A steel wire rope diameter measurement method and system based on event compensation correction

Abstract

The present application relates to steel wire rope state detection technical field, especially in kind based on event compensation correction steel wire rope diameter measurement method and system, its technical scheme includes imaging module, boundary initial extraction module, dynamic motion estimation module, dynamic correction compensation module and diameter calculation module, the imaging module includes traditional camera, event camera, optical component, synchronous trigger device and beam splitter, the traditional camera is used for collecting the frame image of steel wire rope, the present application relies on the complementary advantages of event camera and traditional camera, through event driven dynamic geometric compensation model, the systematic measurement deviation under dynamic working condition is solved from the mechanism, the breakthrough of quasi-static imaging dependence is broken through, without additional constraint or frame rate promotion, control cost and hardware complexity, avoid signal-to-noise ratio decline and other problems, realize high-precision stable measurement under high speed, vibration working condition, adapt to multi-industrial scene online monitoring, practicality and application value are remarkable.

Description

Technical Field

[0001] This invention relates to the field of wire rope condition detection technology, and in particular to a method and system for measuring wire rope diameter based on event compensation correction. Background Technology

[0002] With the development of industrial automation and intelligent detection technologies, online non-contact measurement of wire rope diameter has become a core requirement in key areas such as mine hoisting and elevator traction. As a core load-bearing component, changes in wire rope diameter directly affect load assessment, wear determination, and operational safety; accurate measurement is an important prerequisite for ensuring safe and reliable industrial production.

[0003] Compared to traditional methods such as manual calipers and contact measurement, non-contact measurement based on machine vision has become the mainstream technology due to its advantages such as high efficiency, continuous online monitoring, and adaptability to harsh environments. Existing solutions mostly use area array or line array cameras (collectively referred to as "traditional cameras") to acquire images, and obtain the actual diameter through processes such as preprocessing, edge detection, and fitting conversion. Ideal accuracy can be achieved under quasi-static conditions of low speed, slight vibration, and stable lighting.

[0004] However, in actual working conditions, the wire rope is often in a state of high-speed operation and periodic vibration. Coupled with interference from changes in lighting and metal reflection, the inherent defects of traditional camera measurement become apparent. Traditional cameras use periodic integral imaging, and the movement of the wire rope within the exposure period can easily lead to boundary superposition and blurring, producing systematic offsets that are difficult to eliminate through post-processing. Even if the shutter speed is optimized, the frame rate is increased, or the exposure time is shortened, it is impossible to break through the trade-off between "exposure time - movement speed - accuracy", and it is easy to cause problems such as decreased signal-to-noise ratio and increased cost, making it difficult to operate stably in the long term.

[0005] Existing improvement solutions, such as optimizing shooting angles, adding mechanical constraints, and multi-camera collaboration, all have limitations: mechanical constraints reduce adaptability, multi-camera solutions are complex and costly to synchronize, and none of them completely eliminate dynamic systematic errors, resulting in insufficient stability under vibration conditions. Therefore, suppressing motion errors under dynamic conditions is the core bottleneck for the widespread application of existing technologies.

[0006] In recent years, event cameras have entered the industrial field as a new type of bionic sensor. They operate in an asynchronous pixel mode, outputting event information containing position, microsecond-level timestamps, and polarity only when brightness changes. They have the advantages of high temporal resolution and anti-motion blur, and can accurately capture the transient motion trajectory of targets.

[0007] However, event cameras output sparse event streams, lacking complete grayscale textures and spatial contours, and cannot perform high-precision geometric measurements independently. The two are significantly complementary: traditional cameras are good at acquiring spatial contour references but are easily affected by dynamic errors, while event cameras are good at capturing transient motion but lack spatial information support.

[0008] Current technologies fail to effectively integrate these two approaches, and cannot utilize event cameras to compensate for dynamic biases in traditional vision, resulting in a significant technological gap. Existing solutions either lack sufficient dynamic adaptability or accuracy, making it difficult to balance dynamic adaptability with high-precision measurement requirements. Therefore, a novel technological solution is urgently needed to fill this gap. This application proposes a method and system for measuring the diameter of steel wire ropes based on event compensation correction. Summary of the Invention

[0009] The purpose of this invention is to address the problems in the background technology that traditional camera measurements are prone to systematic errors under dynamic working conditions of wire ropes, that event cameras cannot perform high-precision geometric measurements on their own, and that the two have not been effectively integrated to complement each other's shortcomings. The invention proposes a wire rope diameter measurement method and system based on event compensation correction.

[0010] The technical solution of the present invention: a wire rope diameter measurement system based on event compensation correction, comprising an imaging module, a boundary initial extraction module, a dynamic motion estimation module, a dynamic correction compensation module, and a diameter calculation module;

[0011] The imaging module includes a conventional camera, an event camera, optical components, a synchronization triggering device, and a beam splitter. The conventional camera is used to acquire frame images of the wire rope, the event camera is used to acquire event streams generated by the movement of the wire rope's edges, the optical components and the beam splitter work together to achieve optical path transmission and beam splitting, and the synchronization triggering device is used to ensure that the conventional camera and the event camera acquire images synchronously.

[0012] The initial boundary extraction module is used to preprocess the frame image and extract the boundary to obtain the initial left and right boundary coordinates of the wire rope.

[0013] The dynamic motion estimation module estimates the transient displacement function of the wire rope boundary during the exposure cycle based on the event flow, as well as the displacement correction coordinates of the left and right boundaries;

[0014] The dynamic correction and compensation module is used to perform geometric compensation and boundary correction on the initial left and right boundary coordinates according to the transient displacement function and displacement correction coordinates, so as to obtain the equivalent static left and right boundary coordinates.

[0015] The diameter calculation module calculates the diameter of the steel wire rope in the pixel domain based on the equivalent static left and right boundary coordinates, and then converts it to the actual physical diameter of the steel wire rope by combining the calibration ratio.

[0016] Optionally, the imaging module adopts a dual-camera coaxial heterogeneous optical path structure, including a beam splitter and at least three prisms. The light reflected by the steel wire rope is transmitted to the beam splitter through the prisms. The beam splitter splits the light beams and projects them onto the event camera and the conventional camera through the corresponding prisms, so that the event camera and the conventional camera have the same optical field of view and perspective relationship.

[0017] Optionally, the synchronization triggering device is a hardware-level synchronization and triggering unit based on FPGA. The FPGA generates a high-precision master clock to provide a unified time reference for traditional cameras and event cameras, and timestamps the event stream and frame images at the nanosecond level. When the accumulated events exceed a preset threshold, the FPGA sends a trigger signal to the traditional camera to control the exposure.

[0018] This invention also proposes a method for measuring the diameter of a wire rope based on event compensation correction, applied to the aforementioned system, comprising the following steps:

[0019] Step 1: Synchronous Data Acquisition: The traditional camera and event camera are controlled to acquire data synchronously through a synchronous triggering device. Within an exposure window, frame images and event streams of the wire rope are obtained and uploaded to the computer.

[0020] Step 2: Extract initial boundaries: Preprocess the frame image, extract the set of boundary points of the wire rope, estimate the axial direction vector and normal direction vector of the wire rope, scan along the normal to obtain the gray-level gradient, and use the gray-level gradient maximum point as the initial left and right boundary coordinates;

[0021] Step 3: Estimate transient motion information: Project the event stream onto the image coordinate system, filter the event set that satisfies the time and space constraints in the boundary neighborhood, construct the transient displacement function of the wire rope boundary based on the event set, and calculate the displacement correction coordinates of the left and right boundaries;

[0022] Step 4: Correct the boundary coordinates: Using the exposure midpoint time as a reference time, calculate the average boundary displacement using the transient displacement function, and use the average displacement to perform geometric compensation on the initial left and right boundary coordinates to obtain the equivalent static left and right boundary coordinates.

[0023] Step 5: Calculate the actual diameter: Perform axial discrete sampling on the left and right boundary coordinates of the equivalent stationary point, calculate the pixel domain diameter sequence, and after obtaining the median and average value through sliding window statistics, combine it with the calibration ratio to obtain the actual physical diameter of the wire rope.

[0024] Optionally, the specific process for extracting the initial left and right boundary coordinates in step two is as follows:

[0025] The preprocessed image is obtained after preprocessing frame image I. The initial set of boundary points is obtained through boundary extraction. Calculate the set of boundary points Given the mean coordinates and covariance matrix, perform eigenvalue decomposition on the covariance matrix and take the eigenvector corresponding to the largest eigenvalue as the axial direction vector. :

[0026] That is: axial direction vector

[0027] in, This represents the eigenvector corresponding to the largest eigenvalue; This represents the i-th eigenvalue;

[0028] The vector perpendicular to the axial direction vector is taken as the normal direction vector. :

[0029] That is: normal direction vector ;

[0030] Project the coordinates of any point in the image coordinate system onto the axial and normal directions, scan along the normal direction at each position of the axial direction to obtain a one-dimensional gray-scale profile, and calculate the differential gradient of the gray-scale profile.

[0031] Among them, the one-dimensional grayscale profile Represented as:

[0032] ;

[0033] Represents the coordinates of any point in the image coordinate system; express The projected coordinates; and Representing coordinates Axial coordinates projected along the axial direction and normal coordinates projected along the normal direction;

[0034] Indicates coordinates The pixel grayscale value at that location;

[0035] right The gray-level gradient of the profile along the normal direction can be obtained by taking the difference. ,Right now:

[0036]

[0037] The maximum point of the gray-scale gradient is the initial boundary of the wire rope, which corresponds to the normal coordinates of the initial left and right boundaries, and then the coordinates of the initial left and right boundaries can be calculated.

[0038] Specifically, the initial boundary of the wire rope corresponds to the gray-scale gradient of the cross-section. Let the maximum point be denoted as the normal coordinates of the initial left and right boundaries, respectively. and Then we have:

[0039]

[0040] ;

[0041] in, Indicates the maximum search radius;

[0042] Then the initial left and right boundary coordinates of the wire rope , They can be represented as:

[0043]

[0044] .

[0045] Optionally, the specific conditions for filtering the boundary neighborhood event set in step three are: retaining events whose minimum distance from the event point to the initial boundary point set does not exceed the preset boundary neighborhood radius, and whose event trigger time is within the exposure window;

[0046] Specifically, this includes utilizing the event stream obtained through the steps. Calculate the exposure window Inside, the boundary is in the axial position Transient displacement function in the normal direction at time t Displacement correction coordinates of the left and right boundaries , ;in, This is the initial exposure time; Exposure time;

[0047] After projecting the event onto the image coordinate system, let the first... One event Event points in to the initial set of boundary points The minimum distance is ,Right now:

[0048]

[0049] in, Indicates the first The coordinates of an event in the event camera pixel array; Indicates the first The timestamp at which the event was triggered; Indicates the first The polarity of brightness change for each event; b indicates the polarity of brightness change for each event. Traversing all boundary points;

[0050] Preserving the boundary neighborhood event set , Both time and space constraints must be satisfied, that is:

[0051]

[0052] in, The radius of the boundary neighborhood;

[0053] The main geometric disturbance of the wire rope within the exposure window can often be approximated as a displacement along the normal direction, hence:

[0054]

[0055]

[0056] ;

[0057] in, Indicates the axial position of the left and right boundaries of the wire rope time At that time, relative to the initial left and right boundary coordinates , The transient displacement function;

[0058] Indicates according to estimate The optimization function;

[0059] and These represent the axial positions of the left and right boundaries of the wire rope, respectively. time Corrected coordinates at that time.

[0060] Optionally, when the wire rope undergoes periodic vibration during the exposure time, the transient displacement function is constructed in step three. The process is as follows:

[0061] The axial coordinate of the wire rope is divided into multiple segment intervals. The displacement in each segment interval is represented by a first-order harmonic model. The fundamental frequency of the harmonic model is given by prior mechanical parameters.

[0062] Collect the event set within each segment interval, construct the displacement observation residual for each event, and perform robust minimization of the displacement observation residual based on the Huber loss function to obtain the harmonic coefficients of each segment interval.

[0063] By splicing the displacement models of each segmented interval, the global transient displacement function is obtained.

[0064] Optionally, the specific process of the sliding window statistics in step five is as follows:

[0065] A sliding window process is applied to the pixel domain diameter sequence, and the median of the diameter values ​​within each sliding window is taken as the final pixel domain diameter within the window.

[0066] The median of all windows is averaged to obtain the final pixel domain diameter measurement.

[0067] Compared with the prior art, this application includes at least one of the following beneficial technical effects:

[0068] This invention utilizes the microsecond-level high temporal resolution of an event camera to accurately characterize the transient displacement and vibration patterns of the wire rope boundary within the exposure cycle of a traditional camera. It overcomes the motion blur defect of traditional camera integral imaging from the perspective of imaging mechanism. Through dynamic compensation correction driven by event flow, it does not rely on the quasi-static imaging assumption. Even under the condition of high-speed operation and periodic vibration of the wire rope, it can obtain the equivalent static boundary coordinates, so that the measurement results maintain good consistency between consecutive frames and significantly improve the measurement repeatability.

[0069] This invention breaks through the inherent limitations of traditional measurement technologies, balancing measurement performance and system economy. By drastically shortening the exposure time or significantly increasing the frame rate of traditional cameras, motion blur is suppressed, effectively avoiding the problems caused by increased lighting requirements, decreased image signal-to-noise ratio, increased hardware costs, and a surge in data processing pressure. The event camera only outputs the event stream asynchronously when the brightness changes, resulting in a small data volume and controllable computational burden. Combined with the adaptation design of the dual-camera coaxial heterogeneous optical path structure and FPGA synchronous triggering device, the synergistic optimization of temporal and spatial resolution is achieved without significantly increasing hardware complexity and modification costs.

[0070] This invention eliminates the need for additional mechanical constraints on the operating state of the wire rope, making it suitable for complex industrial scenarios such as mine hoisting, elevator traction, cableway transportation, and marine engineering. It enables continuous online monitoring of the wire rope diameter. The high-precision diameter data it outputs provides reliable support for wire rope wear assessment, life prediction, and equipment safety early warning, filling the technological gap in high-precision non-contact measurement under dynamic working conditions and has broad application prospects.

[0071] This invention is the first to introduce an event camera as a dynamic geometric compensation information source into the measurement of wire rope diameter. It establishes a fusion mechanism between event data and traditional camera spatial geometric information. Through axial segmentation modeling, event flow spatiotemporal filtering, and robust optimization solution, it achieves an organic combination of transient motion estimation and boundary correction, breaking through the dependence of existing methods on a single camera and providing a new technical approach for geometric measurement in dynamic scenes.

[0072] In summary, this invention leverages the complementary advantages of event cameras and traditional cameras, employing an event-driven dynamic geometric compensation model to address systematic measurement deviations under dynamic conditions from a mechanistic perspective, overcoming the dependence on quasi-static imaging. It eliminates the need for additional constraints or increased frame rates, controls costs and hardware complexity, and avoids issues such as reduced signal-to-noise ratio. It achieves high-precision and stable measurements under high-speed and vibration conditions, adapting to online monitoring in multiple industrial scenarios, demonstrating significant practicality and application value. Attached Figure Description

[0073] Figure 1 This is a block diagram illustrating the principle of a wire rope diameter measurement method and system based on event compensation correction.

[0074] Figure 2 This is the optical path diagram of the imaging system proposed in this invention;

[0075] Reference numerals: 1. Prism; 2. Event camera; 3. Conventional camera; 4. Synchronization triggering device; 5. Beam splitter. Detailed Implementation

[0076] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0077] Example

[0078] like Figure 1 As shown, the present invention proposes a wire rope diameter measurement system based on event compensation correction, which includes an imaging module, a boundary initial extraction module, a dynamic motion estimation module, a dynamic correction compensation module, and a diameter calculation module.

[0079] The imaging module includes a conventional camera 3, an event camera 2, optical components, a synchronization triggering device 4, and a beam splitter 5. The conventional camera 3 is used to acquire frame images of the steel wire rope, and the event camera 2 is used to acquire event streams generated by the movement of the steel wire rope's edges. The optical components and the beam splitter 5 work together to achieve optical path transmission and beam splitting. The synchronization triggering device 4 is used to ensure that the conventional camera 3 and the event camera 2 acquire images synchronously. The imaging module adopts a dual-camera coaxial heterogeneous optical path structure, including a beam splitter 5 and at least three prisms 1. The light reflected from the steel wire rope is transmitted to the beam splitter 5 through the prisms 1. The beam splitter 5 splits the light beams and projects them onto the event camera 2 and the conventional camera 3 through the corresponding prisms 1, so that the event camera 2 and the conventional camera 3 have the same optical field of view and perspective relationship. The synchronization triggering device 4 is a hardware-level synchronization and triggering unit based on FPGA. The FPGA generates a high-precision master clock to provide a unified time reference for the conventional camera 3 and the event camera 2, and timestamps the event stream and frame images at the nanosecond level. When the accumulated events exceed a preset threshold, the FPGA sends a trigger signal to the conventional camera 3 to control the exposure.

[0080] The initial boundary extraction module is used to preprocess the frame image and extract the boundary to obtain the initial left and right boundary coordinates of the wire rope.

[0081] The dynamic motion estimation module estimates the transient displacement function of the wire rope boundary during the exposure cycle based on the event flow, as well as the displacement correction coordinates of the left and right boundaries;

[0082] The dynamic correction and compensation module is used to perform geometric compensation and boundary correction on the initial left and right boundary coordinates based on the transient displacement function and displacement correction coordinates, so as to obtain the equivalent static left and right boundary coordinates.

[0083] The diameter calculation module calculates the diameter of the steel wire rope in the pixel domain based on the left and right boundary coordinates of the equivalent static position, and then converts the actual physical diameter of the steel wire rope by combining the calibration ratio.

[0084] In conjunction with the above system, this embodiment also proposes a method for measuring the diameter of a wire rope based on event compensation correction, including the following steps:

[0085] Step 1: Synchronous Data Acquisition: The traditional camera 3 and the event camera 2 are controlled to acquire data synchronously within an exposure window via the synchronization triggering device 4. Get frame images of the wire rope With event flow And upload it to the computer;

[0086] Specifically, this allows a traditional camera to operate within a single exposure window. The frame images captured internally are denoted as The pixel value of each point on the image is denoted as ,in The initial exposure time, For the exposure time, Indicates the pixel position in the image. Represents the horizontal coordinates of a pixel in the image. Represents the coordinates of a pixel in the vertical direction of the image, indicated by the superscript. This indicates the transpose operation.

[0087] Event Camera 2 in the exposure window An event stream sequence was collected internally. ,Include The event will be the first Each event is denoted as , ,in Indicates the first The location of each event in the event camera pixel array Indicates the first The horizontal coordinates of an event in the event camera pixel array. Indicates the first The vertical coordinates of an event in the event camera pixel array. Indicates the first The timestamp when the event was triggered Indicates the first The polarity of brightness changes in each event. This indicates that the brightness at that pixel location has changed positively (becomes brighter). This indicates that the brightness at that pixel location has changed negatively (becomes darker).

[0088] The captured frame images With event flow It is uploaded to the computer via a high-speed interface.

[0089] Step 2: Extract initial boundaries: Preprocess the frame image, extract the set of boundary points of the wire rope, estimate the axial direction vector and normal direction vector of the wire rope, scan along the normal to obtain the gray-level gradient, and use the gray-level gradient maximum point as the initial left and right boundary coordinates;

[0090] Step 2: Extract the initial left and right boundary coordinates , The specific process is as follows:

[0091] In frame image The preprocessed image is obtained after preprocessing. The boundary point set is obtained through boundary extraction. The steel wire rope appears as a long strip in the image, with its center line... The direction vector, i.e., the axial direction vector, can be obtained through... Principal eigenvector estimation, let the set of boundary points If the number of points is N, then the mean coordinates of the boundary points are... Covariance Matrix They can be represented as follows:

[0092]

[0093]

[0094] Among them, mean coordinates This is also the coordinate of the centroid of the wire rope; this point must be located on the center line.

[0095] right By performing eigenvalue decomposition, we can obtain:

[0096] ,

[0097] in, This represents the i-th eigenvector. This represents the i-th eigenvalue;

[0098] Take the eigenvector corresponding to the largest eigenvalue As the axial direction vector ,Right now:

[0099] ;

[0100] With axial direction vector The perpendicular direction vector is denoted as the normal direction vector. ,Right now:

[0101] ;

[0102] Take the center line A reference point coordinate The coordinates of the mean point can usually be taken. For any point coordinate in the image coordinate system Its projected coordinates along the axis in the image coordinate system Projected coordinates along the normal direction It can be represented as:

[0103] ,

[0104] ;

[0105] Then the center line of the wire rope any point corresponding to the above It can be represented as:

[0106]

[0107] Next, on each s, along the normal direction Perform a scan to obtain the normal offset coordinates. ,Right now:

[0108]

[0109] Then, a one-dimensional grayscale profile along the normal direction It can be represented as:

[0110] ,

[0111] in, Indicates coordinates The pixel grayscale value at that location;

[0112] right The gray-level gradient of the profile along the normal direction can be obtained by taking the difference. ,Right now:

[0113] ,

[0114] The initial boundary of the wire rope corresponds to the gray-scale gradient of the cross-section. Let the maximum point be denoted as the normal coordinates of the initial left and right boundaries, respectively. and Then we have:

[0115]

[0116] ;

[0117] in, Indicates the maximum search radius;

[0118] Then the initial left and right boundary coordinates of the wire rope , They can be represented as:

[0119]

[0120] .

[0121] Step 3: Estimate transient motion information: Project the event stream onto the image coordinate system, filter the event set that satisfies the time and space constraints in the boundary neighborhood, construct the transient displacement function of the wire rope boundary based on the event set, and calculate the displacement correction coordinates of the left and right boundaries;

[0122] The specific conditions for filtering the boundary neighborhood event set are: retaining events whose minimum distance from the event point to the initial boundary point set does not exceed the preset boundary neighborhood radius, and whose event trigger time is within the exposure window;

[0123] Specifically, this includes utilizing the event stream obtained through the steps. Calculate the exposure window Inside, the boundary is in the axial position Transient displacement function in the normal direction at time t Displacement correction coordinates of the left and right boundaries , ;

[0124] After projecting the event onto the image coordinate system, let the first... One event Event points in to the initial set of boundary points The minimum distance is ,Right now:

[0125]

[0126] in, Represents the initial boundary set Traverse all boundary points in the array;

[0127] Preserving the boundary neighborhood event set , Both time and space constraints must be satisfied, that is:

[0128]

[0129] in, The radius of the boundary neighborhood;

[0130] The main geometric disturbance of the wire rope within the exposure window can often be approximated as a displacement along the normal direction, hence:

[0131]

[0132]

[0133] ;

[0134] in, Indicates the axial position of the left and right boundaries of the wire rope time At that time, relative to the initial left and right boundary coordinates , The transient displacement function;

[0135] Indicates according to estimate The optimization function;

[0136] and These represent the axial positions of the left and right boundaries of the wire rope, respectively. time Corrected coordinates at time;

[0137] When the wire rope undergoes periodic vibration during the exposure time, a transient displacement function is constructed. The process is as follows:

[0138] The axial coordinate of the wire rope is divided into multiple segment intervals. The displacement in each segment interval is represented by a first-order harmonic model. The fundamental frequency of the harmonic model is given by the prior mechanical parameters.

[0139] Collect the event set within each segment interval, construct the displacement observation residual for each event, and perform robust minimization of the displacement observation residual based on the Huber loss function to obtain the harmonic coefficients of each segment interval.

[0140] By splicing the displacement models of each segmented interval, the global transient displacement function is obtained.

[0141] Step 4: Correct the boundary coordinates: Using the exposure midpoint time as a reference time, calculate the average boundary displacement using the transient displacement function, and use the average displacement to perform geometric compensation on the initial left and right boundary coordinates to obtain the equivalent static left and right boundary coordinates.

[0142] Using the transient displacement function estimated in step three and displacement correction coordinates , For the initial left and right boundary coordinates , Geometric compensation and boundary correction are performed to obtain the equivalent static left and right boundary coordinates. and .

[0143] Select the midpoint of the exposure time. As a reference moment within the exposure window, let the boundary be... The displacement is ,Right now:

[0144]

[0145] Let the average displacement of the boundary in the event flow e within the exposure window be ),Right now:

[0146]

[0147] Then obtained using event estimation For the initial left and right boundary coordinates and Make corrections:

[0148]

[0149]

[0150] in, and These are the corrected equivalent static left and right boundary coordinates, respectively.

[0151] Step 5: Calculate the actual diameter: Perform axial discrete sampling on the left and right boundary coordinates of the equivalent stationary point, calculate the pixel domain diameter sequence, and after obtaining the median and average value through sliding window statistics, combine it with the calibration ratio to obtain the actual physical diameter of the wire rope.

[0152] The specific process of sliding window statistics is as follows:

[0153] A sliding window process is applied to the pixel field diameter sequence, and the median of the diameter values ​​within each sliding window is taken as the final pixel field diameter within the window.

[0154] The median of all windows is averaged to obtain the final pixel field diameter measurement.

[0155] Using the equivalent static boundary coordinates obtained in step four and Calculate the diameter of the steel wire rope in the pixel domain. The actual physical diameter of the wire rope is calculated based on the calibration ratio. .

[0156] Will Discretized Let the coordinates of a discrete point be... Indicates the first Each discrete axial sampling position is substituted into and The discrete boundary axial coordinates are obtained from the calculation formula. and The diameter sequence is obtained through geometric calculation. ,Right now:

[0157]

[0158] ,

[0159]

[0160]

[0161] ;

[0162] in, and This indicates the axial start and end positions of the wire rope in the current image. This represents the corrected set of boundary points, i.e. , Represents the modified boundary set Iterate through all boundary points. Indicates the axial sampling step size. This represents the total number of discrete sampling points.

[0163] To enhance robustness, Perform sliding window statistics and take the median as the final pixel diameter measurement value within the window. ,Right now:

[0164]

[0165] in, This represents median operations. Indicates axial position The corrected diameter value at that location; This represents the index variable within the sliding window; This represents the radius of the sliding window.

[0166] right The average value is taken as the final pixel domain diameter measurement, that is:

[0167]

[0168] If planar calibration is used and the distance from the wire rope to the camera is approximately constant, then the actual physical diameter of the wire rope is... It can be represented as:

[0169] ,

[0170] in, The calibration scale used is defined as 1 pixel in the image corresponding to c millimeters in the physical world.

[0171] This embodiment specifically adopts the following optical path structure:

[0172] The entire imaging system adopts a "dual-camera coaxial heterogeneous" optical path structure, such as Figure 2 As shown, spatial registration is achieved by ensuring that the conventional camera and the event camera observe the steel cable from the same physical viewpoint. The imaging system includes a beam splitter 5, three prisms 1, a conventional camera 3, and an event camera 2.

[0173] The optical path transmission process of the optical system is as follows: First, the light reflected from the steel wire rope reaches the beam splitter 5 through the first prism; the beam splitter 5 projects the same beam of light from the steel wire rope simultaneously and along the same path into two different directions, forming two beams; the conventional camera 3 and the event camera 2 are respectively mounted perpendicularly at the two output optical path ports of the beam splitter 1, one beam is imaged onto the event camera 2 through the second prism, and the other beam is imaged onto the conventional camera 3 through the third prism. This optical path structure ensures that the two cameras have exactly the same optical field of view and perspective relationship, avoiding the registration problem caused by parallax.

[0174] To ensure temporal alignment between the acquired event stream and frame images, a hardware-level synchronization and triggering unit based on FPGA was designed. The FPGA generates a high-precision master clock, providing a unified time reference for both cameras and assigning precise nanosecond-level timestamps to each event and frame image. When the accumulated events caused by the movement of the wire rope edge exceed a preset threshold, the FPGA immediately sends a trigger signal to the conventional camera, controlling it to accurately expose in the next second, achieving adaptive synchronous acquisition.

[0175] Regarding the transient displacement function in step three of the system operation steps in the technical solution If the wire rope undergoes periodic vibration during the exposure time, then Choose one of the following methods:

[0176] axial coordinates Divided into Q segments, that is:

[0177] ;

[0178] in, This represents the q-th axial segmentation interval; Let be the axial coordinate of the q-th segment boundary; then in Internal displacement Approximation is over time The first-order harmonic model, namely:

[0179]

[0180] in, , , Harmonic coefficients The fundamental frequency is given by prior mechanical parameters.

[0181] Because when the vibration occurs, the first One event Event points in The most densely packed area is at the border. event collection within ,Right now:

[0182]

[0183] ;

[0184] in, for Projected coordinates along the axial direction.

[0185] Then construct the first Displacement observation residuals of each event :

[0186] ,

[0187] ,

[0188] ;

[0189] in, express Projected coordinates along the normal direction, Indicates the axial coordinate as The initial normal coordinates at that location, This represents the grayscale gradient of the normal profile, specifically in conjunction with the processing steps in step one. ) indicates in Displacement at time.

[0190] Then the event set Displacement observation residuals within Perform robust minimization, i.e.:

[0191]

[0192] ;

[0193] in, Here is the Huber loss function. This is the Huber inflection threshold.

[0194] Finally, all of them By splicing the images together, the global result can be obtained. .

[0195] The above specific embodiments are merely optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant teachings of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A wire rope diameter measurement system based on event compensation correction, characterized in that, It includes an imaging module, a boundary initial extraction module, a dynamic motion estimation module, a dynamic correction and compensation module, and a diameter calculation module; The imaging module includes a conventional camera (3), an event camera (2), optical components, a synchronization triggering device (4), and a beam splitter (5). The conventional camera (3) is used to acquire frame images of the wire rope, and the event camera (2) is used to acquire event streams generated by the movement of the wire rope's edge. The optical components and the beam splitter (5) work together to achieve optical path transmission and beam splitting. The synchronization triggering device (4) is used to ensure that the conventional camera (3) and the event camera (2) acquire images synchronously. The initial boundary extraction module is used to preprocess the frame image and extract the boundary to obtain the initial left and right boundary coordinates of the wire rope. The dynamic motion estimation module estimates the transient displacement function of the wire rope boundary during the exposure cycle based on the event flow, as well as the displacement correction coordinates of the left and right boundaries; The dynamic correction and compensation module is used to perform geometric compensation and boundary correction on the initial left and right boundary coordinates according to the transient displacement function and displacement correction coordinates, so as to obtain the equivalent static left and right boundary coordinates. The diameter calculation module calculates the diameter of the steel wire rope in the pixel domain based on the equivalent static left and right boundary coordinates, and then converts it to the actual physical diameter of the steel wire rope by combining the calibration ratio.

2. The wire rope diameter measurement system based on event compensation correction according to claim 1, characterized in that, The imaging module adopts a dual-camera coaxial heterogeneous optical path structure, including a beam splitter (5) and at least three prisms (1). The light reflected by the steel wire rope is transmitted to the beam splitter (5) through the prism (1). The beam splitter (5) splits the light beam and projects it to the event camera (2) and the conventional camera (3) through the corresponding prisms (1), so that the event camera (2) and the conventional camera (3) have the same optical field of view and perspective relationship.

3. The wire rope diameter measurement system based on event compensation correction according to claim 1, characterized in that, The synchronization triggering device (4) is a hardware-level synchronization and triggering unit based on FPGA. The FPGA generates a high-precision master clock to provide a unified time reference for the traditional camera (3) and the event camera (2), and stamps the event stream and frame images with nanosecond-level timestamps. When the accumulated events exceed the preset threshold, the FPGA sends a trigger signal to the traditional camera (3) to control the exposure.

4. A method for measuring the diameter of a steel wire rope based on event compensation correction, characterized in that, The system applied to any one of claims 1-3 includes the following steps: Step 1: Synchronous data acquisition: The traditional camera (3) and the event camera (2) are controlled to acquire data synchronously through the synchronous triggering device (4). The frame image and event stream of the wire rope are obtained within an exposure window and uploaded to the computer. Step 2: Extract initial boundaries: Preprocess the frame image, extract the set of boundary points of the wire rope, estimate the axial direction vector and normal direction vector of the wire rope, scan along the normal to obtain the gray-level gradient, and use the gray-level gradient maximum point as the initial left and right boundary coordinates; Step 3: Estimate transient motion information: Project the event stream onto the image coordinate system, filter the event set that satisfies the time and space constraints in the boundary neighborhood, construct the transient displacement function of the wire rope boundary based on the event set, and calculate the displacement correction coordinates of the left and right boundaries; Step 4: Correct the boundary coordinates: Using the exposure midpoint time as a reference time, calculate the average boundary displacement using the transient displacement function, and use the average displacement to perform geometric compensation on the initial left and right boundary coordinates to obtain the equivalent static left and right boundary coordinates. Step 5: Calculate the actual diameter: Perform axial discrete sampling on the left and right boundary coordinates of the equivalent stationary point, calculate the pixel domain diameter sequence, and after obtaining the median and average value through sliding window statistics, combine it with the calibration ratio to obtain the actual physical diameter of the wire rope.

5. The method for measuring the diameter of a wire rope based on event compensation correction according to claim 4, characterized in that, The specific process for extracting the initial left and right boundary coordinates in step two is as follows: The preprocessed image is obtained after preprocessing frame image I. The initial set of boundary points is obtained through boundary extraction. Calculate the set of boundary points Given the mean coordinates and covariance matrix, perform eigenvalue decomposition on the covariance matrix and take the eigenvector corresponding to the largest eigenvalue as the axial direction vector. : That is: axial direction vector in, This represents the eigenvector corresponding to the largest eigenvalue; This represents the i-th eigenvalue; The vector perpendicular to the axial direction vector is taken as the normal direction vector. : That is: normal direction vector ; Project the coordinates of any point in the image coordinate system onto the axial and normal directions, scan along the normal direction at each position of the axial direction to obtain a one-dimensional gray-scale profile, and calculate the differential gradient of the gray-scale profile. Among them, the one-dimensional grayscale profile Represented as: ； Represents the coordinates of any point in the image coordinate system; express The projected coordinates; and Representing coordinates Axial coordinates projected along the axial direction and normal coordinates projected along the normal direction; Indicates coordinates The pixel grayscale value at that location; right The gray-level gradient of the profile along the normal direction can be obtained by taking the difference. ,Right now: The maximum point of the gray-scale gradient is the initial boundary of the wire rope, which corresponds to the normal coordinates of the initial left and right boundaries, and then the coordinates of the initial left and right boundaries can be calculated. Specifically, the initial boundary of the wire rope corresponds to the gray-scale gradient of the cross-section. Let the maximum point be denoted as the normal coordinates of the initial left and right boundaries, respectively. and Then we have: ； in, Indicates the maximum search radius; Then the initial left and right boundary coordinates of the wire rope , They can be represented as: 。 6. The method for measuring the diameter of a wire rope based on event compensation correction according to claim 4, characterized in that, The specific conditions for filtering the boundary neighborhood event set in step three are: retaining events whose minimum distance from the event point to the initial boundary point set does not exceed the preset boundary neighborhood radius, and whose event trigger time is within the exposure window; Specifically, this includes utilizing the event stream obtained through the steps. Calculate the exposure window Inside, the boundary is in the axial position transient displacement function in the normal direction at time t Displacement correction coordinates of the left and right boundaries , ;in, This is the initial exposure time; Exposure time; After projecting the event onto the image coordinate system, let the first... One event Event points in to the initial set of boundary points The minimum distance is ,Right now: in, Indicates the first The coordinates of an event in the event camera pixel array; Indicates the first The timestamp at which the event was triggered; Indicates the first The polarity of brightness change for each event; b indicates the polarity of brightness change for each event. Traversing all boundary points; Preserving the boundary neighborhood event set , Both time and space constraints must be satisfied, that is: in, The radius of the boundary neighborhood; The main geometric disturbance of the wire rope within the exposure window can often be approximated as a displacement along the normal direction, hence: ； in, Indicates the axial position of the left and right boundaries of the wire rope time At that time, relative to the initial left and right boundary coordinates , The transient displacement function; Indicates according to estimate The optimization function; and These represent the axial positions of the left and right boundaries of the wire rope, respectively. time Corrected coordinates at that time.

7. The method for measuring the diameter of a wire rope based on event compensation correction according to claim 4, characterized in that, When the wire rope undergoes periodic vibration during the exposure time, the transient displacement function is constructed in step three. The process is as follows: The axial coordinate of the wire rope is divided into multiple segment intervals. The displacement in each segment interval is represented by a first-order harmonic model. The fundamental frequency of the harmonic model is given by prior mechanical parameters. Collect the event set within each segment interval, construct the displacement observation residual for each event, and perform robust minimization of the displacement observation residual based on the Huber loss function to obtain the harmonic coefficients of each segment interval. By splicing the displacement models of each segmented interval, the global transient displacement function is obtained.

8. The method for measuring the diameter of a wire rope based on event compensation correction according to claim 4, characterized in that, The specific process of the sliding window statistics in step five is as follows: A sliding window process is applied to the pixel domain diameter sequence, and the median of the diameter values ​​within each sliding window is taken as the final pixel domain diameter within the window. The median of all windows is averaged to obtain the final pixel domain diameter measurement.

Images