Workpiece eccentricity detection method based on stroboscopic illumination and intelligent sensor

Abstract

The invention discloses a workpiece eccentricity detection method based on stroboscopic lighting and an intelligent sensor, and relates to the field of stroboscopic lighting and intelligent sensing, and the workpiece eccentricity detection method comprises the following steps: fixing a workpiece at the end part of a rotating shaft through an elastically deformable buffer material layer, so that the workpiece can generate displacement in the radial direction under the action of centrifugal force; arranging a plurality of groups of mark point pairs which are paired relative to a preset rotation center on the to-be-detected surface, and defining different orientations by connecting lines of the mark point pairs; controlling the stroboscopic lighting unit to irradiate the workpiece at a stroboscopic frequency added with the frequency offset; and controlling the intelligent image sensor to perform frame-by-frame exposure under the trigger rhythm synchronous with the stroboscopic illumination unit, performing circular trajectory fitting on each group of mark point pairs in the mark point ghosting image, calculating the center offset of the circular trajectory, and taking the direction of the mark point pair with the maximum offset as the center-of-gravity shift direction of the workpiece. The problem that workpiece eccentricity is difficult to detect is solved by combining stroboflash and intelligent sensing, and online detection can be achieved through the Internet of Things and the Internet.

Description

Technical Field

[0001] This invention relates to the fields of strobe lighting and intelligent sensing, and more specifically, to a method for detecting workpiece eccentricity based on strobe lighting and intelligent sensors. Background Technology

[0002] In rotary machining and assembly, the roundness and straightness of a workpiece are usually assessed through geometric dimensions and form and position tolerances. However, workpiece eccentricity often stems from deviations in mass distribution from the theoretical center of rotation, rather than simply a matter of non-roundness or non-straightness. Many workpieces, when inspected with calipers or dial indicators in a stationary state, do not readily reveal slight shifts in their center of gravity relative to the axis of rotation. These shifts may appear minor in geometric contours, but they can be amplified into significant centrifugal imbalances during high-speed rotation, causing vibration, noise, and premature bearing failure, adversely affecting machine tool accuracy and equipment lifespan. Especially after subsequent machining processes such as turning, milling, and drilling, localized material removal or residual material alters the original mass distribution, causing a shift in the workpiece's center of gravity from the designed theoretical center of rotation. This type of eccentricity is often difficult to detect visually or through simple geometric measurements. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a workpiece eccentricity detection method based on strobe lighting and intelligent sensors, so as to solve the problems mentioned in the background art. To achieve the above objectives, the present invention adopts the following technical solution: A method for detecting workpiece eccentricity based on strobe illumination and smart sensors includes the following steps: 1) The workpiece is fixed to the end of the rotating shaft through a layer of elastically deformable buffer material, so that the workpiece can be displaced in the radial direction relative to the rotating shaft when subjected to centrifugal force; 2) Set at least three sets of marker point pairs on the surface to be inspected of the workpiece. Each set of marker point pairs includes a first marker point and a second marker point arranged in pairs about a preset rotation center. The line connecting each set of marker point pairs defines an orientation direction. 3) Drive the rotating shaft to rotate the workpiece at the target speed, obtain the actual speed of the rotating shaft, calculate the baseline frequency f0 corresponding to the actual speed, and control the strobe illumination unit to illuminate the workpiece at a strobe frequency of f0 plus the frequency offset Δf to form a marker point afterimage with a slow rotation effect. 4) Control the intelligent image sensor to expose frame by frame according to the triggering beat synchronized with the strobe illumination unit. The intelligent image sensor includes an adaptive exposure control module. The adaptive exposure control module adjusts the exposure time according to the strobe frequency and the frequency offset to obtain a sequence of marker point afterimage images. 5) Perform image recognition, including extracting the pixel coordinate sequence of each first marker point and the pixel coordinate sequence of the corresponding second marker point from the marker point afterimage image sequence; 6) Perform image processing, including performing circular trajectory fitting on the pixel coordinate sequence of each pair of marker points to obtain the circular trajectory of the first marker point and the circular trajectory of the second marker point, as well as the center offset of the two circular trajectories; select the target marker point pair with the largest center offset of the circular trajectory among all the marker point pairs, and take the orientation direction of the target marker point pair as the workpiece center of gravity offset direction.

[0004] Preferably, the preset rotation center is the projection point of the theoretical axis of the rotation axis on the surface to be tested. The first and second marker points of each pair of marker points are respectively arranged on both sides of the preset rotation center along the corresponding orientation direction, and the nominal distances of the first and second marker points to the preset rotation center are equal.

[0005] Preferably, the reference line frequency f0 is an integer multiple of the mechanical rotation frequency corresponding to the actual rotation speed, and the frequency offset Δf satisfies 0 < |Δf| ≤ 0.05·f0. The adaptive exposure control module sets the exposure time to a fixed proportion of one strobe cycle length, and reduces the analog gain when the average gray level of the afterimage image brightness statistics result is close to a preset saturation threshold, so that the apparent rotational angular velocity of the marker point afterimage is lower than five percent of the mechanical rotation frequency and the gray level is kept within a preset dynamic range.

[0006] Preferably, extracting the pixel coordinate sequence of each first marker point and the pixel coordinate sequence of the corresponding second marker point from the marker point afterimage image sequence includes: performing grayscale thresholding on each frame of the afterimage image in the marker point afterimage image sequence to obtain candidate marker point regions; performing connected component labeling on the candidate marker point regions to obtain the connected components corresponding to each first marker point and each second marker point; and performing subpixel centroid localization on each connected component to obtain the subpixel precision pixel coordinates of each first marker point and each second marker point in each frame of the afterimage image.

[0007] Preferably, performing circular trajectory fitting on the pixel coordinate sequence of each pair of marker points includes: for each pair of marker points, substituting the pixel coordinate sequence of the corresponding first marker point into a preset least squares circle fitting algorithm to solve for the center coordinates and radius of the circular trajectory of the first marker point; substituting the pixel coordinate sequence of the corresponding second marker point into the least squares circle fitting algorithm to solve for the center coordinates and radius of the circular trajectory of the second marker point; and using the Euclidean distance between the center coordinates of the two as the center offset of the circular trajectory.

[0008] Preferably, after selecting the target marker point pair, the method further includes: comparing the center offset of the circular trajectory of the target marker point pair with a pre-calibrated correspondence curve, wherein the correspondence curve provides a functional relationship between the center offset of the circular trajectory and the center offset of the workpiece, and determining the numerical index of the center offset of the workpiece based on the position of the center offset of the circular trajectory of the target marker point pair in the correspondence curve.

[0009] Preferably, the buffer material layer is an annular elastomer gasket disposed between the workpiece end face and the connecting fixture, and the radial stiffness of the annular elastomer gasket is between 1×10³ N / mm and 1×10 5 The values ​​are between N / mm, and the axial stiffness of the annular elastomer gasket is higher than the radial stiffness.

[0010] Preferably, the first and second markers in the marker pair adopt a graphic structure that can form a high-contrast shadow under the illumination of the strobe lighting unit, wherein the graphic structure is a solid protrusion or solid groove with sharp edges.

[0011] Preferably, the adaptive exposure control module adjusts the exposure time as follows: the current strobe operating frequency f1 is obtained by adding the strobe frequency and the frequency offset, the current strobe cycle length T is defined as 1 / f1, and the exposure time is set to k·T, where the value of k ranges from 0.1 to 0.5.

[0012] Preferably, the method is also used to perform online inspection of multiple workpieces that are being inspected or rotating on the production line. Each workpiece is assigned a workpiece identifier, and the obtained inspection results are combined with the corresponding workpiece identifier to form inspection data. This data is then sent to a remote monitoring server via the Industrial Internet of Things or the Internet. The remote monitoring server centrally receives and stores the inspection data from multiple workstations.

[0013] The advantage of this invention over existing technologies lies in its use of a buffer material layer with specific radial stiffness at the workpiece end. This allows for observable radial displacement of the workpiece under centrifugal force during rotation, transforming the previously difficult-to-quantify center-of-gravity shift into a geometric offset of the marked point trajectory on the end face. Multiple pairs of marked points symmetrical about a preset rotation center are arranged on the surface to be inspected. This ensures that the circular trajectories of each marked point have a unified geometric reference in the unbiased state. Once the center of gravity deviates, the change in the distance between the centers of the marked point pairs along the eccentric direction is most significant. Therefore, by comparing the center offset of the circular trajectories of marked point pairs in different directions, the direction of the workpiece's center-of-gravity shift can be directly determined. The identification results are intuitive and can be used to directly guide the placement of counterweights or material removal. This invention calculates the baseline frequency by acquiring the actual rotational speed of the rotating shaft, and sets the operating frequency of the stroboscopic illumination unit to the sum of the baseline frequency and the frequency offset. This causes the marker point to exhibit a slowly rotating afterimage in the field of view. Simultaneously, the adaptive exposure control module sets the exposure time to a fixed proportion of the stroboscopic cycle length based on the stroboscopic frequency and the frequency offset. This allows the intelligent image sensor to expose frame by frame under a trigger rhythm strictly synchronized with the stroboscopic illumination unit, ensuring that each frame corresponds to only a single stroboscopic event. This avoids cross-cycle ghosting while maintaining the brightness and outline clarity of the marker point, thereby obtaining a stable afterimage image sequence. By combining grayscale thresholding, connected component labeling, and sub-pixel centroid positioning, a high-precision sequence of marker point pixel coordinates can be extracted. After least-squares circle fitting, the center coordinates of the first and second marker point circular trajectories are obtained. The center offset of the circular trajectory is obtained through the center-to-center distance, and the centroid offset value is obtained by comparing it with a pre-calibrated correspondence curve. This invention not only provides the offset direction but also the offset magnitude. Repeated testing and consistency verification at different target speeds can reversely verify whether the stiffness of the buffer material layer is stable within the allowable range, improving the reliability and engineering applicability of the entire testing method within the actual production speed range. The markers employ sharp-edged protrusions or grooves that create high-contrast shadows under strobe illumination, further improving sub-pixel positioning accuracy and reducing the impact of complex backgrounds and reflection conditions on testing accuracy. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall process of the workpiece eccentricity detection method based on strobe lighting and intelligent sensors of the present invention; Figure 2 This is a schematic diagram of the workpiece eccentricity detection system based on strobe lighting and intelligent sensors according to the present invention; Figure 3 This is a schematic diagram of the marker point afterimage image acquisition and image processing flow of the present invention; Figure 4 This is a schematic diagram of the circular trajectory fitting and workpiece center of gravity offset analysis process of the present invention. Detailed Implementation

[0015] The specific embodiments of the present invention will now be described with reference to the accompanying drawings.

[0016] This invention is based on the following physical principle: Imagine a workpiece as a disk that should fit perfectly on its axis. If the workpiece's mass distribution is uneven, its center of gravity will deviate from the theoretical center of rotation, causing the entire object to sway slightly towards the heavier side during rotation. This invention places a pair of markers, arranged in pairs about the center of rotation, along several different directions on the workpiece's end face. When the workpiece actually rotates and becomes eccentric, the actual distances from the pair of markers along the eccentric direction to the actual axis of rotation are no longer perfectly equal, resulting in the most significant difference in the diameter of their corresponding circular tracks. In other directions inconsistent with the eccentricity, this diameter difference is much smaller. Because the rotation is rapid, it is difficult to identify the circular tracks of the markers. Therefore, this invention introduces strobe illumination and a smart image sensor to freeze the circular tracks of these markers into clear afterimages. By comparing the diameter differences of the circular tracks of the pairs of markers in each direction, the pair with the largest diameter difference is identified, which is equivalent to finding the direction of the workpiece's center of gravity deviation, and the degree of deviation can be determined accordingly.

[0017] In a specific embodiment, as shown in Figures 1 and 2, the entire process is completed when the workpiece rotates at a low to medium speed. The target rotational speed is typically controlled between 300 rpm and 3000 rpm, with a typical detection speed between 600 rpm and 1800 rpm. This range generates sufficient centrifugal force to produce a clearly observable radial displacement of the workpiece while completely avoiding the safety risks, vibration noise, and measurement difficulties associated with high-speed rotation. The displacement amplitude is typically between 0.05 mm and 3 mm, specifically determined by the workpiece mass, eccentricity, rotational speed, and the radial stiffness of the buffer layer. In the image, this manifests as a clear separation with a center offset ranging from 0.2 mm to 10 mm, allowing for reliable extraction without requiring sub-pixel precision. Even ordinary industrial cameras can clearly distinguish the displacement, greatly improving physical intuition and practical operability.

[0018] In some embodiments, the workpiece is fixed to the end of the rotating shaft by a specially designed buffer material layer to ensure that the workpiece can be displaced. This buffer material layer uses an annular elastomer gasket, which is placed between the workpiece end face and the connecting fixture to form a complete contact surface. The material is preferably a high-damping polyurethane or silicone rubber composite, and the radial stiffness is preferably selected in the lower limit of 1×10³ N / mm to 5×10³ N / mm. This allows for a significant radial displacement of 0.1-2mm with a typical 20-50kg workpiece and a centrifugal force of 20-500N caused by eccentricity. The axial stiffness is still maintained at 3-5 times higher than the radial stiffness to ensure that the workpiece does not move axially or tilt. The gasket thickness is generally 4-10mm, with the inner diameter precisely matching the shaft and the outer diameter covering more than 80% of the workpiece end face area. This choice of lower radial stiffness is one of the core design considerations: the centrifugal force causes the entire workpiece to produce a macroscopically measurable rigid translation, thereby causing a visible separation of the center of the marked point trajectory. In actual testing, this translation amount is proportional to the centrifugal force and inversely proportional to the radial stiffness. By operating at the lower limit of stiffness, the displacement is amplified to the millimeter level, which is extremely intuitive in physics—it is equivalent to letting the workpiece, under elastic support, throw its center of gravity to a point on the rotation axis. The greater the eccentricity, the higher the rotation speed, and the lower the stiffness, the more obvious the translation will be.

[0019] At least three sets of marker pairs are arranged on the surface to be inspected, with four to eight sets evenly distributed recommended. Each set of marker pairs includes a first marker and a second marker, which are strictly symmetrically arranged about the theoretical center of rotation (i.e., the projection of the axis of rotation onto the end face). The nominal distances to the theoretical center of rotation are exactly equal, typically ranging from 15mm to 80mm. The greater the distance, the more noticeable the center offset in the image, and the higher the measurement sensitivity. A line is drawn through the theoretical center of rotation to define the orientation direction of the set. Multiple orientation directions are evenly distributed across the entire circumference, such as every 45° or 30°. The markers are solid raised cylinders or solid grooves with a diameter of 3-8mm and a height / depth of 1-3mm. The sidewall perpendicularity is better than 3′, which produces extremely sharp shadow edges under strobe illumination, achieving a contrast ratio of over 200:1, almost unaffected by ambient light. In actual processing, this can be achieved through CNC milling or 3D printing of metal inserts, which is inexpensive and durable.

[0020] After the detection is initiated, the motor drives the workpiece to the target speed of 300-3000 rpm. The actual rotational speed n is collected in real time, and the mechanical frequency fr = n / 60 is calculated. The baseline frequency f0 is set to 10-30 times fr to ensure sufficient strobe pulses for dense sampling within one mechanical cycle. The strobe frequency is set to f1 = f0 + Δf, with the absolute value of Δf controlled within 0.02 × f0, typically 0.005-0.015 × f0. As a result, the apparent rotational angular velocity of the afterimage is only 1-3% of the mechanical frequency, and the image sequence presents an approximately static, slowly drifting effect. Even at 600 rpm and a frequency of 10 Hz, the apparent rotation cycle can be as long as 20-30 seconds, facilitating stable long-term acquisition without motion blur.

[0021] The intelligent image sensor employs a high-frame-rate industrial CMOS sensor (200-500fps), providing complete field of view coverage of the area to be inspected. The lens magnification is selected from 0.05 to 0.2x, ensuring that each marker occupies 50-200 pixels in the image. The sensor incorporates an adaptive exposure control module, precisely synchronized with the strobe controller via hardware triggering, ensuring strict alignment of each frame's exposure with the pulse center. The initial exposure time is set to 0.2-0.4 times the current period T = 1 / f1. The module continuously calculates the average grayscale G and saturated pixel percentage P of the previous period's image: if G is below the lower limit of 70, the exposure is extended in steps of 0.05T; if the upper limit is reached but the image is still dark, the analog gain is increased by 1.2-1.5 times; if G is above the upper limit of 170 or P > 2%, the exposure is shortened, and the gain is reduced if necessary. After 4-10 cycles of rapid convergence, the gradient of the shadow edge of the marker point stabilizes at 80-150 gray values / pixel, with a saturation ratio of <0.5%, resulting in extremely high image quality. The outline of the marker point is so sharp that it can be accurately extracted even with a displacement of 2-3mm.

[0022] As shown in Figure 3, in the image processing stage, for each frame of the marker point afterimage image, a segmentation threshold is first adaptively determined based on the gray-level histogram of that frame. Pixels with gray levels higher than the threshold are classified as foreground, and pixels with gray levels lower than the threshold are classified as background, thereby extracting candidate regions containing marker points. Subsequently, eight-neighbor connected component labeling is performed on the binarized result, dividing continuously connected foreground pixels into several independent connected regions. Based on the geometric features such as area and aspect ratio of each connected region, noise spots that are too small or have abnormal shapes are filtered out, retaining only connected regions that meet the size and shape requirements of the marker points. For each retained connected region of the marker point, the position and gray-level value of all pixels within the region are statistically analyzed. Pixels with higher gray-level values ​​are given greater weight, and the brightness centroid of the region is calculated, refining the position of the marker point from an integer pixel grid to between pixel grids. Using this brightness-weighted centroid method, under the condition that the marker point edges are sufficiently sharp and the signal-to-noise ratio is high, the position repetition accuracy can typically reach approximately 0.02 to 0.05 pixels. By repeating the above processing procedure on the continuously acquired 300 to 800 frames of afterimage images, the continuous coordinate sequence of each marker point in the entire detection process can be obtained. Even if the radial displacement of the workpiece in the actual space reaches the millimeter level, the motion trajectory of each marker point in the pixel coordinate system is still approximately circular. Only the center of the circle has an observable offset relative to the theoretical rotation center. This offset of the center of the circle is the core geometric feature used to characterize the direction and amount of the workpiece's center of gravity offset.

[0023] As shown in Figure 4, the circular trajectory fitting uses an improved least squares algorithm. Considering the slight ellipticity that may exist in reality, elliptic constraints are added, or algebraic closed-form solutions are used directly to obtain the precise center (a, b) and radius R of each marker point's trajectory. The Euclidean distance between the center coordinates of the first and second marker points is the center offset of that group. Because the workpiece as a whole undergoes macroscopic radial movement driven by centrifugal force, the centers of the two marker points facing the center of gravity offset will separate outwards. The maximum separation occurs in the direction of center of gravity offset, typically 0.5-15 mm, which can reach 50-500 pixels in the image. The greater the center of gravity offset, the higher the square of the rotational speed, and the lower the stiffness, the more significant the separation. This amplification effect makes the detection extremely sensitive to even small eccentricities, while providing directly visible evidence for larger eccentricities (0.1-2 mm).

[0024] Select the target marker pair with the largest center offset; its orientation is the direction of the center of gravity offset, with a directional accuracy of <2°. If the offset value needs to be quantified, pre-calibrate the relationship curve of "center offset d → center of gravity offset e" using a known eccentric standard component (usually linear with a slope of 1-5 under selected stiffness and rotational speed). In actual testing, e can be directly obtained from a table with an accuracy of 5-20 μm. Even without calibration, the direction and relative magnitude alone are sufficient to guide the counterweight.

[0025] To verify the condition of the buffer layer, repeated tests were conducted at multiple speeds, including 300 rpm, 800 rpm, 1500 rpm, and 2500 rpm. If the deviation in the offset direction at each speed is less than 4°, the offset is proportional to the square of the speed, and the proportionality coefficient is consistent with >95%, then the radial stiffness is considered normal; otherwise, it indicates that the gasket is aging or damaged.

[0026] The entire system is extremely easy to deploy on a production line: each workstation only needs a strobe lighting unit installed inside a protective enclosure, along with an industrial camera or intelligent image sensor facing the workpiece end face and a speed encoder mounted on the spindle or motor. All eccentricity recognition algorithms can be processed in real time on a standard industrial computer, with the time from startup detection to outputting the eccentricity direction and amount results typically within about twenty seconds. The industrial computer at each workstation uploads the detected workpiece identification, center of gravity offset direction, and center of gravity offset to a host monitoring server via an industrial IoT or IoT network. The monitoring server then centrally displays and archives multiple workpieces under inspection or in rotation, achieving remote online inspection and unified management of the entire production line.

[0027] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for detecting workpiece eccentricity based on strobe illumination and intelligent sensors, characterized in that, Includes the following steps: 1) The workpiece is fixed to the end of the rotating shaft through a layer of elastically deformable buffer material, so that the workpiece can be displaced in the radial direction relative to the rotating shaft when subjected to centrifugal force; 2) Set at least three sets of marker point pairs on the surface to be inspected of the workpiece. Each set of marker point pairs includes a first marker point and a second marker point arranged in pairs about a preset rotation center. The line connecting each set of marker point pairs defines an orientation direction. 3) Drive the rotating shaft to rotate the workpiece at the target speed, obtain the actual speed of the rotating shaft, calculate the baseline frequency f0 corresponding to the actual speed, and control the strobe illumination unit to illuminate the workpiece at a strobe frequency of f0 plus the frequency offset Δf to form a marker point afterimage with a slow rotation effect. 4) Control the intelligent image sensor to expose frame by frame according to the triggering beat synchronized with the strobe illumination unit. The intelligent image sensor includes an adaptive exposure control module. The adaptive exposure control module adjusts the exposure time according to the strobe frequency and the frequency offset to obtain a sequence of marker point afterimage images. 5) Perform image recognition, including extracting the pixel coordinate sequence of each first marker point and the pixel coordinate sequence of the corresponding second marker point from the marker point afterimage image sequence; 6) Perform image processing, including performing circular trajectory fitting on the pixel coordinate sequence of each pair of marker points to obtain the circular trajectory of the first marker point and the circular trajectory of the second marker point, as well as the center offset of the two circular trajectories; select the target marker point pair with the largest center offset of the circular trajectory among all the marker point pairs, and take the orientation direction of the target marker point pair as the workpiece center of gravity offset direction.

2. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, The preset rotation center is the projection point of the theoretical axis of the rotation axis on the surface to be tested. The first and second marker points of each pair of marker points are arranged on both sides of the preset rotation center along the corresponding orientation direction, and the nominal distances of the first and second marker points to the preset rotation center are equal.

3. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, The baseline frequency f0 is an integer multiple of the mechanical rotation frequency corresponding to the actual rotation speed, and the frequency offset Δf satisfies 0 < |Δf| ≤ 0.05·f0. The adaptive exposure control module sets the exposure time to a fixed proportion of one strobe cycle length, and reduces the analog gain when the average gray level of the afterimage image brightness statistics result is close to the preset saturation threshold, so that the apparent rotational angular velocity of the marker point afterimage is less than five percent of the mechanical rotation frequency and the gray level is kept within the preset dynamic range.

4. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, Extracting the pixel coordinate sequence of each first marker point and the pixel coordinate sequence of the corresponding second marker point from the marker point afterimage image sequence includes: performing grayscale thresholding on each frame of the afterimage image in the marker point afterimage image sequence to obtain candidate marker point regions; performing connected component labeling on the candidate marker point regions to obtain the connected components corresponding to each first marker point and each second marker point; and performing subpixel centroid localization on each connected component to obtain the subpixel precision pixel coordinates of each first marker point and each second marker point in each frame of the afterimage image.

5. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, Performing circular trajectory fitting on the pixel coordinate sequence of each pair of marker points includes: for each pair of marker points, substituting the pixel coordinate sequence of the corresponding first marker point into a preset least squares circle fitting algorithm to solve for the center coordinates and radius of the circular trajectory of the first marker point; substituting the pixel coordinate sequence of the corresponding second marker point into the least squares circle fitting algorithm to solve for the center coordinates and radius of the circular trajectory of the second marker point; and using the Euclidean distance between the center coordinates of the two as the center offset of the circular trajectory.

6. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, After selecting the target marker point pair, the method further includes: comparing the center offset of the circular trajectory of the target marker point pair with a pre-calibrated correspondence curve, wherein the correspondence curve provides a functional relationship between the center offset of the circular trajectory and the center offset of the workpiece, and determining the numerical index of the center offset of the workpiece based on the position of the center offset of the circular trajectory of the target marker point pair in the correspondence curve.

7. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, The buffer material layer is an annular elastomer gasket disposed between the workpiece end face and the connecting fixture, and the radial stiffness of the annular elastomer gasket is between 1×10³ N / mm and 1×10 5 The values ​​are between N / mm, and the axial stiffness of the annular elastomer gasket is higher than the radial stiffness.

8. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, The first and second markers in the marker pair adopt a graphic structure that can form a high-contrast shadow under the illumination of the strobe lighting unit. The graphic structure is a solid protrusion or solid groove with sharp edges.

9. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, The adaptive exposure control module adjusts the exposure time as follows: the current strobe operating frequency f1 is obtained by adding the strobe frequency and the frequency offset, the current strobe cycle length T is defined as 1 / f1, and the exposure time is set to k·T, where the value of k ranges from 0.1 to 0.

5.

10. The workpiece eccentricity detection method based on strobe illumination and intelligent sensors according to claim 1, characterized in that, The method is also used to perform online inspection of multiple workpieces that are being inspected or rotating on the production line. Each workpiece is assigned a workpiece identifier, and the obtained inspection results are combined with the corresponding workpiece identifier to form inspection data. This data is then sent to a remote monitoring server via the Industrial Internet of Things or the Internet. The remote monitoring server centrally receives and stores the inspection data from multiple workstations.

Images