of the drive parameter must be forcibly capped. In the specific implementation scenario, the preset limiting threshold is set to 80%.

[0036] When the absolute value of the motor duty cycle correction parameter exceeds the preset limit threshold, the system activates a numerical protection algorithm to forcibly clamp the output duty cycle to the preset limit threshold. The specific clamping assignment formula is as follows: In the formula This represents the duty cycle scalar value of the leveling control command that is permitted to be issued after safety verification, in percentage form. This represents the preset threshold value, expressed as a percentage. Simultaneously, the system extracts the sign bits of the original motor duty cycle correction parameters and maps them to the motor's rotation direction bits; for example, a positive sign maps to forward rotation (upward), and a negative sign maps to reverse rotation (downward). After completing the scalar clamping and direction mapping, the system packages and encapsulates both according to a message format such as the Modbus communication protocol, generating the final leveling control command and sending it to the subsequent lifting control module.

[0037] When the absolute value of the motor duty cycle correction parameter is not greater than the preset limit threshold, it indicates that the drive demand calculated by the front-end module is completely within the safe operating envelope of the hardware. At this time, the system directly extracts the absolute value of the original input, and the assignment formula is... In the formula This represents the signed motor duty cycle correction parameters generated by the pre-leveling calculation module, in percentage form. The system synchronously extracts... The sign bit is mapped to the rotation direction bit, and then the absolute value and direction bit are encapsulated to generate a leveling control command and sent to subsequent modules.

[0038] By constructing a dual protection logic of device status and command limiting between the software algorithm and the underlying actuator, the abnormal device overload action caused by extreme parameter calculation is avoided while clarifying the physical drive direction of the motor, thus ensuring the safety and stability of the platform operation process.

[0039] The lifting control module receives leveling control commands, adjusts the speed of multiple lifting motors on the work platform according to the leveling control commands, and outputs a leveling completion signal. Furthermore, the lifting control module includes the following steps: adjusting the speed of multiple lifting motors on the work platform according to the leveling control command and outputting a leveling completion signal. Analyze the leveling control commands and extract the target duty cycle of each lifting motor distributed on the work platform; Based on the preset smooth speed regulation curve, the current duty cycle of each lifting motor is controlled to gradually approach the target duty cycle; Once the current duty cycle reaches the target duty cycle and the platform tilt angle remains within the preset horizontal range for a set time, a leveling completion signal is output.

[0040] Specifically, the lifting control module receives the leveling control command sent by the pre-installed safety verification module via the communication bus. The system parses the leveling control command message to extract the target duty cycle and corresponding rotation direction of each lifting motor located at the four corners of the work platform. The parsed target duty cycle is the safety scalar value after the amplitude limiting clamping process in the pre-installed steps. The system assigns the extracted target duty cycle and rotation direction to the corresponding motor drive control channel, preparing to execute the physical drive action.

[0041] To avoid severe shaking and mechanical impact on the flexible suspension ropes of the work platform caused by sudden abrupt changes in motor speed, the system controls the current duty cycle of each lifting motor to gradually approach the target duty cycle based on a preset smooth speed regulation curve. In specific implementation, a first-order inertial digital filtering algorithm is used to achieve the smooth approximation process, and its discrete iterative calculation formula is as follows: In the formula, Represents the current control cycle number in the discrete-time domain; it is dimensionless. Represents the sequence number of the previous control cycle; The actual output duty cycle of the current control cycle, expressed as a percentage. This represents the actual output duty cycle of the previous control cycle, expressed as a percentage. This represents the target duty cycle extracted from the parsing, expressed as a percentage. The dimensionless smoothing damping coefficient is configured to 0.15 in practice. This iterative calculation is performed cyclically within each constant control cycle at the system's underlying layer, ensuring that the duty cycle of the actual output to the driver changes smoothly with the characteristics of an exponential curve, preventing sudden increases or decreases in transient drive energy.

[0042] While performing smooth speed control, the system monitors the status of drive parameters and the spatial attitude of the work platform in real time. The system calculates the absolute value of the difference between the actual output duty cycle and the target duty cycle for the current cycle. When this absolute value is less than the system's set approximation tolerance (e.g., less than 1%), it determines that the current duty cycle has reached the target duty cycle in a physical engineering sense, and then initiates the attitude stabilization judgment logic. The system continuously obtains the current absolute values ​​of the actual tilt angles of the work platform along the X-axis and Y-axis from the pose sensing module and determines whether both absolute values ​​remain within a preset horizontal range. In practice, this preset horizontal tolerance range is defined as less than or equal to 0.3 degrees. If the absolute values ​​of the actual tilt angles of both the X-axis and Y-axis are within this range, the system's internal stabilization timer accumulates the time periodically; if the absolute value of the tilt angle in any axis momentarily exceeds this range, the stabilization timer is immediately reset to zero. When the stabilization timer has accumulated for a set time (e.g., 2000 milliseconds), the system determines that the physical shaking generated by this closed-loop correction has completely decayed, the platform has reached its final static state, and then generates a leveling completion signal and outputs it to the subsequent image acquisition module.

[0043] By smoothly converting control data into underlying mechanical motion, the rigid impact of the motor drive stage is eliminated, and the static state of the platform is confirmed at the physical level through dual-axis joint judgment, providing a stable hardware environment for subsequent high-quality image capture.

[0044] The image acquisition module receives the leveling completion signal, triggers multiple cameras to acquire images in response to the leveling completion signal, and generates a buffered image with an acquisition timestamp. Furthermore, in the image acquisition module, the step of triggering multiple cameras to acquire images and generating a buffered image with an acquisition timestamp in response to the leveling completion signal specifically includes: During system standby, control multiple cameras to continuously pull preview video streams and cyclically overwrite the current preview frame and its corresponding timestamp in memory; When the leveling completion signal is received, the physical hardware shooting action of the multiple cameras is suspended. The preview frame with the largest frame timestamp is directly extracted from memory as the captured image, and a cached image with the capture timestamp is generated by combining the frame timestamp with the capture timestamp.

[0045] Furthermore, the step of directly extracting the preview frame with the largest frame timestamp from memory as the captured image specifically includes: Obtain the reference timestamp when the leveling completion signal is triggered; Extract the maximum frame timestamp of each of the multiple cameras in memory, and calculate the absolute value of the time difference between each maximum frame timestamp and the base timestamp; Determine whether the absolute value of all time differences is less than the preset concurrent synchronization tolerance threshold; If so, then simultaneously extract the preview frame with the largest image timestamp from multiple cameras as the captured image; If any time difference has an absolute value greater than or equal to the concurrent synchronization tolerance threshold, the current extraction operation is discarded and the process waits for the next poll.

[0046] Specifically, the image acquisition module receives the leveling completion signal output by the front-mounted lifting control module. To overcome the warm-up delay and mechanical shutter asynchrony issues encountered when multiple cameras are triggered concurrently in traditional hardware, the system pre-activates the underlying visual polling mechanism during standby and platform movement phases. The main control unit controls multiple cameras to continuously pull preview video streams facing the bridge pier facade and allocates an independent buffer for each camera in system memory, cyclically overwriting the current preview frame. Simultaneously, the system reads the main control unit's system clock the instant the image is stored in memory, generates a corresponding image timestamp, and binds it to that frame, ensuring that all time data is within the same system clock domain.

[0047] When the system receives the leveling completion signal, it indicates that the physical attitude of the work platform has reached a stable state. At this time, the system suspends the physical hardware shooting instructions of multiple cameras at the software logic layer, abandons the time-consuming hardware image capture interface, and instead triggers a memory-based concurrent synchronous extraction algorithm. The system first reads the absolute time of the main control unit system at the moment the leveling completion signal is triggered as the reference timestamp, denoted as . The unit is milliseconds.

[0048] The system then iterates through the memory cache, extracting the timestamp of the image with the maximum value corresponding to each of the multiple cameras, and denoting it as... In the formula, the subscript This represents the camera channel number in a multi-camera array, and its value range is... ,in This represents the total number of multi-channel cameras configured in the system, and is a positive integer; for example, it is set to 12 in the implementation. Representing the The system time, in milliseconds, is the time it takes for the latest preview frame within each channel to be written to memory. The system calculates the absolute value of the time difference between the maximum frame timestamp and the baseline timestamp for each channel using the following formula: In the formula Representing the The absolute time deviation between the image generation time of each camera channel and the platform's physical stability reference point, expressed in milliseconds.

[0049] After calculating the absolute value of the time difference for all channels, the system performs a batch comparison with a preset concurrent synchronization tolerance threshold. Let the concurrent synchronization tolerance threshold be... In the implementation of the panoramic stitching project for the bridge piers, to ensure that the stitching edges of the images captured by adjacent cameras do not experience motion tearing, this... Set to 30 milliseconds. The system determines whether this applies to all... ,That All strictly less than If so, it indicates that the latest images currently residing in memory from all cameras were generated within a stable and valid time window of the platform. The system then simultaneously extracts the preview frame with the largest image timestamp from multiple cameras as the actual captured image, and combines it with the corresponding image timestamp to generate a cached image with an acquisition timestamp, which is then pushed to the subsequent filtering and compositing process.

[0050] If during the comparison process, there is any channel... Greater than or equal to This indicates that due to network latency or bus bandwidth congestion, the multi-camera images in memory currently exhibit discrete deviations on the time axis. To avoid generating misaligned panoramic images, the system directly discards this global extraction operation and waits for the image refresh in the next visual polling cycle until the absolute value of the time difference of all channels meets the synchronization condition.

[0051] By using a memory preview frame cyclic overwrite and a strict absolute difference verification algorithm based on the same system clock, the traditional physical hardware triggering process is replaced, enabling ultra-fast synchronous extraction of multiple cameras and ensuring high consistency of the output image array in both time and space dimensions.

[0052] The filtering and compositing module acquires cached images and receives the current height coordinates. It extracts the sharpness and exposure parameters of the cached images. When the sharpness and exposure parameters meet the preset image thresholds, it generates valid inspection images, associates the valid inspection images with the current height coordinates, and outputs the inspection records.

[0053] Furthermore, in the screening and synthesis module, the step of extracting the sharpness and exposure parameters of the cached image and generating a valid inspection image when the sharpness and exposure parameters meet a preset image threshold specifically includes: The gray-level variance of the cached image is calculated as the sharpness parameter, and the mean gray-level histogram of the cached image is calculated as the exposure parameter. Compare the difference between the collected timestamp and the current system time to determine whether the difference is within the preset validity period; When the difference is within the preset validity period, and the sharpness parameter is greater than the preset sharpness threshold and the exposure parameter is within the preset exposure range, the cached image is extracted and marked as a valid inspection image.

[0054] Specifically, the filtering and synthesis module receives cached images with acquisition timestamps pushed by the front-end image acquisition module, and simultaneously receives the current height coordinates continuously updated by the pose perception module on the system data bus. To avoid the extracted images being outdated due to multi-threaded queuing processing, the system first compares the difference between the acquisition timestamp carried by the cached image and the current system time to determine if the difference is within a preset validity period. The calculation formula is as follows: In the formula Represents the survival time difference, in milliseconds; This represents the current system time of the main control unit when this screening action is performed, in milliseconds; This represents the capture timestamp carried by the cached image, in milliseconds. (System judgment) The system checks whether the difference is less than or equal to a preset validity period, which is configured to be 5000 milliseconds in practice. If the difference is within the validity period, the system allows the image to proceed to the subsequent pixel-level quality evaluation process.

[0055] To remove blur caused by minor vibrations from the bridge surface and unusable images due to sudden changes in outdoor lighting, the system extracts the sharpness and exposure parameters from the cached images. The system converts the cached images into single-channel grayscale images, first calculating the mean of the grayscale histogram of the entire image as the exposure parameter. The calculation formula is as follows: In the formula This represents the calculated exposure parameter, which is dimensionless and represents the global average brightness. Represents the total horizontal pixel width of the image, in pixels; Represents the total vertical pixel height of the image, in pixels; The x-coordinate index of the current pixel is a positive integer. The y-coordinate index of the current pixel is a positive integer. The image represents the horizontal axis. and the vertical axis The pixel grayscale value at that location is an integer between 0 and 255.

[0056] Subsequently, the system calculates the grayscale variance of the entire image based on this exposure parameter as the sharpness parameter, using the following formula: In the formula This represents the calculated sharpness parameter, which is dimensionless. This variance value objectively reflects the dispersion of image pixel gray levels; the larger the value, the richer the high-frequency details such as the inner edge contours of the image, and the sharper the image.

[0057] After obtaining the sharpness and exposure parameters, the system compares them with a preset image threshold. The system then determines the sharpness parameter. Whether it exceeds a preset sharpness threshold, which is set to 800 in practice; simultaneously, the exposure parameter is also checked. Whether it is within the preset exposure range, in practice, the preset exposure range is defined as greater than or equal to 80 and less than or equal to 180. When the time survival difference meets the preset validity period, the sharpness parameter is greater than the preset sharpness threshold, and the exposure parameter is within the preset exposure range, the system determines that the image is not affected by physical motion blur and overexposure interference, and then extracts the cached image and marks it as a valid inspection image.

[0058] Finally, the system associates the valid inspection image with the current elevation coordinates at the file level. In one specific implementation, the system renames the valid inspection image according to a structured naming convention, with the naming format including the device number, acquisition timestamp, and current elevation coordinates. For example, a string containing millimeter-level elevation data is appended to the image file name. In another implementation, the system extracts the exchanged image file format information area of ​​the valid inspection image and writes the current elevation coordinates into this header information area. The system generates an inspection file containing elevation positioning data through the above association method, outputs it, and stores it as the final pier inspection record.

[0059] By using time lifecycle verification and pixel-level image quality quantization algorithms, images with dynamic blur and abnormal exposure were filtered out, and the defective images were linked to the physical height of the bridge piers, thus improving the usability of inspection data.

[0060] Example 2: This invention also provides a high-precision inspection method for bridge piers, such as... Figure 2 As shown, the method includes the following steps: S1. Collect multi-axis tilt angle data and lifting displacement data of the work platform, generate tilt angle deviation value based on multi-axis tilt angle data, generate current height coordinate value based on lifting displacement data, and output sensor connection status information; S2. Receive tilt angle deviation value. When the tilt angle deviation value is greater than the preset dead zone threshold, calculate the attitude correction amount and generate motor duty cycle correction parameters. S3. Receive motor duty cycle correction parameters and sensor connection status information. When the sensor connection status information meets the preset conditions and the motor duty cycle correction parameters do not exceed the preset limit threshold, generate a leveling control command. S4. Receive leveling control command, adjust the speed of multiple lifting motors on the work platform according to the leveling control command, and output leveling completion signal; S5. Receive the leveling completion signal, and in response to the leveling completion signal, trigger multiple cameras to acquire images and generate a buffered image with an acquisition timestamp. S6. Obtain the cached image and receive the current height coordinate value. Extract the sharpness and exposure parameters of the cached image. When the sharpness and exposure parameters meet the preset image threshold, generate a valid inspection image. Associate the valid inspection image with the current height coordinate value and output the inspection record.

[0061] Specifically, the system first collects multi-axis tilt angle data and lifting displacement data from the work platform. Based on the multi-axis tilt angle data, it calculates and generates a tilt angle deviation value. Based on the lifting displacement data, it converts this into the current height coordinate value and outputs the underlying sensor connection status information in real time. Then, the system receives the tilt angle deviation value. When the tilt angle deviation value is determined to be greater than a preset dead zone threshold, the control algorithm is activated to calculate the attitude correction amount and convert it into motor duty cycle correction parameters. Based on this, the system receives the motor duty cycle correction parameters and sensor connection status information. When it confirms that the sensor connection status information meets preset normal conditions and the motor duty cycle correction parameters do not exceed a preset limit threshold, a safe leveling control command is generated. Next, the underlying drive module receives the adjustment... The system uses leveling control commands to smoothly adjust the speed of multiple lifting motors on the work platform, and outputs a leveling completion signal after the physical platform's attitude stabilizes. Upon receiving the leveling completion signal, the system triggers multiple cameras to concurrently acquire images, extracts and generates cached images with acquisition timestamps from memory, and finally acquires the cached images while simultaneously receiving the current height coordinates. The system extracts the sharpness and exposure parameters of the cached images, and determines and generates valid inspection images when the sharpness and exposure parameters meet preset image thresholds. The valid inspection images are then linked and bound to the current height coordinates at the file level, thereby outputting a bridge pier inspection record containing precise elevation positioning.

[0062] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A high-precision bridge pier inspection system, characterized in that, include: The pose perception module collects multi-axis tilt angle data and lifting displacement data of the work platform, generates tilt angle deviation value based on the multi-axis tilt angle data, generates current height coordinate value based on the lifting displacement data, and outputs sensor connection status information. The leveling calculation module receives the tilt angle deviation value, calculates the attitude correction amount when the tilt angle deviation value is greater than the preset dead zone threshold, and generates motor duty cycle correction parameters. The safety verification module receives the motor duty cycle correction parameter and the sensor connection status information. When the sensor connection status information meets the preset conditions and the motor duty cycle correction parameter does not exceed the preset limit threshold, it generates a leveling control command. The lifting control module receives the leveling control command, adjusts the speed of multiple lifting motors on the work platform according to the leveling control command, and outputs a leveling completion signal. The image acquisition module receives the leveling completion signal, triggers multiple cameras to acquire images in response to the leveling completion signal, and generates a cached image with an acquisition timestamp. The filtering and synthesis module acquires the cached image and receives the current height coordinate value. It extracts the sharpness and exposure parameters of the cached image. When the sharpness and exposure parameters meet a preset image threshold, it generates a valid inspection image. It associates the valid inspection image with the current height coordinate value and outputs the inspection record.

2. The high-precision bridge pier inspection system according to claim 1, characterized in that, The step of outputting the sensor connection status information in the pose perception module specifically includes: Send data read requests to the tilt sensor according to the preset polling cycle; Determine whether a valid response frame is received from the tilt sensor within a preset timeout window; If the valid response frame is received, the multi-axis tilt data is parsed and output, and the sensor connection status information is marked as normal. If no valid response frame is received and the number of consecutive timeouts reaches the preset disconnection threshold, the sensor connection status information will be marked as abnormal.

3. The high-precision bridge pier inspection system according to claim 1, characterized in that, The step of calculating the attitude correction amount and generating the motor duty cycle correction parameters in the leveling calculation module specifically includes: Obtain the absolute value of the tilt angle deviation and compare it with a preset dead zone threshold; When the absolute value is less than or equal to the preset dead zone threshold, the attitude correction amount is set to zero; When the absolute value is greater than the preset dead zone threshold, the tilt angle deviation value is multiplied by a preset proportional coefficient to obtain the attitude correction amount; The attitude correction amount is superimposed on the preset base operating duty cycle to generate motor duty cycle correction parameters.

4. The high-precision bridge pier inspection system according to claim 3, characterized in that, The step of multiplying the tilt angle deviation value by a preset scaling factor to obtain the attitude correction amount specifically includes: Calculate the difference in tilt angle deviation values ​​within two consecutive acquisition cycles to obtain the deviation change rate; The deviation change rate is compared with a preset wind load sudden change threshold; When the deviation change rate is less than or equal to the preset wind load sudden change threshold, the basic proportional coefficient is used as the preset proportional coefficient multiplied by the tilt angle deviation value; When the deviation change rate is greater than the preset wind load sudden change threshold, the basic proportional coefficient is multiplied by the dynamic gain compensation value that is positively correlated with the deviation change rate to generate an updated preset proportional coefficient, and the updated preset proportional coefficient is multiplied by the tilt deviation value to obtain the attitude correction amount.

5. The high-precision bridge pier inspection system according to claim 1, characterized in that, In the safety verification module, the step of generating a leveling control command when the sensor connection status information meets preset conditions and the motor duty cycle correction parameter does not exceed a preset limit threshold specifically includes: Determine whether the sensor connection status information is normal. If it is abnormal, intercept the command and forcibly generate a safety braking command with a duty cycle of zero. If it is normal, then determine whether the motor duty cycle correction parameter is greater than the preset limit threshold. If the value is greater than the preset limit threshold, the motor duty cycle correction parameter is forcibly clamped to the preset limit threshold, and a leveling control command is generated. If the value is not greater than the preset limit threshold, a leveling control command is generated directly based on the motor duty cycle correction parameter.

6. The high-precision bridge pier inspection system according to claim 1, characterized in that, In the lifting control module, the step of adjusting the speed of multiple lifting motors on the work platform according to the leveling control command and outputting a leveling completion signal specifically includes: The leveling control command is analyzed to extract the target duty cycle of each of the lifting motors distributed on the work platform; Based on a preset smooth speed regulation curve, the current duty cycle of each of the lifting motors is controlled to gradually approach the target duty cycle; When the current duty cycle reaches the target duty cycle and the platform tilt angle remains within the preset horizontal range for a set time, a leveling completion signal is output.

7. The high-precision bridge pier inspection system according to claim 1, characterized in that, In the image acquisition module, the step of triggering multiple cameras to acquire images and generating a buffered image with an acquisition timestamp in response to the leveling completion signal specifically includes: During system standby, the multi-camera system continuously pulls preview video streams and cyclically overwrites the current preview frame and its corresponding timestamp in memory. When the leveling completion signal is received, the physical hardware shooting action of the multi-channel camera is suspended. The preview frame with the largest frame timestamp is directly extracted from the memory as the captured image, and a cached image with the capture timestamp is generated by combining the frame timestamp with the capture timestamp.

8. A high-precision bridge pier inspection system according to claim 7, characterized in that, The step of directly extracting the preview frame with the largest frame timestamp from the memory as the captured image specifically includes: Obtain the reference timestamp when the leveling completion signal is triggered; Extract the maximum frame timestamp of each of the multiple cameras in the memory, and calculate the absolute value of the time difference between each maximum frame timestamp and the reference timestamp; Determine whether the absolute value of all the time differences is less than a preset concurrent synchronization tolerance threshold; If so, then simultaneously extract the preview frame with the largest image timestamp from the multiple cameras as the captured image; If any of the time differences is greater than or equal to the concurrent synchronization tolerance threshold, the current extraction operation is discarded and the process waits for the next polling.

9. A high-precision bridge pier inspection system according to claim 1, characterized in that, In the filtering and synthesis module, the step of extracting the sharpness and exposure parameters of the cached image, and generating a valid inspection image when the sharpness and exposure parameters meet a preset image threshold, specifically includes: The gray-level variance of the cached image is calculated as a sharpness parameter, and the mean gray-level histogram of the cached image is calculated as an exposure parameter. Compare the difference between the collected timestamp and the current system time to determine whether the difference is within a preset validity period; When the difference is within the preset validity period, and the sharpness parameter is greater than the preset sharpness threshold and the exposure parameter is within the preset exposure range, the cached image is extracted and marked as a valid inspection image.

10. A high-precision inspection method for bridge piers, characterized in that, The method for a high-precision bridge pier inspection system according to any one of claims 1-9 includes the following steps: S1. Collect multi-axis tilt angle data and lifting displacement data of the work platform, generate tilt angle deviation value based on the multi-axis tilt angle data, generate current height coordinate value based on the lifting displacement data, and output sensor connection status information; S2. Receive the tilt angle deviation value. When the tilt angle deviation value is greater than the preset dead zone threshold, calculate the attitude correction amount and generate the motor duty cycle correction parameter. S3. Receive the motor duty cycle correction parameter and the sensor connection status information. When the sensor connection status information meets the preset conditions and the motor duty cycle correction parameter does not exceed the preset limit threshold, generate a leveling control command. S4. Receive the leveling control command, adjust the speed of multiple lifting motors on the work platform according to the leveling control command, and output a leveling completion signal; S5. Receive the leveling completion signal, and in response to the leveling completion signal, trigger multiple cameras to acquire images and generate a buffered image with an acquisition timestamp. S6. Obtain the cached image and receive the current height coordinate value, extract the sharpness parameter and exposure parameter of the cached image, generate a valid inspection image when the sharpness parameter and the exposure parameter meet the preset image threshold, associate the valid inspection image with the current height coordinate value, and output the inspection record.

Images