Wireless transmission calibration method for visual data in prefabricated steel structure hoisting

By synchronizing the visual data acquisition clock with the hoisting swing phase, and utilizing multipath delay spectrum analysis and differential error correction coding, the mismatch between visual data acquisition and transmission in prefabricated steel structure hoisting was solved, improving image quality and transmission reliability.

CN123075486APending Publication Date: 2026-11-10CHINA RAILWAY BEIJING ENG BUREAU GP OR GRP BEIJING CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610913378.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-06-24
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

During the hoisting of prefabricated steel structures, the timing of visual data acquisition does not match the hoisting swing rhythm, resulting in image quality degradation and poor wireless transmission reliability. Existing error correction coding strategies lack adaptability and cannot effectively cope with complex multipath environments.

Method used

By analyzing the hoisting motion to obtain the swing period and amplitude, synchronizing the visual data acquisition clock with the hoisting swing phase, constructing a dynamic confidence interval using multipath delay spectrum analysis, and performing differentiated forward error correction coding to generate a calibrated visual data stream.

Benefits of technology

It improves the image clarity of hoisting nodes and the anti-interference capability of data links, ensuring real-time and reliable transmission of visual data under complex and dynamic working conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123075486A_ABST
    Figure CN123075486A_ABST
Patent Text Reader

Abstract

This invention discloses a method for wireless transmission calibration of visual data for prefabricated steel structure hoisting, belonging to the field of visual monitoring technology for prefabricated steel structure hoisting. The method includes: analyzing real-time hoisting actions during the prefabricated steel structure hoisting process to obtain the swing period and swing amplitude; based on the swing period and swing amplitude, synchronizing the sampling clock of the visual data acquisition terminal with the swing phase to generate a synchronization trigger signal; acquiring raw visual data packets from the hoisting nodes based on the synchronization trigger signal; performing multipath delay spectrum analysis on the raw visual data packets to extract signal characteristic parameters for each transmission path; constructing a dynamic confidence interval based on the signal characteristic parameters, and segmenting the raw visual data packets according to the dynamic confidence interval; and using the results of the segmented confidence marking, performing differentiated forward error correction coding on the raw visual data packets to generate a calibrated visual data stream.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of visual monitoring technology for prefabricated steel structure hoisting, specifically to a method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting. Background Technology

[0002] During the hoisting of prefabricated steel structures, real-time monitoring of the attitude and position of hoisting nodes is required using wirelessly transmitted visual data. The hoisted object oscillates periodically under wind loads and sling tension, causing motion blur and framing errors in the acquired visual images. Simultaneously, the wireless channel is affected by multipath reflections and obstructions at the hoisting site, making data packets prone to errors and loss during transmission.

[0003] Existing visual data acquisition methods use a fixed-frequency sampling clock, which is not coupled with the swing rhythm of the hoisting equipment. This results in uneven effective information density in the acquired image frames, insufficient visual data acquisition at key swing positions, and increased bias in subsequent pose calculations. Wireless transmission typically employs a uniform forward error correction coding strategy, which lacks adaptability to channel state changes and cannot provide targeted protection for high-error-risk data segments, leading to a trade-off between transmission robustness and spectral efficiency. Channel state acquisition methods primarily rely on statistical average bit error rate feedback, lacking real-time mapping and fine-grained utilization of multipath delay structures. Confidence assessment granularity is coarse, making it difficult to support differentiated transmission protection mechanisms.

[0004] To suppress image quality degradation caused by hoisting sway, it is necessary to address the separation between visual acquisition timing and sway phase, ensuring that the exposure time is accurately locked within a specific phase range of the sway trajectory. To improve the reliability of wireless transmission in complex multipath environments, it is necessary to address the disconnect between confidence markers and channel physical characteristics, enabling error correction and protection strategies to implement segmented, granular differential coding of visual data based on changes in the channel's multipath structure. Summary of the Invention

[0005] This invention provides a method for wireless transmission calibration of visual data for prefabricated steel structure hoisting. The purpose is to incorporate the hoisting swing characteristics into the visual acquisition clock and wireless transmission protection mechanism to achieve phase synchronization between the visual data acquisition rhythm and the hoisting swing. Furthermore, based on the multipath delay spectrum, a dynamic confidence interval is constructed to mark the confidence level of the data segments. Then, differentiated forward error correction coding is performed on the high and low confidence data segments to improve the acquisition quality and transmission reliability of visual data under complex dynamic working conditions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a method for wireless transmission calibration of visual data for prefabricated steel structure hoisting, the method comprising: analyzing the real-time hoisting action during the prefabricated steel structure hoisting process to obtain the swing period and swing amplitude of the hoisting action, thereby providing a precise swing time-frequency reference for the synchronous calibration of subsequent visual data acquisition and wireless transmission; according to the swing period and swing amplitude, performing swing phase synchronization binding on the sampling clock of the visual data acquisition end to generate a synchronization trigger signal, thereby effectively eliminating image motion blur and inter-frame jitter caused by large-amplitude swing by locking the exposure time of the visual sensor to a specific phase point of the hoisting swing, ensuring the clarity and consistency of visual data from the source of acquisition; acquiring the original visual data packets of the hoisting node based on the synchronization trigger signal; and further... The original visual data packets undergo multipath delay spectrum analysis to extract signal characteristic parameters for each transmission path, enabling real-time perception of the complex multipath propagation state of the wireless channel caused by component obstruction and reflection in the steel structure hoisting environment. A dynamic confidence interval is constructed based on the signal characteristic parameters, and the original visual data packets are segmented and labeled with confidence levels according to these dynamic confidence intervals. The reliability of the data after multipath fading is quantified segment by segment, accurately correlated with the risk level of the data segment. Using the results of the segmented confidence labeling, differentiated forward error correction coding is performed on the original visual data packets. Stronger error correction redundancy is configured for low-confidence data segments, while lower redundancy overhead is configured for high-confidence data segments, thereby achieving an optimized balance between overall coding efficiency and error resistance performance, generating a calibrated visual data stream.

[0007] Preferably, the real-time hoisting action analysis of the prefabricated steel structure hoisting process specifically includes: acquiring triaxial angular velocity and triaxial acceleration data output by an inertial measurement unit installed at the root of the hoisting ropes; performing time series integration on the triaxial angular velocity and triaxial acceleration data to obtain the swing trajectory curve of the hoisted object in three-dimensional space; extracting the time interval between two adjacent zero-crossing points from the swing trajectory curve as the swing period, and calculating the peak offset within a single swing period as the swing amplitude. The inertial measurement unit employs a six-axis MEMS sensor and outputs data at a sampling frequency of not less than 200Hz, ensuring high resolution and real-time performance of the swing parameter extraction.

[0008] As a technical solution of the present invention, the step of generating a synchronization trigger signal by synchronizing the sampling clock of the visual data acquisition terminal with the swing phase according to the swing period and swing amplitude includes: using the swing period as the length parameter of the sampling time window and the swing amplitude as the phase offset base of the sampling time window; using a phase-locked loop circuit to track the real-time change of the swing period and generate a reference clock with the same frequency as the swing period; using the swing amplitude as the offset, generating a pulse at the corresponding phase point in each cycle of the reference clock; and shaping and amplifying the pulse to serve as the synchronization trigger signal, which is then output to the shutter control pin of the visual data acquisition terminal. The proportional-integral parameter of the phase-locked loop circuit is adaptively adjusted according to the rate of change of the swing period to keep the phase error between the reference clock and the swing period within ±2 degrees, ensuring that the synchronization trigger pulse is highly locked to the swing phase of the suspended object. Furthermore, when acquiring the original visual data packets of the hoisting node based on the synchronous trigger signal, the synchronous trigger signal is used as the frame exposure start signal of the visual sensor. Pixel charge accumulation is performed during the high level of the synchronous trigger signal, and the analog-to-digital converter is triggered on the falling edge to convert the accumulated pixel charge into digital image data. Multiple frames of digital image data acquired in multiple consecutive synchronous trigger signal cycles are packaged into an original visual data packet in chronological order, and the swing period value and swing amplitude value corresponding to the current acquisition cycle are embedded in the header of each original visual data packet, so that the data packet itself carries the swing state information at the acquisition time, providing a correlation basis for subsequent channel estimation and coding decisions.

[0009] Furthermore, the method for performing multipath delay spectrum analysis on the original visual data packets and extracting signal characteristic parameters for each transmission path is as follows: a probe reference signal is broadcast to the wireless channel in the hoisting area, and channel impulse response sequences from multiple receiving antennas are received; a sliding correlation operation is performed on the channel impulse response sequences to separate multipath components at different time delays; the arrival time and power attenuation value of each multipath component are recorded as the signal characteristic parameters; the multipath components are sorted according to the arrival time to generate a multipath delay spectrum, and the primary path component with the smallest power attenuation value and the secondary path component with the largest power attenuation value are marked. The probe reference signal adopts the Zadoff-Chu sequence and is broadcast within a guard interval before each original visual data packet is sent to avoid interference with the effective data payload and to obtain an accurate estimate of the current channel impulse response.

[0010] When constructing a dynamic confidence interval and performing segmented confidence labeling, the arrival time of the primary path component is used as the center point, and the difference between the arrival time of the secondary path component and the arrival time of the primary path component is used as the interval radius to construct the dynamic confidence interval. The original visual data packet is divided into multiple data segments by bytes, and each data segment's corresponding delay during transmission is checked to see if it falls within the dynamic confidence interval. If it falls within the interval, the data segment is labeled with a high confidence label; otherwise, it is labeled with a low confidence label. This dynamic confidence interval is constructed based on the real-time extracted primary and secondary path delay differences, and can adaptively follow the instantaneous changes in channel multipath delay spread, so that the confidence label accurately reflects the severity of inter-symbol interference suffered by each data segment.

[0011] Based on this, the step of performing differentiated forward error correction coding using the results of the segmented confidence marking includes: extracting all data segments carrying high-confidence labels from the original visual data packet to form a first data set, and extracting all data segments carrying low-confidence labels to form a second data set; applying a first redundancy coding rate to the first data set to generate first check redundancy data, and applying a higher second redundancy coding rate to the second data set to generate second check redundancy data; interleaving the first data set, the second data set, the first check redundancy data, and the second check redundancy data according to the order of the original data segments to form the calibrated visual data stream. By nesting error correction codes with different protection strengths in the data stream, the risk of error propagation of low-confidence data segments during wireless transmission is significantly reduced, while high-confidence data segments are transmitted with low redundancy, saving channel bandwidth.

[0012] Preferably, after generating the calibrated visual data stream, the method further includes: sending the calibrated visual data stream into an orthogonal frequency division multiplexing modulator and mapping it onto multiple subcarriers; dynamically adjusting the modulation and coding scheme level of the multiple subcarriers according to the swing period and swing amplitude. When the swing amplitude exceeds a preset severe swing threshold, the modulation and coding scheme level is switched to the most robust level, and the cyclic prefix length of the orthogonal frequency division multiplexing modulator is reduced to cope with the adverse effects of increased channel delay spread. The specific process of dynamically adjusting the modulation and coding scheme level is as follows: establishing a mapping table containing a ternary correspondence between the swing amplitude range, the swing period change rate range, and the modulation and coding scheme level; monitoring the current swing amplitude and swing period in real time and calculating the swing period change rate; querying the mapping table using the current swing amplitude and swing period change rate as indexes to obtain the target modulation and coding scheme level and performing the switch; simultaneously adjusting the subcarrier power allocation ratio so that the power allocated to the subcarriers carrying low-confidence tag data segments is higher than the power allocated to the subcarriers carrying high-confidence tag data segments, thereby providing additional energy protection for error-prone data segments and improving the overall reliability of the transmission link.

[0013] As a supplementary embodiment of the present invention, after performing differentiated forward error correction coding, a closed-loop verification step is further included: the receiving end decodes the calibrated visual data stream to recover the original visual data packets and the corresponding dynamic confidence interval information; the bit error rate of the low-confidence label data segment is calculated based on the dynamic confidence interval information; the bit error rate is compared with a preset bit error rate target value to generate an error signal; the error signal is transmitted back to the transmitting end through the reverse control channel to correct the radius of the dynamic confidence interval constructed in the next round of multipath delay spectrum analysis. This closed-loop correction mechanism enables the construction of the confidence interval to be adaptively optimized according to the actual decoding bit error situation, so that the entire calibration process can maintain the optimal state continuously under dynamic changes in hoisting and strong interference environments.

[0014] Through the coordinated efforts of the above-mentioned steps, this method achieves integrated closed-loop calibration of precise synchronous acquisition of visual data for prefabricated steel structure hoisting under swaying conditions, channel status perception, and differentiated protection transmission. It effectively overcomes the problems of image quality degradation and transmission reliability deterioration caused by large swaying during hoisting and complex multipath channels. While ensuring the real-time wireless transmission of visual data, it significantly improves the clarity of reconstructed images and the anti-interference capability of data links.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: The swing period and amplitude are obtained by analyzing the hoisting motion. The swing period is used as the sampling time window length, and the swing amplitude is used as the phase offset base. A phase-locked loop (PLL) circuit generates a reference clock with the same frequency as the swing period, and generates a pulse at the corresponding phase point within each period as a synchronization trigger signal for the visual data acquisition end. The synchronization trigger signal locks the exposure time of the visual sensor at a predetermined phase interval of the hoisting swing. When the swing amplitude is large, the exposure point automatically shifts to the phase with a lower swing speed, directly reducing the degree of motion blur within the frame and improving the clarity of image features and pose extraction accuracy of the hoisting node. The proportional-integral parameters of the PLL circuit are adaptively adjusted according to the rate of change of the swing period, keeping the tracking error between the reference clock and the swing phase within a very small range, avoiding phase loss due to sudden changes in wind conditions, and ensuring the long-term stability of the synchronization trigger time.

[0016] Multipath delay spectrum analysis is performed on the raw visual data packets to extract the arrival times and power attenuation values ​​of the primary and secondary path components. A dynamic confidence interval is constructed with the arrival time of the primary path as the center and the difference between the arrival times of the primary and secondary paths as the radius. The raw visual data packets are divided into data segments by bytes. Based on whether the transmission delay of each data segment falls within the dynamic confidence interval, a high-confidence label or a low-confidence label is applied. A lower redundancy coding rate is applied to the data segments carrying high-confidence labels, and a higher redundancy coding rate is applied to the data segments carrying low-confidence labels. Differentiated verification redundancy data is generated and interleaved to form the calibrated visual data stream. The dynamic confidence interval reflects the delay distribution of the multipath channel in real time. The confidence label is directly related to the physical characteristics of the channel, which concentrates error correction protection resources on data segments with large fading depth. Under the same coding overhead, the residual bit error rate of low-confidence data segments is significantly reduced, improving the integrity of the visual data recovered by the receiver. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0018] Figure 1 This is a flowchart of a method for wireless transmission calibration of visual data for hoisting prefabricated steel structures; Figure 2 This is a flowchart of the calculation of the swing trajectory and extraction of swing features of a hoisted object based on an inertial measurement unit; Figure 3 It is a timing flowchart for synchronously triggering signals to control the exposure of the vision sensor, analog-to-digital conversion, and data packaging; Figure 4 This is a schematic diagram of the displacement and swing trajectory of the hoisted object and its zero-crossing point; Figure 5 This is a schematic diagram of the swing period phase-locked loop reference clock and synchronous trigger signal; Figure 6 It is a curve showing the change in the swing period and swing amplitude of the hoisting action. Detailed Implementation

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

[0020] See Figure 1 This invention provides a method for wireless transmission calibration of visual data for prefabricated steel structure hoisting, comprising: Analyze the real-time hoisting motions during the hoisting process of prefabricated steel structures to obtain the swing period and swing amplitude of the hoisting motions; Based on the swing period and swing amplitude, the sampling clock of the visual data acquisition terminal is synchronized with the swing phase to generate a synchronization trigger signal. Based on the aforementioned synchronous trigger signal, the original visual data packets of the hoisting node are acquired; Multipath delay spectrum analysis is performed on the original visual data packets to extract the signal feature parameters of each transmission path; Based on the signal feature parameters, a dynamic confidence interval is constructed, and the original visual data packet is segmented and labeled with confidence level according to the dynamic confidence interval; Using the results of the segmented confidence marking, differential forward error correction coding is performed on the original visual data packet to generate a calibrated visual data stream.

[0021] Example 1: In specific implementation, please refer to Figure 2 The system acquires triaxial angular velocity and triaxial acceleration data from an inertial measurement unit (IMU) mounted at the base of the lifting rope. The IMU employs a six-axis MEMS sensor and outputs the triaxial angular velocity and acceleration data at a sampling frequency of at least 200Hz. An optional sampling frequency of 200Hz means the IMU outputs a set of triaxial angular velocity data and a set of triaxial acceleration data every 5 milliseconds. The triaxial angular velocity data includes angular velocity values ​​around the X, Y, and Z axes of the sensor coordinate system, and the triaxial acceleration data includes acceleration values ​​along the X, Y, and Z axes of the sensor coordinate system. The IMU is rigidly fixed to the base of the lifting rope near the hook, ensuring that the IMU's coordinate system is rigidly correlated with the motion attitude of the lifted object.

[0022] Time-series integration of triaxial angular velocity and triaxial acceleration data yields the swing trajectory curve of the hoisted object in three-dimensional space. In practice, the triaxial angular velocity data is integrated, and the attitude quaternion of the hoisted object is calculated based on the time-accumulated angular velocity information. Using the attitude quaternion at each moment, the corresponding triaxial acceleration data is rotated from the sensor coordinate system to the world coordinate system. The gravitational acceleration vector is subtracted from the three axes in the world coordinate system to obtain the components of the hoisted object's motion acceleration in the three axes. The motion acceleration components are numerically integrated twice: the first integration yields the velocity component, and the second integration yields the displacement component. The displacement components in the three axes, time-series, constitute the swing trajectory curve of the hoisted object in three-dimensional space. The numerical integration uses the composite trapezoidal rule, with the integration step size being the reciprocal of the sampling period of the inertial measurement unit.

[0023] The time interval between two adjacent zero-crossing points is extracted from the swing trajectory curve and taken as the swing period. The world coordinate system axis corresponding to the main swing direction of the hoisted object is selected, and the displacement time series of the swing trajectory curve on that axis is obtained. The displacement time series is traversed, and the instant when the displacement value changes from negative to positive or from positive to negative is recorded as the zero-crossing point. For two adjacent zero-crossing points with the same direction of sign change, the difference between the two zero-crossing points is calculated; this difference is one swing period. To suppress false zero-crossing point judgments caused by random noise, the displacement time series is low-pass filtered to remove components higher than a preset cutoff frequency, which is selected as 5 times the estimated swing frequency.

[0024] Calculate the peak offset of the oscillation trajectory curve within a single oscillation cycle, which is taken as the oscillation amplitude. For each defined oscillation cycle, extract the displacement segment of the oscillation trajectory curve along the corresponding coordinate axis within that cycle. Search for the maximum and minimum values ​​within the displacement segment, and calculate the oscillation amplitude using the following formula:

[0025] in, Indicates the amplitude of the swing; The function representing the displacement over time along the selected coordinate axis is determined by the projection of the oscillating trajectory curve onto that coordinate axis. Indicates within a single oscillation period The maximum value; This represents the minimum value within a single oscillation period; The value range is limited to the time interval of this single oscillation cycle. Each time a new oscillation cycle is obtained, the oscillation amplitude is calculated once, thus obtaining a oscillation amplitude sequence that is updated synchronously with the oscillation cycle.

[0026] See Figure 4 This figure illustrates the time-varying characteristics of the swing trajectory curve of the hoisted object acquired by the inertial measurement unit in Example 1. The horizontal axis represents time, ranging from 0 to 10 seconds; the vertical axis represents displacement, in millimeters. The original swing trajectory curve is represented by a black dashed line, which contains significant high-frequency noise and interference. The solid black line represents the swing trajectory curve after low-pass filtering of the original trajectory, effectively filtering out interference signals higher than the preset cutoff frequency, exhibiting smooth periodic swing characteristics. Hollow dots mark the zero-crossing points of the filtered swing trajectory curve; the zero-crossing point is the instantaneous time point when the displacement changes from positive to negative or vice versa.

[0027] As shown in the figure, the filtered swing trajectory curve exhibits a decreasing amplitude trend overall, reflecting the dynamic change characteristic of the swing amplitude of the hoisted object gradually decreasing over time. The zero-crossing points are evenly distributed with approximately constant intervals, indicating that the swing period is basically stable within this time window. This swing trajectory curve and its zero-crossing point extraction results provide accurate time series data for subsequent calculations of the swing period and amplitude, supporting the swing phase synchronization binding of the sampling clock and the generation of synchronization trigger signals, ensuring precise timing control of clock synchronization and data acquisition at the visual data acquisition end.

[0028] Example 2: In specific implementation, the oscillation period is used as the length parameter of the sampling time window, and the oscillation amplitude is used as the phase offset base of the sampling time window. The latest oscillation period value is extracted from the oscillation period sequence obtained in Example 1 and directly assigned to the length parameter of the sampling time window. The oscillation amplitude value corresponding to the latest oscillation period is extracted from the oscillation amplitude sequence obtained in Example 1 and used as the phase offset base to input into the phase offset calculation stage.

[0029] A phase-locked loop (PLL) circuit is used to track the real-time changes in the oscillation period, generating a reference clock with the same frequency as the oscillation period. The PLL circuit includes a phase detector, a loop filter, a voltage-controlled oscillator (VCO), and a frequency divider. One input of the phase detector receives the real-time oscillation period signal calculated by the inertial measurement unit (IMU), and the other input receives the signal fed back from the frequency divider. The phase detector outputs an error voltage proportional to the phase difference between the two input signals. This error voltage is fed into the loop filter, which employs a proportional-integral (PI) structure, comprising a proportional branch and an integral branch.

[0030] The proportional-integral (PI) parameters of the phase-locked loop (PLL) circuit are adaptively adjusted based on the rate of change of the oscillation period to keep the phase error between the reference clock and the oscillation period within ±2 degrees. In practice, the PLI parameters include a proportional gain and an integral gain. The time derivative of the oscillation period is calculated in real time to obtain the rate of change of the oscillation period. A preset threshold is used; when the rate of change of the oscillation period exceeds the threshold, the proportional gain is increased by multiplying it by an amplification factor; when the rate of change of the oscillation period falls below the threshold, the proportional gain is restored to the reference value. The reference value of the proportional gain is determined as follows: with the PLL circuit locked, the output jitter amplitude of the loop filter is measured, and the proportional gain value that makes the output jitter amplitude less than a preset jitter threshold is selected as the reference value. The preset jitter threshold is set to one-thousandth of the reference clock period. The integral gain is set to one-eighth of the reference value of the proportional gain. The amplification factor is set to 3 to ensure that the phase error is still limited to within ±2 degrees even under maximum oscillation acceleration. After the proportional gain and integral gain are determined, they are written into the proportional branch gain register and the integral branch gain register of the loop filter. The voltage-controlled oscillator (VCO) receives the control voltage after it has been processed by a loop filter and outputs an oscillation signal whose frequency varies with the control voltage. This oscillation signal is then divided by a frequency divider and fed back to a phase detector, forming a closed-loop lock-in. Once the phase-locked loop is locked, the oscillation signal output by the VCO becomes the reference clock, and the frequency of the reference clock is consistent with the real-time oscillation frequency.

[0031] Using the swing amplitude as the offset, a pulse is generated at the corresponding phase point within each cycle of the reference clock. In specific implementations, the swing amplitude is converted into a phase offset using the following formula:

[0032] in, This indicates the phase offset, in degrees. This indicates the amplitude of the swing corresponding to the current swing period; This indicates the maximum swing amplitude recorded during the hoisting process, which is pre-obtained through a full-range swing test during the system calibration phase. The digital phase accumulator uses a reference clock as its counting clock, accumulating from zero at the beginning of each reference clock cycle. When the accumulated value reaches a certain value... When the corresponding count value is reached, a pulse generator is triggered to produce a high-level pulse with a width of one-fiftieth of the reference clock cycle.

[0033] The pulse, after shaping and amplification, is used as a synchronization trigger signal and output to the shutter control pin of the visual data acquisition terminal. In specific implementation, the pulse shaping process includes passing the pulse generated by the pulse generator through a Schmitt trigger to eliminate edge jitter, and then adjusting the pulse voltage amplitude to the voltage range required by the shutter control pin of the visual data acquisition terminal through a level conversion circuit. The level-converted pulse signal is then amplified by a driver amplifier to enhance its driving capability. The output of the driver amplifier is connected to the shutter control pin of the visual data acquisition terminal through a coaxial shielded cable. The pulse signal output by the driver amplifier is the synchronization trigger signal.

[0034] See Figure 5 In the figure, the horizontal axis represents time in seconds, and the vertical axis represents amplitude normalization. The dashed curve is the swing phase reference signal, which is a periodic sine wave reflecting the periodic characteristics of the hoisted object's swing. The swing period of the dashed curve is approximately 1 second, consistent with the latest swing period value calculated by the inertial measurement unit in Example 2, used to determine the sampling time window length. The black dashed square wave curve is the phase-locked loop (PLL) reference clock, whose high-level duration is synchronized with the swing period, and its phase closely tracks the swing phase reference signal, indicating that the PLL circuit successfully locked and tracked the swing period, achieving synchronous synchronization between the reference clock and the real-time swing period. The phase error is controlled within ±2 degrees to ensure synchronization accuracy. The black solid dots in the figure represent the synchronization trigger signal. The trigger signal is generated at the end of the high level of the PLL reference clock, corresponding to a specific phase position of the swing phase reference signal, consistent with the mechanism in Example 2 where the synchronization signal is triggered by the phase offset of the swing amplitude conversion. The period of the synchronization trigger signal is consistent with the swing period, and the trigger time is dynamically adjusted to ensure that the shutter control of the visual data acquisition end is synchronously bound with the swing phase of the hoisted object, thereby achieving accurate sampling of visual data. The overall curve trend and key nodes show the specific implementation process of Example 2, which generates a reference clock through a phase-locked loop circuit, calculates the phase offset by combining the swing amplitude, and finally outputs a synchronous trigger signal for frame exposure start control at the visual data acquisition end.

[0035] Example 3: In specific implementation, please refer to Figure 3The synchronization trigger signal is used as the frame exposure start signal of the vision sensor, controlling the vision sensor to perform pixel charge accumulation during the high level of the synchronization trigger signal. The vision sensor is a global shutter image sensor, and the frame exposure start pin of the global shutter image sensor receives the synchronization trigger signal generated in Example 2. When the synchronization trigger signal transitions from low to high, the row and column gating circuits inside the vision sensor simultaneously turn on the transmission gates of all pixel units. The photodiodes of all pixel units begin to convert incident photons into charge, and the charge accumulates on the storage capacitors inside the pixel units. The duration of the high level of the synchronization trigger signal determines the duration of pixel charge accumulation. The high level duration is set to one-tenth of the synchronization trigger signal period, so that pixel charge accumulation is completed within a specific phase interval of the swing period. When the synchronization trigger signal transitions from high to low, the transmission gates of all pixel units simultaneously turn off, pixel charge accumulation stops, and the accumulated charge is latched in the storage capacitors of each pixel unit.

[0036] The falling edge of the synchronization trigger signal triggers the analog-to-digital converter (ADC) to convert accumulated pixel charge into digital image data. In specific implementations, the vision sensor integrates a column-level ADC array, whose conversion start pin is connected to the falling edge detection circuit of the synchronization trigger signal. The falling edge detection circuit generates a start pulse at the falling edge of the synchronization trigger signal, which simultaneously triggers all column-level ADCs to begin analog-to-digital conversion. The column-level ADCs read the analog charge voltage on the storage capacitor of each pixel unit in the corresponding column of the pixel array row by row, quantizing the analog charge voltage into digital grayscale values. The digital grayscale value has a bit width of 12 bits and a quantization range of 0 to 4095. The digital grayscale values ​​of all pixel units in a frame of image data constitute a frame of digital image data. The data output buffer inside the vision sensor temporarily stores the complete digital image data frame, waiting for external reading.

[0037] Multiple frames of digital image data acquired within consecutive synchronization trigger signal cycles are packaged into a single raw visual data packet in chronological order. Optionally, the number of consecutive synchronization trigger signal cycles is set to 6, meaning one raw visual data packet contains 6 frames of digital image data. In specific implementations, a data packet controller is configured. The data packet controller uses the raw visual data packet as the basic transmission unit to perform packet assembly operations on the digital image data frames. The data packet controller internally maintains a frame counter, which increments by 1 each time a frame of digital image data is received. When the frame counter reaches 6, the data packet controller resets the frame counter to zero and simultaneously concatenates the currently buffered 6 frames of digital image data in chronological order of acquisition time to form the payload of a raw visual data packet. The 6-frame setting is based on the fact that at a 200Hz synchronization trigger signal frequency, 6 frames correspond to a 30-millisecond acquisition time, which is less than one-fifth of the wireless channel coherence time, ensuring that the channel response of each frame within the raw visual data packet remains approximately unchanged.

[0038] The swing period value and swing amplitude value corresponding to the current acquisition cycle are embedded in the header of each raw visual data packet. In specific implementation, a data packet header structure is defined, with a length of 16 bytes. The first 2 bytes of the header structure are the synchronization word, which is fixed as the hexadecimal value 0xEB90; the next 2 bytes are the frame number field, filled with the frame number 6 of the digital image data contained in the current raw visual data packet; the next 2 bytes are the swing period value field corresponding to the acquisition cycle, with the swing period value in milliseconds and represented by a 16-bit unsigned integer; the next 2 bytes are the swing amplitude value field corresponding to the acquisition cycle, with the swing amplitude value in millimeters and represented by a 16-bit unsigned integer; the remaining 8 bytes of the header structure are reserved. During the packet assembly process, the data packaging controller reads the latest swing period value and the latest swing amplitude value corresponding to the current acquisition cycle from the swing period sequence and swing amplitude sequence obtained in Example 1, fills in the corresponding fields mentioned above, and then assembles the complete raw visual data packet.

[0039] See Figure 6 The figure uses the original visual data packet sequence number as the x-axis, the oscillation period value (in milliseconds) as the left y-axis, and the oscillation amplitude value (in millimeters) as the right y-axis, plotting the oscillation period value curve (dashed line) and the oscillation amplitude value curve (solid line). The x-axis range covers 600 consecutive original visual data packet sequences, the oscillation period value on the y-axis fluctuates between 1850 ms and 2150 ms, and the oscillation amplitude value varies between 0 mm and approximately 450 mm.

[0040] The curves show that the oscillation period value exhibits a periodic fluctuation trend with relatively small amplitude, generally fluctuating between 1900 and 2100 milliseconds. This fluctuation corresponds to the periodic change in the oscillation of the hoisted object, reflecting the dynamic adjustment of the oscillation period in the real-time calculated oscillation trajectory. The oscillation amplitude curve shows significant fluctuations and a wide range of variation. In some intervals, the oscillation amplitude approaches zero, indicating that the oscillation amplitude of the hoisted object is small within certain acquisition cycles. In other intervals, the oscillation amplitude increases significantly, exceeding 400 mm, reaching the preset threshold for severe oscillation, reflecting changes in the intensity of the hoisted object's oscillation.

[0041] From data packet sequence number 0 to approximately 60, the swing amplitude increases rapidly, reaching a threshold of about 400 mm. Subsequently, within the entire sequence range, the swing amplitude curve exhibits similar sharp fluctuations multiple times, reflecting the periodic changes in the swing intensity of the hoisted object during the hoisting process. Correspondingly, although the swing period curve fluctuates, the amplitude is small, and there is a certain phase difference between it and the swing amplitude curve, indicating that the swing period and swing amplitude are two related but independently changing parameters.

[0042] The sway amplitude exceeding 400 mm in the figure corresponds to the severe sway threshold range defined in Examples 4 and 5. This threshold range triggers the system to switch to the most robust modulation and coding scheme level and adjust the cyclic prefix length. The sway amplitude is divided into multiple ranges: 0 to 50 mm, 50 to 100 mm, 100 to 200 mm, 200 to 400 mm, and exceeding 400 mm. The curve covers the entire range shown in the figure, indicating that the system can dynamically adapt to different sway intensities.

[0043] Example 4: In practice, a probe reference signal is broadcast to the wireless channel in the hoisting area, and channel impulse response sequences from multiple receiving antennas are received. The probe reference signal uses a Zadoff-Chu sequence and is broadcast within a guard interval before each raw visual data packet is sent. The transmitting end is equipped with a probe reference signal generation module, which internally stores pre-calculated real and imaginary sample values ​​of the Zadoff-Chu sequence. In the Zadoff-Chu sequence generation formula, the sequence length is 839, and the root sequence index is 3. The root sequence index is set to 3 because, in link budget simulation, the ratio of the cross-correlation peak value to the autocorrelation peak value corresponding to root sequence index 3 is lower than a preset cross-correlation threshold, which is set to 0.1. The guard interval is set to 20 microseconds. During the guard interval, the transmitting end does not send visual data, only the probe reference signal. The 20-microsecond duration of the guard interval is greater than the preset estimate of the maximum delay spread of the multipath channel in the hoisting area, which is 15 microseconds. The transmitting end is equipped with a single omnidirectional transmitting antenna, and the receiving end is equipped with four omnidirectional receiving antennas, which are arranged in a uniform linear array with a half-wavelength spacing. Each omnidirectional receiving antenna independently receives wireless signals and generates a channel impulse response sequence in its respective receiving link.

[0044] A sliding correlation operation is performed on the channel impulse response sequence to separate multipath components at different time delays. In specific implementation, on the receiving link corresponding to each omnidirectional receiving antenna at the receiving end, the received discrete baseband signal sequence is subjected to sliding correlation with the conjugate sequence of the locally stored Zadoff-Chu sequence. The step size of the sliding correlation is set to one sampling period, which is 10 nanoseconds corresponding to a baseband sampling rate of 100MHz. The sliding correlation operation outputs a sequence of correlation value amplitudes varying with the delayed sampling points, and each peak in the correlation value amplitude sequence corresponds to a multipath component. A peak detection threshold is set, which is determined according to the constant false alarm rate (CFAR) criterion, with the CFAR set to 0.001. The peak detection threshold is obtained by statistically sorting the correlation value amplitudes within the pure noise interval. The sampling point position where the correlation value amplitude exceeds the peak detection threshold is the arrival time of the multipath component, and the arrival time of the multipath component is represented by the ordinal number of the delayed sampling point; the negative half of the logarithm of the ratio of the correlation value amplitude at this sampling point position to the known power of the transmitted detection reference signal is used as the power attenuation value of the multipath component.

[0045] The arrival time and power attenuation value of each multipath component are recorded as signal characteristic parameters. In specific implementation, after performing sliding correlation and peak detection on the four omnidirectional receiving antennas, all multipath components detected on the four base station omnidirectional receiving antennas are aggregated into a multipath component list. Each multipath component entry in the multipath component list contains three information cells: arrival time information cell, storing the ordinal number of the delayed sampling point; power attenuation value information cell, storing the power attenuation value in decibels; and antenna index information cell, storing the number of the omnidirectional receiving antenna that detected the multipath component. Entries in the multipath component list with the same ordinal number of delayed sampling points but different antenna index information cells are merged. The power attenuation value of the merged multipath component is taken as the arithmetic mean of the power attenuation values ​​of each antenna branch, and the arrival time of the merged multipath component is taken as the ordinal number of the delayed sampling point. The merged multipath component list is output as a set of signal characteristic parameters.

[0046] Multipath components are sorted according to their arrival times to generate a multipath delay spectrum, and the primary path component with the smallest power attenuation and the secondary path component with the largest power attenuation are marked. In practice, the multipath component entries in the multipath component list are sorted in ascending order of their arrival time cell values. The sorted multipath component sequence constitutes the multipath delay spectrum. In the multipath delay spectrum, the multipath component with the smallest arrival time cell is the first arriving multipath component, and the multipath component with the largest arrival time cell is the last arriving multipath component. The sorted multipath component sequence is traversed, and the power attenuation cell values ​​of each multipath component are compared. The multipath component with the smallest power attenuation cell is identified and marked as the primary path component; the multipath component with the largest power attenuation cell is identified and marked as the secondary path component. If the maximum power attenuation cell value corresponds to multiple multipath components, the multipath component with the largest arrival time cell is selected and marked as the secondary path component.

[0047] A dynamic confidence interval is constructed using the arrival time of the primary path component as the center point and the difference between the arrival times of the secondary path components and the primary path component as the interval radius. In practice, the arrival time information cell value of the primary path component is read from the multipath delay spectrum and denoted as... Read the arrival time cell value of the secondary path component and record it as... The center point of the dynamic confidence interval. Set as The radius of the dynamic confidence interval Calculated using the following formula:

[0048] in, Represents the radius of the dynamic confidence interval, in units of the number of delayed sampling points; This indicates the arrival time of the secondary path component, expressed in units of the number of delayed sampling points. This represents the arrival time of the principal diameter component, expressed in units of the number of delayed sampling points. The lower bound of the dynamic confidence interval on the time axis is [value], and the upper bound is [value]. .

[0049] The original visual data packet is divided into multiple data segments by bytes. In specific implementation, the total number of bytes in the original visual data packet is divided by the fixed byte length of each data segment to obtain the total number of data segments, and the fixed byte length of each data segment is set to 64 bytes. If the total number of bytes in the original visual data packet is not divisible by 64, the last data segment has fewer than 64 bytes and is still treated as an independent data segment. Each data segment is assigned a segment number, starting from 1 and incrementing sequentially according to the byte order within the original visual data packet.

[0050] The process involves determining whether the transmission delay of each data segment falls within the dynamic confidence interval. In practice, the transmission delay for each data segment is obtained by estimating the difference between the transmission and reception times during wireless transmission. The receiver uses a clock synchronized with the transmitter to record the reception time when each data segment is received. It also reads the embedded transmission time from the frame structure of the data segment and subtracts the transmission time from the reception time to obtain the delay for that data segment. The delay for each data segment is then compared with the lower and upper bounds of the dynamic confidence interval.

[0051] If the latency corresponding to a data segment falls within the dynamic confidence interval (i.e., the latency is greater than or equal to the lower bound and less than or equal to the upper bound), the data segment is labeled with a high-confidence tag. If the latency corresponding to a data segment does not fall within the dynamic confidence interval (i.e., the latency is less than the lower bound or greater than the upper bound), the data segment is labeled with a low-confidence tag. High-confidence tags are encoded using binary values ​​of 1, and low-confidence tags are encoded using binary values ​​of 0. The encoded tags are arranged in order of data segment sequence to form a confidence tag bitmap. This confidence tag bitmap is stored in association with the original visual data packet for subsequent differential forward error correction coding.

[0052] Example 5: In specific implementation, all data segments carrying high-confidence labels are extracted from the original visual data packet to form a first data set. The confidence marker bitmap generated in Example 4 is obtained. The confidence marker bitmap is a binary sequence, where each bit corresponds to a data segment divided into bytes in the original visual data packet. Each bit of the confidence marker bitmap is traversed; when a bit's value is binary 1, the data segment corresponding to that bit in the original visual data packet is extracted and placed into the first data set. The data segments within the first data set are arranged in ascending order of their data segment numbers.

[0053] Extract all data segments carrying low-confidence labels from the original visual data packet to form a second data set. Continue iterating through each bit of the confidence label bitmap; when a bit has a value of binary 0, extract the data segment corresponding to that bit number from the original visual data packet and add it to the second data set. The data segments within the second data set are also arranged in ascending order of their data segment numbers.

[0054] A first redundancy coding rate is applied to the first data set to generate first parity redundant data. A second redundancy coding rate is applied to the second data set to generate second parity redundant data, wherein the second redundancy coding rate is higher than the first redundancy coding rate. In specific implementation, the forward error correction coding uses a low-density parity-check code. The first redundancy coding rate is set to 3 / 4, which means that 1 parity bit is generated for every 3 information bits, corresponding to a low-density parity-check code parity-check matrix with a row weight of 6 and a column weight of 4. The second redundancy coding rate is set to 1 / 2, which means that 1 parity bit is generated for every 1 information bit, corresponding to a low-density parity-check code parity-check matrix with a row weight of 8 and a column weight of 4. The reason for setting the second redundancy coding rate of 1 / 2 higher than the first redundancy coding rate of 3 / 4 is that data segments carrying low-confidence tags experience worse channel conditions during wireless transmission, requiring stronger error correction protection capabilities to reduce the bit error rate at the receiver. Selecting a low-density parity check code with a code rate of 1 / 2 can improve the bit error rate performance by about two orders of magnitude compared to a code rate of 3 / 4, maintaining the consistency of the bit error rate after the two data streams are recovered. The first data set is fed into a low-density parity check code encoder with a code rate parameter configured to 3 / 4, outputting the first parity redundancy data. The second data set is fed into a low-density parity check code encoder with a code rate parameter configured to 1 / 2, outputting the second parity redundancy data.

[0055] The first data set, the second data set, the first check redundancy data, and the second check redundancy data are interleaved and arranged in the order of the original data segments to form the calibrated visual data stream. In specific implementation, an output data buffer is established, and the capacity of the output data buffer is greater than the sum of the number of bytes in the first data set, the second data set, the first check redundancy data, and the second check redundancy data. The total number of data segments in the original visual data packet is obtained and denoted as . Create an interleaved write pointer, with a step size equal to the length of one data segment (in bytes). Starting from index 1, traverse the data segments sequentially. While traversing each data segment index, determine the set to which the corresponding data segment belongs: if the index bit in the confidence flag bitmap is binary 1, the data segment belongs to the first data set. Take a data segment from the first data set in sequence and write it to the current position in the output data buffer, then move the interleaved write pointer forward by one data segment byte. If the index bit in the confidence flag bitmap is binary 0, the data segment belongs to the second data set. Take a data segment from the second data set in sequence and write it to the current position in the output data buffer, then move the interleaved write pointer forward by one data segment byte. After traversing all data segment indices, all data segments in the first and second data sets are arranged in their original order. Write the first parity redundancy data sequentially to the output data buffer, and move the interleaved write pointer accordingly. Write the second parity redundancy data sequentially to the output data buffer. The final, complete data arranged in the output data buffer is the calibrated visual data stream.

[0056] After generating the calibrated visual data stream, it is fed into an orthogonal frequency division multiplexing modulator and mapped onto multiple subcarriers. In this specific implementation, the orthogonal frequency division multiplexing modulator is configured with a total of 256 subcarriers, including 192 effective subcarriers, 8 pilot subcarriers, and 56 unused subcarriers. After serial-to-parallel conversion, the calibrated visual data stream is distributed to the 192 effective subcarriers in ascending order of subcarrier number, with each effective subcarrier carrying a certain number of bits.

[0057] Based on the swing period and swing amplitude, the modulation and coding scheme levels of multiple subcarriers are dynamically adjusted. A mapping table is established, containing a ternary correspondence between swing amplitude ranges, swing period change rate ranges, and modulation and coding scheme levels. The mapping table is structured as a two-dimensional lookup table, where the row index is the swing amplitude range number, the column index is the swing period change rate range number, and the content of the lookup table cell is the modulation and coding scheme level index. In specific implementation, the swing amplitude range is divided into 5 ranges: range 0 is a swing amplitude less than 50 mm, range 1 is a swing amplitude greater than or equal to 50 mm and less than 100 mm, range 2 is a swing amplitude greater than or equal to 100 mm and less than 200 mm, range 3 is a swing amplitude greater than or equal to 200 mm and less than 400 mm, and range 4 is a swing amplitude greater than or equal to 400 mm. The swing period change rate range is divided into 3 ranges: range 0 is a swing period change rate with an absolute value less than 5 milliseconds per second, range 1 is a swing period change rate with an absolute value greater than or equal to 5 milliseconds per second and less than 20 milliseconds per second, and range 2 is a swing period change rate with an absolute value greater than or equal to 20 milliseconds per second. The modulation and coding scheme level index ranges from 0 to 9. A modulation and coding scheme level index of 0 indicates binary phase shift keying with a 1 / 2 rate low-density parity check code, while a modulation and coding scheme level index of 9 indicates 64 quadrature amplitude modulation with a 3 / 4 rate low-density parity check code.

[0058] The system monitors the current oscillation amplitude and oscillation period in real time and calculates the rate of change of the oscillation period. In a specific implementation, the oscillation period values ​​at two consecutive moments are obtained from the oscillation period sequence output in Example 1. The difference between the oscillation period value at the later moment and the oscillation period value at the earlier moment is divided by the time interval between the two moments to obtain the rate of change of the oscillation period. The time interval between the two moments is taken as the update interval of the oscillation period sequence, which is 5 milliseconds.

[0059] Using the current swing amplitude and swing period change rate as indexes, the target modulation and coding scheme level is obtained by querying the mapping table. The current swing amplitude value is compared with the upper and lower bounds of the swing amplitude interval to determine the swing amplitude interval into which the current swing amplitude falls, and the swing amplitude interval number is recorded. The absolute value of the calculated swing period change rate is compared with the upper and lower bounds of the swing period change rate interval to determine the swing period change rate interval into which the swing period change rate falls, and the swing period change rate interval number is recorded. The target modulation and coding scheme level index of the corresponding unit is retrieved by looking up the swing amplitude interval number and the swing period change rate interval number in the mapping table.

[0060] The current modulation and coding scheme level is adjusted to the target modulation and coding scheme level, and the subcarrier power allocation ratio is adjusted simultaneously so that the power allocated to subcarriers carrying low-confidence tag data segments is higher than the power allocated to subcarriers carrying high-confidence tag data segments. The subcarrier power allocation ratio is adjusted as follows: For all 192 valid subcarriers, the confidence tag corresponding to the data segment carried by each subcarrier is obtained. If the confidence tag is a high-confidence tag, the transmit power allocated to that subcarrier is the base power; if the confidence tag is a low-confidence tag, the transmit power allocated to that subcarrier is the base power multiplied by a power enhancement factor. The power enhancement factor is set to 2. The reason for setting the power enhancement factor to 2 is that, under the total transmit power constraint, the additional 3dB power allocated to the low-confidence tag subcarrier can improve the received signal-to-noise ratio by 3dB, compensating for the power backoff margin caused by the high redundant coding rate in the link budget. The total transmit power of all subcarriers does not exceed the maximum linear output power of the transmitter's RF front-end.

[0061] When the swing amplitude exceeds a preset severe swing threshold, the modulation and coding scheme level is switched to the most robust level, and the cyclic prefix length of the orthogonal frequency division multiplexing modulator is reduced. The preset severe swing threshold is set to 400 mm, which is consistent with the lower limit of swing amplitude interval 4 in the mapping table. When the current swing amplitude exceeds 400 mm, regardless of the current swing period change rate, the modulation and coding scheme level index is directly set to 0. Modulation and coding scheme level index 0 corresponds to binary phase shift keying modulation plus 1 / 2 code rate low-density parity check code, which is the most robust modulation and coding scheme level in the mapping table. Meanwhile, the cyclic prefix length of the orthogonal frequency division multiplexing modulator is reduced from 64 sampling points in the conventional mode to 16 sampling points. The basis for reducing the cyclic prefix length is that under severe swing conditions, the multipath delay spread is concentrated in a small range due to the regular swing of the hoisting object. The measured value of the multipath delay spread does not exceed 10 times the baseband sampling period. The 16 sampling point cyclic prefix is ​​sufficient to eliminate inter-symbol interference. After the cyclic prefix length is reduced, the period of a single orthogonal frequency division multiplexing symbol is shortened, and the time-domain symbol rate is increased without changing the subcarrier spacing to counteract the rapid time-varying channel.

[0062] The receiving end decodes the calibrated visual data stream to recover the original visual data packets and corresponding dynamic confidence interval information. In specific implementations, the receiving end sequentially performs orthogonal frequency division multiplexing demodulation, deinterleaving, and low-density parity check (LDPC) decoding on the calibrated visual data stream. The LPC decoder iteratively decodes the code blocks corresponding to the first and second data sets according to the parameter configuration of a first redundancy coding rate of 3 / 4 and a second redundancy coding rate of 1 / 2, respectively, outputting the recovered data segment set. Following the reverse process of the interleaving arrangement, the recovered data segment set is reassembled into the original visual data packets based on the confidence marker bitmap synchronized locally at the receiving end. The dynamic confidence interval information includes the center point. and radius The dynamic confidence interval information is obtained independently by the receiver from the received multipath delay spectrum using the same calculation method as the transmitter.

[0063] The receiving end calculates the bit error rate (BER) of the low-confidence tag data segment based on the dynamic confidence interval information. In practice, the receiving end records the number of error bits for each code block during the low-density parity-check code decoding process. The total number of code blocks and the total number of error bits corresponding to the low-confidence tag data segment are accumulated in the BER statistics module. After receiving the decoding results of all code blocks of a complete original visual data packet, the BER statistics module calculates the BER of the low-confidence tag data segment using the formula: the total number of error bits after decoding the low-confidence tag data segment divided by the total number of bits after decoding the low-confidence tag data segment.

[0064] The bit error rate is compared with a preset target value to generate an error signal. The preset target value is... The preset target bit error rate The selection criterion is the maximum acceptable bit error rate of visual data that meets the image quality requirements for monitoring the hoisting of prefabricated steel structures. The error signal is calculated using the following formula:

[0065] in, This represents the error signal and is dimensionless. Represents the measured bit error rate of low-confidence labeled data segments, dimensionless; This represents the preset target bit error rate value, which takes the value of Dimensionless. When A positive value indicates that the measured bit error rate is higher than the target, and the radius of the dynamic confidence interval needs to be expanded; when A negative value indicates that the measured bit error rate is lower than the target, which can reduce the radius of the dynamic confidence interval.

[0066] The error signal is transmitted back to the transmitter via a reverse control channel to correct the radius of the dynamic confidence interval constructed in the next round of multipath delay spectrum analysis. In practice, an independent wireless reverse control channel is established between the transmitter and receiver, employing binary frequency shift keying modulation. The receiver then receives the calculated error signal. The quantization encoding is converted into an 8-bit control command word, which is transmitted back once every original visual data packet transmission cycle via the reverse control channel. Upon receiving the control command word, the transmitting end decodes it to recover the error signal. The current dynamic confidence interval radius Correct according to the following rules: Adjust the error signal... Multiplying by a correction step size factor, where the correction step size factor is 1 (the number of delayed sampling points), yields the radius correction increment. This radius correction increment is then multiplied by the current dynamic confidence interval radius. The sums are used to obtain the dynamic confidence interval radius for the next round. The correction step size coefficient is set to 1, which is the number of delayed sampling points, so that the radius adjustment is performed in units of the smallest granularity of delayed sampling points, ensuring that the precision of the confidence interval adjustment matches the time resolution of the multipath delay spectrum.

[0067] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for wireless transmission and calibration of visual data for hoisting prefabricated steel structures, characterized in that, The method includes: Analyze the real-time hoisting motions during the hoisting process of prefabricated steel structures to obtain the swing period and swing amplitude of the hoisting motions; Based on the swing period and swing amplitude, the sampling clock of the visual data acquisition terminal is synchronized with the swing phase to generate a synchronization trigger signal. Based on the aforementioned synchronous trigger signal, the original visual data packets of the hoisting node are acquired; Multipath delay spectrum analysis is performed on the original visual data packets to extract the signal feature parameters of each transmission path; Based on the signal feature parameters, a dynamic confidence interval is constructed, and the original visual data packet is segmented and labeled with confidence level according to the dynamic confidence interval; Using the results of the segmented confidence marking, differential forward error correction coding is performed on the original visual data packet to generate a calibrated visual data stream.

2. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 1, characterized in that, The steps for analyzing the real-time hoisting motion of the prefabricated steel structure hoisting process and obtaining the swing period and swing amplitude of the hoisting motion specifically include: Acquire triaxial angular velocity and triaxial acceleration data output by the inertial measurement unit installed at the root of the hoisting rope; By integrating the triaxial angular velocity and triaxial acceleration data over time, the swing trajectory curve of the hoisted object in three-dimensional space is obtained; The time interval between two adjacent zero-crossing points is extracted from the swing trajectory curve and used as the swing period. The peak offset of the swing trajectory curve within a single swing cycle is calculated as the swing amplitude.

3. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 2, characterized in that, The inertial measurement unit uses a six-axis MEMS sensor and outputs triaxial angular velocity and triaxial acceleration data at a sampling frequency of not less than 200Hz.

4. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 1, characterized in that, The steps for generating a synchronization trigger signal by synchronizing the sampling clock of the visual data acquisition terminal with the swing phase based on the swing period and swing amplitude specifically include: The oscillation period is used as the length parameter of the sampling time window, and the oscillation amplitude is used as the phase offset base of the sampling time window; A phase-locked loop circuit is used to track the real-time changes in the oscillation period and generate a reference clock with the same frequency as the oscillation period. Using the swing amplitude as an offset, a pulse is generated at the corresponding phase point within each cycle of the reference clock; The pulse is shaped and amplified, and then used as the synchronization trigger signal, which is output to the shutter control pin of the visual data acquisition terminal.

5. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 4, characterized in that, The proportional-integral parameters of the phase-locked loop circuit are adaptively adjusted according to the rate of change of the oscillation period, so that the phase error between the reference clock and the oscillation period is kept within ±2 degrees.

6. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 1, characterized in that, The steps for acquiring the original visual data packets of the hoisting node based on the aforementioned synchronization trigger signal specifically include: The synchronization trigger signal is used as the frame exposure start signal of the visual sensor, and the visual sensor is controlled to perform pixel charge accumulation during the high level of the synchronization trigger signal; The falling edge of the synchronization trigger signal triggers the analog-to-digital converter, which converts the accumulated pixel charge into digital image data. Multiple frames of digital image data acquired within several consecutive synchronous trigger signal cycles are packaged into a single raw visual data packet in chronological order. The swing period value and swing amplitude value corresponding to the current acquisition cycle are embedded in the header of each raw visual data packet.

7. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 1, characterized in that, The steps of performing multipath delay spectrum analysis on the original visual data packets and extracting signal feature parameters for each transmission path specifically include: Broadcast a probe reference signal to the wireless channel in the hoisting area and receive channel impulse response sequences from multiple receiving antennas; Perform sliding correlation operation on the channel impulse response sequence to separate multipath components at different time delays; The arrival time and power attenuation value of each multipath component are recorded as signal characteristic parameters; The multipath components are sorted according to their arrival times to generate a multipath time delay spectrum, and the primary path component with the smallest power attenuation value and the secondary path component with the largest power attenuation value are marked.

8. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 7, characterized in that, The detection reference signal uses the Zadoff-Chu sequence and is broadcast during a guard interval before each raw visual data packet is sent.

9. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 7, characterized in that, The steps of constructing dynamic confidence intervals based on the signal feature parameters and segmenting the original visual data packets according to the dynamic confidence intervals specifically include: The dynamic confidence interval is constructed with the arrival time of the primary diameter component as the center point and the difference between the arrival time of the secondary diameter component and the arrival time of the primary diameter component as the interval radius. The original visual data packet is divided into multiple data segments by bytes; Determine whether the time delay corresponding to each data segment during transmission falls within the dynamic confidence interval; If the delay corresponding to the data segment falls within the dynamic confidence interval, the data segment is labeled with a high confidence level. If the delay corresponding to the data segment does not fall within the dynamic confidence interval, the data segment is labeled with a low confidence level.

10. The method for wireless transmission and calibration of visual data for prefabricated steel structure hoisting according to claim 9, characterized in that, The steps of performing differential forward error correction coding on the original visual data packet using the results of the segmented confidence marking to generate a calibrated visual data stream specifically include: Extract all data segments carrying high-confidence labels from the original visual data packet to form a first data set; Extract all data segments carrying low-confidence labels from the original visual data packet to form a second data set; Apply a first redundancy coding rate to the first data set to generate first check redundancy data; A second redundancy coding rate is applied to the second data set to generate second check redundancy data, wherein the second redundancy coding rate is higher than the first redundancy coding rate; The first data set, the second data set, the first check redundancy data, and the second check redundancy data are interleaved and arranged in the order of the original data segments to form the calibrated visual data stream.