An intelligent shooting control method and system for a fan blade in a non-stop state

Abstract

The present application relates to the technical field of image shooting control, in particular to an intelligent shooting control method and system for fan blades in non-stop state, comprising obtaining the geometric characteristics of the blades of a target fan, dividing the blades into multiple speed characteristic intervals along the radial direction, and establishing the corresponding relationship between the speed characteristic intervals and the radial positions; the rotor speed and the blade angle position are monitored in real time through a sensor, the real-time linear speed of the speed characteristic interval is calculated based on the rotor speed and the radial position, the optimal exposure time corresponding to the speed characteristic interval is calculated, and the aperture value and the sensitivity parameter corresponding to the speed characteristic interval are determined in combination with the current lighting conditions and the imaging quality; based on the rotor speed and the blade angle position, the time for the target blade speed characteristic interval to reach a preset shooting position is predicted, a shooting control signal is generated, the shooting parameters are switched when the target blade reaches the shooting position, the speed characteristic interval image is obtained, and the shooting parameter range is adjusted when the image quality is lower than a preset threshold.

Description

Technical Field

[0001] This invention relates to the field of image capture control technology, specifically to an intelligent capture control method and system for wind turbine blades without stopping operation. Background Technology

[0002] As a crucial component of clean energy, the safe and reliable operation of wind power equipment is of paramount importance. Wind turbine blades, as the core components of wind turbine generators, are subjected to complex alternating loads and harsh environmental conditions over long periods, making them prone to defects such as surface wear, cracks, and lightning strike damage. If these defects are not detected and addressed promptly, they can lead to serious accidents such as blade breakage, causing significant economic losses and safety hazards.

[0003] However, existing non-stop inspection technologies mainly employ long-distance imaging, which faces numerous technical challenges. The most prominent issue is the significant difference in linear velocity between different parts of the blade during rotation: taking a typical 120-meter diameter impeller as an example, the linear velocity at the blade root is close to zero, while the linear velocity at the blade tip can reach over 80 meters per second, a difference of several times or even tens of times. This velocity difference leads to severely inconsistent imaging results for different parts of the blade when using uniform imaging parameters: the blade tip is prone to motion blur, while the blade root may appear dark due to insufficient exposure time.

[0004] Existing technologies typically employ fixed shooting parameters or simple segmented processing methods, which cannot effectively address the speed differences across the entire blade length, resulting in uneven image quality and affecting the accuracy and reliability of defect identification. Therefore, this paper proposes an intelligent shooting control method and system for wind turbine blades without shutting down the turbine. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent shooting control method and system for wind turbine blades without shutting down the machine. The method includes acquiring the geometric characteristics of the target wind turbine blades, dividing the blades radially into multiple velocity characteristic intervals, and establishing a correspondence between the velocity characteristic intervals and radial positions; monitoring the impeller speed and blade angle position in real time using sensors, calculating the real-time linear velocity of the velocity characteristic intervals based on the impeller speed and radial position, calculating the optimal exposure time corresponding to each velocity characteristic interval, and determining the aperture value and ISO parameters corresponding to each velocity characteristic interval based on current lighting conditions and image quality; predicting the time it takes for the target blade's velocity characteristic interval to reach a preset shooting position based on the impeller speed and blade angle position, generating a shooting control signal, switching shooting parameters when the target blade reaches the shooting position, acquiring images of the velocity characteristic intervals, and adjusting the shooting parameter range when the image quality is below a preset threshold.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for intelligent image capture control of wind turbine blades without shutting down the machine includes:

[0008] Obtain the blade geometry characteristics of the target wind turbine, divide the blade radially into multiple velocity characteristic intervals, each velocity characteristic interval corresponds to a different radial position, and establish the correspondence between the velocity characteristic intervals and the radial positions;

[0009] The impeller speed and blade angle position are monitored in real time by sensors, and the real-time linear velocity of the speed characteristic range is calculated based on the impeller speed and radial position.

[0010] Based on the real-time linear velocity, the optimal exposure time corresponding to each speed characteristic range is calculated, and the aperture value and ISO parameter corresponding to the speed characteristic range are determined in combination with the current lighting conditions and image quality.

[0011] Based on the impeller speed and blade angle position, predict the time when the target blade speed characteristic range reaches the preset shooting position, generate a shooting control signal, and switch the shooting parameters of the corresponding speed characteristic range when the target blade speed characteristic range reaches the shooting position in sequence to obtain the speed characteristic range image.

[0012] Analyze the sharpness and contrast of images within the speed characteristic range. When the image quality index is lower than the preset threshold, adjust the shooting parameter range for the corresponding speed characteristic range.

[0013] Preferably, the specific steps for establishing the correspondence between velocity characteristic intervals and radial positions include: acquiring blade length, blade root radius, and blade tip radius data, and calculating the blade radial length; setting a threshold for the linear velocity difference between adjacent intervals, and dividing radial position points starting from the blade root according to the increasing linear velocity difference threshold; defining the area between adjacent radial position points as a velocity characteristic interval, and recording the starting radial position, ending radial position, and center radial position for each velocity characteristic interval; and establishing a database of velocity characteristic interval numbers and corresponding radial position ranges to form a mapping relationship between velocity characteristic intervals and radial positions.

[0014] Preferably, the real-time linear velocity acquisition process includes: acquiring impeller revolutions per minute data through a speed sensor and converting the revolutions data into angular velocity; acquiring the current angular position of the blade through a blade position sensor and determining the blade's spatial attitude; for each velocity characteristic interval, extracting the central radial position data, multiplying the central radial position by the angular velocity to obtain the instantaneous linear velocity of the velocity characteristic interval; establishing a linear velocity data cache, performing smoothing filtering on linear velocity data for multiple consecutive cycles, and outputting a stable real-time linear velocity value.

[0015] Preferably, the steps for calculating the optimal exposure time and determining the aperture value and ISO parameters are as follows: Based on the principle of motion fuzzy control, a maximum allowable pixel fuzzy distance is set, and the maximum pixel fuzzy distance is divided by the real-time linear velocity of the corresponding velocity characteristic range to obtain the optimal exposure time for the velocity characteristic range; the current ambient light intensity data is obtained through the light sensor, and a reference aperture value is calculated in combination with the target imaging brightness requirements; the aperture value is adjusted according to the ratio of the optimal exposure time to the reference exposure time, and the aperture diameter is increased accordingly when the exposure time is shortened; when the aperture value reaches the physical limit of the lens, exposure compensation is performed by increasing the ISO value.

[0016] Preferably, the process of predicting the time for the target blade's velocity characteristic range to reach the preset shooting position is as follows: set the optimal shooting angle position of the blade and establish the angular coordinates of the shooting position; acquire the current angular position and angular velocity data of the target blade in real time, and calculate the angular difference between the current position of the blade and the preset shooting position; divide the angular difference by the current angular velocity to obtain the estimated time for the blade to reach the shooting position; considering the camera parameter switching time and response delay, trigger the time amount in advance based on the estimated time to generate the shooting control signal corresponding to each velocity characteristic range; establish a shooting timing queue and arrange the control signals according to the order in which each velocity characteristic range reaches the shooting position.

[0017] Preferably, the process of acquiring images of the velocity characteristic range is as follows: after receiving the shooting control signal, the camera is automatically switched to the exposure time, aperture value, and ISO parameters corresponding to the velocity characteristic range; when the velocity characteristic range of the target blade enters the shooting field of view, the camera shutter is triggered to take a picture; the raw image data obtained by the shooting is preprocessed, including noise filtering and brightness equalization; the region corresponding to the velocity characteristic range of the image is extracted, and the image region is identified and numbered in combination with the blade position information; the processed velocity characteristic range image is stored in the image database, and a record of the association between the image and the blade position and shooting time is established.

[0018] Preferably, the specific adjustment process for adjusting the shooting parameter range is as follows: Analyze the sharpness of the images within the speed characteristic range by calculating the image gradient amplitude to obtain sharpness values; analyze the contrast of the images within the speed characteristic range by calculating the standard deviation of the image brightness distribution to obtain contrast values; set acceptable thresholds for sharpness and contrast, and determine that the image quality is unacceptable when either indicator is below the acceptable threshold; for ranges with unacceptable sharpness, shorten the upper limit of the exposure time within the speed characteristic range to reduce the impact of motion blur; for ranges with unacceptable contrast, adjust the aperture value range and / or ISO range within the speed characteristic range to optimize the imaging brightness distribution; update the adjusted parameter range to the shooting parameter database for subsequent shooting control within that speed characteristic range.

[0019] An intelligent imaging control system for wind turbine blades without shutting down the machine includes: a blade radial partitioning module: acquiring the geometric characteristics of the target wind turbine blades, dividing the blades radially into multiple velocity characteristic intervals, each corresponding to a different radial position, and establishing a correspondence between the velocity characteristic intervals and the radial positions; a motion parameter monitoring module: monitoring the impeller speed and blade angle position in real time through sensors, and calculating the real-time linear velocity of the velocity characteristic intervals based on the impeller speed and radial position; a regional imaging parameter calculation module: calculating the optimal exposure time corresponding to each velocity characteristic interval based on the real-time linear velocity, and determining the aperture value and ISO parameters corresponding to the velocity characteristic intervals in conjunction with the current lighting conditions and image quality; a regional imaging image acquisition module: predicting the time for the target blade velocity characteristic intervals to reach the preset imaging positions based on the impeller speed and blade angle position, generating an imaging control signal, and switching the imaging parameters of the corresponding velocity characteristic intervals when the target blade velocity characteristic intervals sequentially reach the imaging positions to acquire the velocity characteristic interval images; and a parameter optimization module: performing sharpness and contrast analysis on the velocity characteristic interval images, and adjusting the imaging parameter range of the corresponding velocity characteristic intervals when the image quality index is lower than a preset threshold.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0021] 1. This invention divides the wind turbine blades radially into multiple speed characteristic ranges and, combined with real-time speed and angle monitoring, establishes corresponding linear velocity calculation models for different regions, enabling precise imaging without shutting down the wind turbine. Through partitioning, this invention dynamically matches the optimal exposure time, aperture value, and ISO for different speed ranges, ensuring both sharpness and brightness requirements in the imaging, thus significantly improving image usability. It fulfills the online monitoring needs in daily wind turbine operation and maintenance, improving data acquisition efficiency while reducing interference with equipment operation, resulting in higher engineering application value and economic benefits.

[0022] 2. This invention not only considers the leaf motion characteristics during the shooting process but also introduces a light intensity monitoring and image quality feedback adjustment mechanism. By setting the maximum pixel blur distance and calculating the optimal exposure time, combined with real-time light intensity data, the aperture value and ISO are dynamically adjusted. This invention can maintain good image brightness and clarity in complex environments such as strong light, shadow, and backlight. Simultaneously, this invention designs a quality analysis module based on image clarity and contrast indicators. When the image is detected to be below a preset threshold, it can automatically shorten the exposure time, adjust the aperture, or increase the ISO range, thereby adaptively correcting the shooting parameters. Through this dynamic optimization mechanism, imaging distortion caused by environmental fluctuations or the complexity of leaf motion is avoided, improving the reliability and stability of monitoring data during long-term operation and maintenance, and providing more robust data support for subsequent defect identification and condition assessment.

[0023] 3. This invention monitors the angular position and angular velocity of the blades in real time, predicts the time it takes for the blade velocity characteristic range to reach the preset shooting position, and generates shooting control signals in advance by combining camera response delay and parameter switching time, thus establishing a partitioned shooting timing queue. Compared with traditional methods of continuous high-speed shooting or manual timed triggering, the timing prediction mechanism of this invention can accurately capture the target range during blade movement, improve the imaging hit rate of a single shutter speed, and avoid a large amount of redundant shooting and storage waste.

[0024] 4. Simultaneously, due to significant differences in parameters across different intervals, this invention can quickly access the corresponding parameter library during blade interval switching, achieving precise switching of partition parameters and thus ensuring the matching degree and balance of images across each interval. This not only improves the efficiency and accuracy of image acquisition but also reduces resource consumption under prolonged high-speed operation. Through the synergistic optimization of partitioning, prediction, and control, this invention achieves automation and intelligence in the imaging process, providing a scalable solution for remote centralized monitoring of large-scale wind turbine clusters. Attached Figure Description

[0025] Figure 1 A flowchart of an intelligent shooting control method for wind turbine blades in a non-stop state provided by the present invention;

[0026] Figure 2 This is a schematic diagram of the shooting time acquisition process provided by the present invention;

[0027] Figure 3 A flowchart illustrating the process of adjusting the range of shooting parameters provided by the present invention;

[0028] Figure 4 The present invention provides a structural diagram of an intelligent shooting control system for wind turbine blades in a non-stop state. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0030] Please see Figure 1This invention provides an intelligent shooting control method for wind turbine blades without shutting down the machine. The technical solution is as follows: The geometric characteristics of the target wind turbine blades are acquired, and the blades are divided radially into multiple velocity characteristic intervals, each corresponding to a different radial position, establishing a correspondence between the velocity characteristic intervals and the radial positions; the impeller speed and blade angle position are monitored in real time using sensors, and the real-time linear velocity of the velocity characteristic intervals is calculated based on the impeller speed and radial position; based on the real-time linear velocity, the optimal exposure time corresponding to each velocity characteristic interval is calculated, and the aperture value and ISO parameters corresponding to the velocity characteristic interval are determined in conjunction with the current lighting conditions and image quality; based on the impeller speed and blade angle position, the time for the target blade velocity characteristic interval to reach the preset shooting position is predicted, and a shooting control signal is generated; when the target blade velocity characteristic intervals sequentially reach the shooting positions, the shooting parameters of the corresponding velocity characteristic intervals are switched to acquire the velocity characteristic interval images; the sharpness and contrast of the velocity characteristic interval images are analyzed, and when the image quality index is lower than a preset threshold, the shooting parameter range of the corresponding velocity characteristic interval is adjusted.

[0031] Specifically, the process involves: acquiring data on blade length, blade root radius, and blade tip radius, and calculating the blade radial length; setting a threshold for the linear velocity difference between adjacent intervals, and dividing radial position points starting from the blade root according to the increasing threshold; defining the region between adjacent radial position points as a velocity characteristic interval, and recording the starting radial position, ending radial position, and center radial position for each velocity characteristic interval; and establishing a database of velocity characteristic interval numbers and corresponding radial position ranges to form a mapping relationship between velocity characteristic intervals and radial positions.

[0032] This application establishes a radial partitioning method for blades, scientifically dividing them according to the linear velocity difference threshold. This ensures the consistency of motion characteristics within each velocity characteristic range, providing a reliable basis for subsequent differentiated shooting parameter settings and improving the accuracy and systematic nature of overall shooting control.

[0033] The real-time linear velocity acquisition process includes: acquiring impeller revolutions per minute data through a speed sensor and converting the revolutions data into angular velocity; acquiring the current angular position of the blade through a blade position sensor to determine the blade's spatial attitude; for each velocity characteristic interval, extracting the central radial position data, multiplying the central radial position by the angular velocity to obtain the instantaneous linear velocity of the velocity characteristic interval; establishing a linear velocity data cache, performing smoothing filtering on linear velocity data from multiple consecutive cycles, and outputting a stable real-time linear velocity value.

[0034] This application obtains stable and reliable real-time linear velocity data through multi-sensor fusion and data filtering, eliminating measurement errors caused by speed fluctuations and vibrations during wind turbine operation, providing high-quality basic data for calculating shooting parameters, and improving the stability and consistency of image shooting.

[0035] Furthermore, the steps for calculating the optimal exposure time and determining the aperture value and ISO parameters are as follows: Based on the principle of motion fuzzy control, a maximum allowable pixel blur distance is set, and the maximum pixel blur distance is divided by the real-time linear velocity of the corresponding velocity characteristic range to obtain the optimal exposure time for the velocity characteristic range; the current ambient light intensity data is obtained through the light sensor, and the reference aperture value is calculated in combination with the target imaging brightness requirements; the aperture value is adjusted according to the ratio of the optimal exposure time to the reference exposure time, and the aperture diameter is increased accordingly when the exposure time is shortened; when the aperture value reaches the physical limit of the lens, exposure compensation is performed by increasing the ISO value.

[0036] In this application, a parameter calculation method based on the principle of motion fuzzy control is used to achieve intelligent coordinated configuration of exposure time, aperture value and sensitivity, effectively balancing motion fuzzy control and imaging brightness requirements, ensuring that clear and bright leaf images can still be obtained under high-speed motion conditions, and greatly improving the controllability of image quality.

[0037] The process of predicting the time it takes for the target blade's velocity characteristic range to reach the preset shooting position is referred to Figure 2 Specifically, the process involves: setting the optimal shooting angle for the blade and establishing the angular coordinates of the shooting position; acquiring the current angular position and angular velocity data of the target blade in real time, and calculating the angular difference between the current position of the blade and the preset shooting position; dividing the angular difference by the current angular velocity to obtain the estimated time for the blade to reach the shooting position; considering the camera parameter switching time and response delay, triggering the time amount in advance based on the estimated time to generate shooting control signals corresponding to each speed characteristic range; and establishing a shooting timing queue, arranging the control signals according to the order in which each speed characteristic range arrives at the shooting position.

[0038] In this application, the timing control problem of high-speed moving blade imaging is solved by using time prediction and early triggering mechanisms, ensuring that the imaging parameters can be switched to the correct position in a timely and accurate manner, avoiding imaging errors caused by response delays, and significantly improving the success rate of imaging and the reliability of image acquisition.

[0039] The process of acquiring images of the velocity characteristic range is as follows: after receiving the shooting control signal, the camera is automatically switched to the exposure time, aperture value, and ISO parameters corresponding to the velocity characteristic range; when the target blade's velocity characteristic range enters the shooting field of view, the camera shutter is triggered to take a picture; the raw image data obtained by the shooting is preprocessed, including noise filtering and brightness equalization; the region corresponding to the velocity characteristic range of the image is extracted, and the image region is identified and numbered in combination with the blade position information; the processed velocity characteristic range image is stored in the image database, and a record of the association between the image and the blade position and shooting time is established.

[0040] In this embodiment, a complete image acquisition and management process was established. Through automatic parameter switching, preprocessing, and data association, a high degree of automation of the shooting process was achieved, reducing the need for manual intervention. At the same time, the integrity and traceability of image data were ensured, providing a reliable data foundation for subsequent image analysis and blade condition assessment.

[0041] The specific steps for performing sharpness and contrast analysis on images within the velocity characteristic range include:

[0042] Multi-scale sharpness analysis was performed on the images in the velocity characteristic range. The Sobel operator, Laplacian operator and Canny edge detection operator were used to calculate the gradient magnitude of the image in the horizontal, vertical and diagonal directions, respectively. The weighted average of the gradient magnitude in the three directions was taken as the comprehensive sharpness value.

[0043] Contrast analysis is performed on the image of the velocity characteristic range, the image gray-level histogram is calculated, and the mean, standard deviation and dynamic range of the image brightness distribution are extracted. The product of the standard deviation and the dynamic range is used as the contrast value.

[0044] A comprehensive image quality evaluation model was established. After normalizing the sharpness and contrast values, the comprehensive image quality index was calculated according to a weight ratio of 6:4.

[0045] The acceptable thresholds are set as follows: sharpness 75, contrast 40, and overall quality index 0.8. If any index falls below the corresponding threshold, the image quality is deemed unacceptable.

[0046] It also includes an image quality assessment step based on frequency domain analysis, specifically:

[0047] A two-dimensional Fourier transform is performed on the velocity characteristic interval image to obtain the frequency domain distribution characteristics of the image. The ability to preserve details of the image is evaluated by analyzing the energy distribution of high-frequency components.

[0048] Calculate the center frequency and spectral width of the image spectrum. The center frequency reflects the main detail features of the image, and the spectral width reflects the richness of detail in the image.

[0049] A motion blur frequency domain feature recognition model is established. The degree of motion blur in the image is identified by the stripe distribution feature of the spectrum. When a clear stripe spectrum distribution is detected, motion blur is determined to exist.

[0050] By combining spatial domain sharpness analysis and frequency domain blur detection results, a dual verification mechanism is established to improve the accuracy and reliability of image quality assessment.

[0051] When frequency domain analysis detects motion blur, it automatically marks the corresponding velocity characteristic range where the exposure time needs to be shortened. When frequency domain analysis shows insufficient high-frequency components, it marks the need to increase the intensity of image sharpening.

[0052] It also includes a machine learning-based image quality prediction and optimization step, specifically: collecting image samples of velocity characteristic ranges under different combinations of shooting parameters to establish a training dataset of shooting parameters and image quality; constructing a deep neural network model, with the input layer including linear velocity, exposure time, aperture value, ISO, and environmental parameters, and the output layer predicting image sharpness and contrast indicators; training the neural network model using historical shooting data, optimizing network weights through backpropagation algorithm to achieve a non-linear mapping between shooting parameters and image quality; before actual shooting, inputting the shooting parameters of the current velocity characteristic range into the trained model to predict image quality indicators; when the predicted image quality is lower than a threshold, calculating the optimal combination of shooting parameters through model inversion and adjusting the shooting parameters in advance; and periodically retraining the model using new shooting data to continuously improve prediction accuracy and parameter optimization effects.

[0053] The machine learning-based predictive optimization function realizes the transformation from passive adjustment to active prediction. By using deep learning models to predict image quality in advance and optimize parameters, it avoids the generation of low-quality images, improves shooting efficiency, and continuously improves the system's intelligence level and optimization effect through continuous learning.

[0054] Adjust shooting parameter range reference Figure 3 The specific adjustment process is as follows: Sharpness analysis is performed on images within the speed characteristic range, and sharpness values ​​are obtained by calculating the image gradient amplitude; contrast analysis is performed on images within the speed characteristic range, and contrast values ​​are obtained by calculating the standard deviation of the image brightness distribution; acceptable thresholds for sharpness and contrast are set, and the image quality is deemed unacceptable when either indicator falls below the acceptable threshold; for ranges with unacceptable sharpness, the upper limit of exposure time within the speed characteristic range is shortened to reduce the impact of motion blur; for ranges with unacceptable contrast, the aperture range and / or ISO range within the speed characteristic range are adjusted to optimize the image brightness distribution; the adjusted parameter ranges are updated to the shooting parameter database for subsequent shooting control within that speed characteristic range.

[0055] In this embodiment, the shooting system achieves self-optimization and continuous improvement through real-time image quality assessment and parameter feedback adjustment mechanism. It can adapt to different operating conditions and environmental changes, ensure the stability of image quality during long-term operation, and significantly improve the system's intelligence level and adaptability.

[0056] It also includes an environmental adaptive shooting parameter adjustment step, specifically: real-time collection of wind speed, temperature, humidity, and weather data through environmental sensors to establish an environmental condition database; establishment of an environmental correction coefficient table based on the influence of different environmental conditions on the reflective characteristics of the blade surface and imaging contrast; when calculating shooting parameters within the speed characteristic range, extraction of the correction coefficient corresponding to the current environmental conditions from the environmental correction coefficient table to correct the base exposure time, aperture value, and ISO; addition of a blade vibration compensation algorithm for strong wind environments, monitoring the blade vibration frequency through an accelerometer, and introducing vibration phase compensation into the shooting sequence; and establishment of a correlation model between historical environment and image quality to achieve automatic optimization of shooting parameters under different environmental conditions.

[0057] The environmental adaptive function enables the system to automatically respond to the effects of different weather and environmental conditions. Through environmental correction coefficients and vibration compensation algorithms, it effectively eliminates the interference of external environmental factors on image quality, achieving stable shooting in all weather and all environments, and greatly expanding the application scope and practicality of the system.

[0058] This application also includes a multi-blade collaborative shooting control step, specifically: identifying the number of wind turbine blades, establishing a synchronous monitoring system for the angle positions of multiple blades, and acquiring the spatial attitude of each blade in real time; calculating the time series of the speed characteristic intervals of each blade reaching the shooting position based on the angle intervals between blades, and generating a multi-blade shooting time sequence table; establishing a shooting resource scheduling algorithm, and sorting the shooting sequence according to the blade number and interval priority when the speed characteristic intervals of multiple blades arrive at the shooting position simultaneously; adopting a high-speed continuous shooting mode to achieve single-trigger multi-frame image acquisition for the case where blades rapidly and continuously pass through the shooting position under high speed conditions; establishing a unified management mechanism for multi-blade image data, and marking each acquired blade speed characteristic interval image with blade number and timestamp recording; and identifying the differences between blades through comparative analysis of multi-blade image data, providing data support for blade health status assessment.

[0059] The multi-blade collaborative control function enables unified and efficient detection of all blades of the wind turbine. Through intelligent scheduling and resource optimization, it significantly improves detection efficiency. At the same time, through multi-blade data comparison and analysis, it can more accurately identify differences and potential problems between blades, providing more comprehensive data support for the overall health status assessment of the wind turbine.

[0060] This invention proposes an intelligent imaging control method for wind turbine blades under continuous operation. Through radial blade partitioning, real-time velocity monitoring, motion fuzz control, and intelligent parameter optimization, it achieves high-quality imaging of high-speed rotating blades. This method utilizes sensor fusion and filtering to obtain stable linear velocity data, combined with a predictive triggering mechanism to ensure precise shooting timing and avoid shooting errors caused by delays. Intelligent coordination of exposure time, aperture value, and ISO balances motion blur and brightness requirements, significantly improving image clarity and contrast. Simultaneously, the system possesses machine learning predictive optimization, image quality feedback adaptive adjustment, and environmental correction functions, enabling proactive parameter prediction and optimization to adapt to different operating and weather conditions, ensuring long-term stable imaging. A multi-blade collaborative control mechanism further improves detection efficiency, supports inter-blade difference analysis, and provides comprehensive data support for blade health status assessment. Overall, this invention achieves efficient, intelligent, and environmentally adaptive blade imaging control, significantly improving the accuracy and reliability of continuous wind turbine detection.

[0061] Example 2:

[0062] This invention provides a specific implementation of an intelligent shooting control system for wind turbine blades without shutting down, referring to... Figure 4 The detailed content includes: a blade radial partitioning module: acquiring the geometric characteristics of the target wind turbine blades, dividing the blades radially into multiple velocity characteristic intervals, each corresponding to a different radial position, and establishing a correspondence between the velocity characteristic intervals and radial positions; a motion parameter monitoring module: monitoring the impeller speed and blade angle position in real time through sensors, and calculating the real-time linear velocity of the velocity characteristic interval based on the impeller speed and radial position; a regional shooting parameter calculation module: calculating the optimal exposure time corresponding to each velocity characteristic interval based on the real-time linear velocity, and determining the aperture value and ISO parameters corresponding to the velocity characteristic interval based on the current lighting conditions and image quality; a regional shooting image acquisition module: predicting the time for the target blade velocity characteristic interval to reach the preset shooting position based on the impeller speed and blade angle position, generating a shooting control signal, and switching the shooting parameters of the corresponding velocity characteristic interval when the target blade velocity characteristic intervals sequentially reach the shooting position to acquire the velocity characteristic interval image; and a parameter optimization module: performing sharpness and contrast analysis on the velocity characteristic interval image, and adjusting the shooting parameter range of the corresponding velocity characteristic interval when the image quality index is lower than a preset threshold.

[0063] This embodiment uses an intelligent shooting platform with a gimbal camera mounted on a drone, and the specific configuration is as follows:

[0064] The drone platform utilizes a DJI M350 RTK quadcopter, equipped with high-precision RTK positioning capabilities, achieving centimeter-level positioning accuracy and providing a reliable position reference for precise hovering and path planning. The aircraft has an IP55 protection rating, making it suitable for operation in complex environments such as at sea and on land, and can meet the inspection needs of large wind farms. The gimbal camera system employs a specially designed EO61a gimbal camera, integrating dual-spectrum imaging capabilities and laser ranging functions.

[0065] The gimbal camera comprises the following core components: 1. Visible light wide-angle camera module: Utilizing a 48-megapixel CMOS sensor with a field of view of 101.26°×84.9°, it achieves wide-range scene coverage, suitable for overall wind turbine outline recognition and orientation determination. 2. Visible light telephoto camera module: Equipped with a full-frame sensor (35.7×23.8mm Exmor R CMOS) and an 85mm F1.8 ultra-large aperture telephoto lens. This module features multiple focus modes, including single focus (AF_S), autofocus (AF_A), continuous focus (AF_C), full-time manual focus (DMF), manual focus mode (MF), and continuous preset focus (PF), ensuring accurate focus during high-speed blade movement photography. The maximum mechanical shutter speed reaches 1 / 4000 second, effectively controlling motion blur and meeting the requirements for clear imaging of high-speed areas at blade tips. 3. Laser Ranging Module: Integrates a 905nm wavelength laser rangefinder with a measurement range of 0-180 meters and a ranging frequency of 7000Hz, enabling real-time acquisition of precise distance information between the camera and the blade surface. This module works in conjunction with the shooting parameter calculation module to dynamically adjust focal length and exposure parameters based on real-time distance data. 4. Three-Axis Stabilization Gimbal System: The gimbal stabilization accuracy reaches ±0.01°, with a rotation range of -115° to +45° on the pitch axis and 0° to 330° on the yaw axis. It is driven by a brushless motor and supports independent control of the horizontal, pitch, and roll axes. The gimbal system can maintain camera stability during UAV flight, eliminating the impact of aircraft vibration on image quality. This embodiment also features a domestically produced edge computing board with 6T computing power for real-time image data processing, AI algorithm execution, and gimbal control, ensuring the system's real-time responsiveness.

[0066] This embodiment uses a 2.5 MW wind turbine generator set in an offshore wind farm as the application object. The rotor diameter is 110 meters, the rated speed is 16 revolutions per minute, and the blade length is 53 meters.

[0067] In this embodiment, the geometric characteristic data of the target wind turbine blades are first obtained. The wind turbine's technical specifications indicate a blade length of 53 meters, a root radius of 1.5 meters, and a tip radius of 55 meters, resulting in a calculated radial length of 53.5 meters. Considering the high-speed characteristics of offshore wind turbines, a threshold of 8 meters per second is set for the linear velocity difference between adjacent zones to ensure relative consistency of motion characteristics within each zone. Radial division is performed starting from the blade root, establishing a velocity characteristic zone database numbered Zone1 to Zone5.

[0068] In this embodiment, a speed sensor is installed in the wind turbine nacelle to monitor the impeller speed in real time; the blade angle position information is obtained by combining a laser displacement sensor and a photoelectric encoder; under normal wind turbine operation, the speed sensor collects data 100 times per second and converts the revolutions per minute into angular velocity; at the same time, the blade position sensor obtains the current angle position of the blade at a frequency of 200 times per second.

[0069] For each velocity characteristic interval, the corresponding central radial position data is extracted, and the central radial position is multiplied by the real-time angular velocity to calculate the instantaneous linear velocity of each interval. A linear velocity data buffer with a capacity of 500 data points is established, and a combination of moving average filtering and Kalman filtering is used to smooth the linear velocity data of multiple consecutive cycles, eliminating velocity fluctuations caused by gusts or mechanical vibrations, and outputting stable and reliable real-time linear velocity values.

[0070] Based on the principle of motion fuzzy control, this embodiment sets the maximum allowable pixel fuzzy distance to 2 pixels and obtains the actual distance greater than 2 pixels. For Zone 1, when the real-time linear velocity is 2.1 meters per second, the optimal exposure time is calculated to be 1 / 400 second; for Zone 2, when the linear velocity is 4.8 meters per second, the optimal exposure time is 1 / 800 second; for Zone 3, when the linear velocity is 8.5 meters per second, the optimal exposure time is 1 / 1600 second; for Zone 4, when the linear velocity is 13.2 meters per second, the optimal exposure time is 1 / 2500 second; and for Zone 5, when the linear velocity is 16.8 meters per second, the optimal exposure time is 1 / 3200 second.

[0071] Ambient light monitoring employs a broadband light sensor with a measurement range of 0.1 lux to 100,000 lux. Based on the target imaging brightness requirements, a baseline aperture value of f / 8.0 is calculated under standard lighting conditions. The aperture value is adjusted according to the ratio of the optimal exposure time to the baseline exposure time for each zone, following the exposure compensation principle: Zone 1 aperture value is adjusted to f / 5.6; Zone 2 to f / 6.3; Zone 3 to f / 7.1; Zone 4 to f / 8.0; and Zone 5 to f / 9.0. When the aperture value reaches the lens's physical limit of f / 2.8, compensation is achieved by increasing the ISO sensitivity, dynamically adjusted within the ISO range of 100 to 3200.

[0072] In this embodiment, the optimal shooting angle for the blade is set to 90 degrees to the right horizontally, establishing an angular coordinate system for the shooting position. The system acquires the target blade's current angular position and angular velocity data in real time. When the blade's angle difference from the preset shooting position reaches 30 degrees, the system begins calculating the arrival time. Considering that the camera parameter switching time is approximately 0.02 seconds and the image sensor response delay is approximately 0.01 seconds, the shooting control signal is triggered 0.05 seconds before the estimated arrival time.

[0073] A time-series shooting queue is established, with the system sending shooting control signals for the corresponding zones sequentially as the blade passes through the shooting position, according to the order from Zone 1 to Zone 5. The shooting time window for each zone is approximately 0.08 seconds, ensuring that the images of the corresponding speed characteristic zones can be captured completely.

[0074] After receiving the shooting control signal, the camera automatically switches to the shooting parameter combination corresponding to the speed characteristic range; when the target blade speed characteristic range enters the shooting field of view of the 200mm focal length lens, the shutter is triggered to take a picture; the raw image data obtained by shooting is first preprocessed, including Gaussian noise filtering and histogram equalization processing, to improve the image signal-to-noise ratio and dynamic range.

[0075] Based on the blade geometry model and the current angular position, the system automatically extracts the rectangular region corresponding to the velocity characteristic range in the image, and identifies and numbers the image region in combination with the blade position information; the processed velocity characteristic range image is stored in the image database in TIFF format, and a complete association record is established between the image and the blade position and the shooting time.

[0076] The quality of the acquired speed characteristic range images is evaluated using the Sobel operator to calculate the image gradient magnitude. The sharpness index is quantified by the gradient average, with a sharpness pass threshold set at 80. Contrast analysis is performed by calculating the standard deviation of the image grayscale histogram, with a contrast pass threshold set at 45. When the sharpness measurement value of the Zone 4 image is 72, which is below the pass threshold, the system automatically shortens the upper limit of the exposure time for this zone from 1 / 2500 second to 1 / 3200 second to reduce motion blur. When the contrast measurement value of the Zone 1 image is 38, which is below the pass threshold, the system adjusts the aperture value for this zone from f / 5.6 to f / 4.5, and simultaneously increases the ISO sensitivity from 100 to 200.

[0077] The adjusted parameter range is updated in real time to the shooting parameter database for subsequent shooting control within this speed characteristic range, forming an adaptive and optimized closed-loop control system.

[0078] This embodiment is equipped with an environmental sensor array, including an anemometer, temperature and humidity sensors, and weather condition monitoring equipment, to collect marine environmental data in real time. Based on historical data analysis, an environmental correction factor table is established: the correction factor is 1.0 for sunny conditions; 1.2 for cloudy conditions; 1.5 for overcast conditions; and 2.0 for foggy conditions. When calculating shooting parameters, the correction factor corresponding to the current weather conditions is extracted from the correction factor table to correct the base exposure time and ISO.

[0079] To address the challenges of strong winds at sea, the system incorporates a blade vibration compensation algorithm. A triaxial accelerometer installed at the blade root monitors the blade's vibration frequency under wind load. Vibration phase compensation is integrated into the timing calculation for image capture; when the blade is detected to be at a vibration trough, image capture is triggered, reducing image blurring caused by vibration.

[0080] This embodiment focuses on a three-bladed fan, establishing a multi-blade angle and position synchronous monitoring system to acquire the spatial attitude of the three blades in real time. Based on 120-degree angle intervals, the time series of each blade's velocity characteristic range reaching the shooting position is calculated, generating a multi-blade shooting time sequence table. When multiple blade velocity characteristic ranges arrive at the shooting position simultaneously, they are sorted according to blade number priority and range importance, with priority given to shooting the high-speed range of the blade tip area.

[0081] Under high-speed operation, the blades rapidly and continuously pass through the shooting position. The system adopts a high-speed continuous shooting mode of 10 frames per second to achieve the acquisition of multiple frames of images in a single trigger, thereby improving the success rate of shooting. A unified management mechanism for multi-blade image data is established, and the acquired images of each blade speed characteristic range are identified by blade number and recorded with millisecond-level timestamps. Through comparative analysis of multi-blade image data, geometric differences and surface condition differences between blades are identified, providing data support for blade health status assessment and fault early warning.

[0082] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent shooting control of wind turbine blades without stopping the machine, characterized in that, include: Obtain the blade geometry characteristics of the target wind turbine, divide the blade radially into multiple velocity characteristic intervals, each velocity characteristic interval corresponds to a different radial position, and establish the correspondence between the velocity characteristic intervals and the radial positions; The impeller speed and blade angle position are monitored in real time by sensors, and the real-time linear velocity of the speed characteristic range is calculated based on the impeller speed and radial position. Based on the real-time linear velocity, the optimal exposure time corresponding to each speed characteristic range is calculated, and the aperture value and ISO parameter corresponding to the speed characteristic range are determined in combination with the current lighting conditions and image quality. Based on the impeller speed and blade angle position, predict the time when the target blade speed characteristic range reaches the preset shooting position, generate a shooting control signal, and switch the shooting parameters of the corresponding speed characteristic range when the target blade speed characteristic range reaches the shooting position in sequence to obtain the speed characteristic range image. Analyze the sharpness and contrast of images within the speed characteristic range. When the image quality index is lower than the preset threshold, adjust the shooting parameter range for the corresponding speed characteristic range.

2. The intelligent shooting control method for wind turbine blades under non-stop operation as described in claim 1, characterized in that: The specific steps for establishing the correspondence between velocity characteristic intervals and radial positions include: acquiring blade length, blade root radius, and blade tip radius data, and calculating the blade radial length; setting a threshold for the linear velocity difference between adjacent intervals, and dividing radial position points starting from the blade root according to the increasing linear velocity difference threshold; defining the area between adjacent radial position points as a velocity characteristic interval, and recording the starting radial position, ending radial position, and center radial position for each velocity characteristic interval; and establishing a database of velocity characteristic interval numbers and corresponding radial position ranges to form a mapping relationship between velocity characteristic intervals and radial positions.

3. The intelligent shooting control method for wind turbine blades under non-stop operation as described in claim 1, characterized in that: The real-time linear velocity acquisition process includes: acquiring impeller revolutions per minute data through a speed sensor and converting the revolutions data into angular velocity; acquiring the current angular position of the blade through a blade position sensor to determine the blade's spatial attitude; for each velocity characteristic interval, extracting the central radial position data, multiplying the central radial position by the angular velocity to obtain the instantaneous linear velocity of the velocity characteristic interval; establishing a linear velocity data cache, performing smoothing filtering on linear velocity data from multiple consecutive cycles, and outputting a stable real-time linear velocity value.

4. The intelligent shooting control method for wind turbine blades under non-stop operation as described in claim 3, characterized in that: The steps for calculating the optimal exposure time and determining the aperture value and ISO parameters are as follows: Based on the principle of motion fuzz control, set the maximum allowable pixel fuzz distance, divide the maximum pixel fuzz distance by the real-time linear velocity of the corresponding velocity characteristic range, and obtain the optimal exposure time of the velocity characteristic range. The system acquires ambient light intensity data using a light sensor, calculates a reference aperture value based on the target imaging brightness requirements, and adjusts the aperture value according to the ratio of the optimal exposure time to the reference exposure time, increasing the aperture diameter when the exposure time is shortened. When the aperture value reaches the lens's physical limit, exposure compensation is performed by increasing the ISO value.

5. The intelligent shooting control method for wind turbine blades under non-stop operation as described in claim 1, characterized in that: The process of predicting the time for the target blade to reach the preset shooting position within its velocity characteristic range is as follows: set the optimal shooting angle position for the blade and establish the angular coordinates of the shooting position; acquire the current angular position and angular velocity data of the target blade in real time and calculate the angle difference between the current position of the blade and the preset shooting position; divide the angle difference by the current angular velocity to obtain the estimated time for the blade to reach the shooting position. Considering camera parameter switching time and response delay, the timing is triggered in advance based on the expected time to generate shooting control signals corresponding to each speed characteristic range; a shooting timing queue is established to arrange the control signals according to the order in which each speed characteristic range arrives at the shooting position.

6. The intelligent shooting control method for wind turbine blades under non-stop operation as described in claim 5, characterized in that: The process of acquiring images within a speed characteristic range is as follows: after receiving the shooting control signal, the camera automatically switches to the corresponding speed characteristic range for exposure time, aperture value, and ISO parameters; When the target blade's velocity characteristic range enters the field of view, the camera shutter is triggered to take a picture; the raw image data is preprocessed, including noise filtering and brightness equalization; the region corresponding to the velocity characteristic range of the image is extracted, and the image region is identified and numbered in combination with the blade position information; the processed velocity characteristic range image is stored in the image database, and a record of the association between the image and the blade position and shooting time is established.

7. The intelligent shooting control method for wind turbine blades under non-stop operation as described in claim 6, characterized in that: The specific adjustment process for the shooting parameter range is as follows: Analyze the sharpness of images within the speed characteristic range by calculating the image gradient amplitude to obtain sharpness values; analyze the contrast of images within the speed characteristic range by calculating the standard deviation of the image brightness distribution to obtain contrast values; set acceptable thresholds for sharpness and contrast, and determine image quality as unacceptable when either indicator falls below the acceptable threshold; for ranges with unacceptable sharpness, shorten the upper limit of exposure time within the speed characteristic range; for ranges with unacceptable contrast, adjust the aperture range and / or ISO range within the speed characteristic range to optimize the image brightness distribution; update the adjusted parameter range to the shooting parameter database for subsequent shooting control within that speed characteristic range.

8. An intelligent shooting control system for wind turbine blades without stopping the machine, characterized in that, include: Blade radial partitioning module: acquires the blade geometry characteristics of the target wind turbine, divides the blade radially into multiple velocity characteristic intervals, each velocity characteristic interval corresponds to a different radial position, and establishes the correspondence between velocity characteristic intervals and radial positions; Motion parameter monitoring module: monitors the impeller speed and blade angle position in real time through sensors, and calculates the real-time linear velocity of the velocity characteristic interval based on the impeller speed and radial position; The regional shooting parameter calculation module calculates the optimal exposure time for each speed characteristic range based on the real-time linear velocity, and determines the aperture value and ISO parameters for each speed characteristic range in conjunction with the current lighting conditions and image quality. The regional shooting image acquisition module predicts the time it takes for the target blade's speed characteristic range to reach the preset shooting position based on the impeller rotation speed and blade angle position, generates a shooting control signal, and switches the shooting parameters for the corresponding speed characteristic range when the target blade's speed characteristic range sequentially reaches the shooting position, acquiring the speed characteristic range image. The parameter optimization module performs sharpness and contrast analysis on the speed characteristic range image; when the image quality index is lower than a preset threshold, it adjusts the shooting parameter range for the corresponding speed characteristic range.

Images