Monomer assembly debugging and positioning method based on photoelectric sensor

Abstract

The invention discloses a single component debugging and positioning method based on a photoelectric sensor. The method comprises the steps of S1, multi-spectral data synchronous acquisition, S2, adaptive parameter optimization, S3, multi-physics field error compensation and S4, iterative positioning correction. According to the invention, material insensitive measurement is realized through time-space synchronous acquisition of dual-band laser ranging and visible light imaging, a dynamic temperature compensation algorithm is adopted, the influence of ambient temperature excursion is eliminated, a surface reflectivity adaptive mechanism is innovated, the measurement stability is further improved, and a phase laser ranging and image edge identification fusion technology is adopted, so that the measurement accuracy is improved. The problems that traditional mechanical positioning is large in accumulative error, poor in flexibility and adaptability and the like are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of precision assembly automation, and particularly to a debugging and positioning method for single components based on optoelectronic sensors. Background Art

[0002] With the rapid development of industrial automation and intelligent manufacturing technologies, optoelectronic sensors are increasingly widely used in industrial production, especially in the fields of component positioning, debugging, and detection. In the fields of industrial automation and precision measurement, the debugging and positioning of single components (such as optical elements, mechanical parts, electronic components, etc.) is a key task that directly affects the overall performance and working efficiency of the system.

[0003] However, the existing debugging and positioning methods of optoelectronic sensors have the following problems:

[0004] The debugging efficiency of the existing methods is relatively low and cannot meet the requirements of high precision and rapid positioning; some methods rely on complex hardware devices or cumbersome manual operations, resulting in increased costs and prone to human errors;

[0005] When dealing with positioning tasks in multi-dimensional or dynamic environments, the accuracy and stability of the existing methods have significant deficiencies; the existing methods lack intelligent analysis of optoelectronic signals and are difficult to achieve efficient signal processing and position calculation. Summary of the Invention

[0006] (1) Technical Problems to be Solved

[0007] Aiming at the deficiencies of the existing technology, the present invention collects three-dimensional position information of components through a multi-spectral optoelectronic sensor array and combines an adaptive calibration algorithm to achieve sub-millimeter positioning accuracy, solving problems such as large cumulative errors and poor flexible adaptability existing in traditional mechanical positioning.

[0008] (2) Technical Solutions

[0009] To achieve the above object, the present invention provides the following technical solutions: A debugging and positioning method for single components based on optoelectronic sensors, including the following steps:

[0010] S1. Synchronous multi-spectral data acquisition:

[0011] Obtain three-dimensional point cloud data of the component through a phase-type laser ranging unit (wavelength 850nm / 1550nm switchable), cooperate with a visible light imaging unit (frame rate ≥ 60fps) to capture surface texture features, and use the PTP precise time protocol to achieve μs-level data synchronization;

[0012] S2. Adaptive parameter optimization:

[0013] Measure the surface reflectivity R. When R < 15%, switch to the 1550nm long-wavelength mode and increase the supplementary light intensity by 30%. When specular reflection is detected, automatically adjust the laser incident angle θ = arcsin(0.7R);

[0014] S3. Multi-physical field error compensation:

[0015] Establish a temperature-displacement compensation model: ΔL = α(T - T 0 )L 0 +β(T - T 0 ) 2

[0016] (α = 2.3×10 -6 / ℃, β = 1.7×10 -9 / ℃ 2 );

[0017] Perform mechanical backlash compensation: δ = γ·v 2 ·e^(-kv), γ = 0.12, k = 0.05;

[0018] S4. Iterative positioning correction:

[0019] Calculate the pose deviation based on the Levenberg-Marquardt algorithm, generate a fifth-order polynomial motion trajectory for smooth correction, and perform three-point verification measurement after positioning. If the deviation > threshold, trigger secondary compensation.

[0020] As a preferred solution, based on the debugging positioning system, it includes:

[0021] Distributed multi-spectral optoelectronic sensor array, including 3 groups of phase laser ranging units and 2 groups of visible light imaging units;

[0022] Dynamic compensation module, integrating a temperature compensation circuit and a surface reflectivity adaptive adjustment unit;

[0023] Intelligent control unit, equipped with a multi-modal data fusion algorithm and a motion trajectory planning module;

[0024] Six-degree-of-freedom precision motion platform, with XYZ-axis linear guide rails and a rotation adjustment mechanism.

[0025] As a preferred solution, the working wavelengths of the phase laser ranging units include a dual-band of 850nm and 1550nm, and the modulation frequency range is adjustable from 10MHz to 100MHz.

[0026] As a preferred solution, the multi-modal data fusion algorithm uses weighted Kalman filtering, and the weight coefficients are dynamically adjusted according to the real-time signal-to-noise ratio.

[0027] As a preferred solution, the surface reflectance adaptive adjustment unit includes a programmable LED array, which can output 32-level illuminance adjustment within the range of 0 - 1000 lux.

[0028] As a preferred solution, the specific process of S1 data synchronous acquisition includes:

[0029] Benchmark establishment, loading the CAD model of the component to generate a theoretical coordinate system;

[0030] Arranging a sensor array, with 2 groups of laser rangefinders + 1 group of wide-angle cameras at the top, and 1 group of obliquely installed laser sensors (the inclination angle can be adjusted from 15° to 45°) on each of the four sides;

[0031] Performing spatial calibration, using a standard calibration sphere (diameter 50 mm ± 1 μm), fitting the sensor coordinate system by the least squares method, with a residual error < 5 μm;

[0032] Data acquisition and processing, which consists of a rough scanning mode, a fine scanning mode, and feature fusion processing.

[0033] As a preferred solution, its three-dimensional point cloud sampling density is dynamically adjusted according to the positioning stage. The point distance in the rough positioning stage is 2 mm, and the point distance in the fine positioning stage is 0.2 mm. The original point cloud data is subjected to multi-scale noise reduction processing using the db9 wavelet basis.

[0034] As a preferred solution, the process of S4 is intelligent positioning correction, which specifically includes:

[0035] Deviation calculation, centroid offset: ΔX = Σ(x_i - x_ref) / N; angular deviation: Δθ = arctan((y_max - y_min) / (x_max - x_min));

[0036] Motion compensation, generating an S-shaped acceleration and deceleration curve to drive the six-axis platform to perform compensation motion;

[0037] Verification and iteration, using the three-point verification method to measure the actual pose. If |ΔX| > 0.03 mm, start secondary compensation: new compensation amount = original compensation amount × 0.7 + current deviation × 0.3.

[0038] (III) Beneficial effects

[0039] Compared with the prior art, the present invention provides a debugging and positioning method for single components based on optoelectronic sensors, which has the following beneficial effects:

[0040] Through the spatio-temporal synchronous acquisition of dual-band laser ranging and visible light imaging, the present invention realizes insensitive measurement of materials, adopts a dynamic temperature compensation algorithm to eliminate the influence of environmental temperature drift, innovates the surface reflectivity adaptive mechanism to further improve the measurement stability, and adopts the fusion technology of phase laser ranging and image edge recognition to solve the problems of large cumulative error and poor flexible adaptability existing in traditional mechanical positioning. Compared with the prior art, the positioning speed is increased by more than 40%, the repeated positioning accuracy reaches ±0.05 mm, and it can adapt to the differences in surface reflection characteristics of components made of different materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic flow chart of the method steps of the present invention;

[0042] Figure 2 It is a schematic diagram of the system modules of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] In order to better understand the purpose, structure and function of the present invention, the single-component debugging and positioning method based on photoelectric sensors of the present invention will be further described below in conjunction with the drawings and specific embodiments.

[0044] Embodiment 1

[0045] Referring to Figure 1-2 , the single-component debugging and positioning method based on photoelectric sensors of the present invention includes the following steps:

[0046] S1. Synchronous acquisition of multi-spectral data:

[0047] Obtain the three-dimensional point cloud data of the component through a phase laser ranging unit (wavelength 850 nm / 1550 nm switchable), cooperate with a visible light imaging unit (frame rate ≥ 60 fps) to capture surface texture features, and use the PTP precision time protocol to achieve μs-level data synchronization;

[0048] S2. Adaptive parameter optimization:

[0049] Measure the surface reflectivity R. When R < 15%, switch to the 1550 nm long-wavelength mode and increase the fill light intensity by 30%. When specular reflection is detected, automatically adjust the laser incident angle θ = arcsin(0.7R);

[0050] S3. Multi-physical field error compensation:

[0051] Establish a temperature-displacement compensation model: ΔL = α(T - T 0 )L 0 + β(T - T 0 ) 2

[0052] (α = 2.3×10 -6 / °C, β = 1.7×10 -9 / °C 2 );

[0053] Perform mechanical backlash compensation: δ = γ·v 2 ·e^(-kv), γ = 0.12, k = 0.05;

[0054] S4. Iterative positioning correction:

[0055] Calculate the pose deviation based on the Levenberg - Marquardt algorithm, generate a quintic polynomial motion trajectory for smooth correction, perform three - point verification measurement after positioning, and trigger secondary compensation if the deviation > threshold.

[0056] Specifically, the S1 data synchronous acquisition specifically includes:

[0057] Establish a reference, load the CAD model of the component to generate a theoretical coordinate system;

[0058] Arrange the sensor array, 2 groups of laser rangefinders + 1 group of wide - angle cameras on the top, and 1 group of obliquely installed laser sensors (the inclination angle is adjustable from 15° to 45°) on each of the four sides;

[0059] Perform spatial calibration, use a standard calibration sphere (diameter 50mm ± 1μm), fit the sensor coordinate system by the least - squares method, and the residual error < 5μm;

[0060] Data acquisition and processing, which consists of a rough scanning mode, a fine scanning mode, and feature fusion processing.

[0061] Its three - dimensional point cloud sampling density is dynamically adjusted according to the positioning stage. The point distance is 2mm in the rough positioning stage and 0.2mm in the fine positioning stage. The original point cloud data is processed by multi - scale noise reduction using the db9 wavelet basis.

[0062] The process of S4 is intelligent positioning correction, specifically including:

[0063] Deviation calculation, centroid offset: ΔX = Σ(x_i - x_ref) / N; angle deviation: Δθ = arctan((y_max - y_min) / (x_max - x_min));

[0064] Motion compensation, generate an S - shaped acceleration - deceleration curve, and drive the six - axis platform to perform compensation motion;

[0065] Verification iteration, measure the actual pose using the three - point verification method. If |ΔX| > 0.03mm, start secondary compensation: new compensation amount = original compensation amount × 0.7+ current deviation × 0.3.

[0066] Furthermore, the specific process of the method of the present invention is:

[0067] Phase 1: Benchmark Establishment

[0068] 1. Load the CAD model of the component to generate the theoretical coordinate system

[0069] 2. Arrange the sensor array:

[0070] Top: 2 groups of laser rangefinders + 1 group of wide-angle cameras

[0071] Four sides: 1 group of inclined-mounted laser sensors on each side (the inclination angle is adjustable from 15° to 45°)

[0072] 3. Perform spatial calibration

[0073] Use a standard calibration sphere (diameter 50mm ± 1μm)

[0074] Fit the sensor coordinate system by the least squares method, with the residual < 5μm

[0075] Phase 2: Data Acquisition and Processing

[0076] 1. Coarse scanning mode:

[0077] Laser power 10mW, scanning speed 200mm / s

[0078] Generate low-density point cloud (point spacing 2mm)

[0079] Identify the approximate orientation of the component (error ±0.5mm)

[0080] 2. Fine scanning mode:

[0081] Switch the laser power to 30mW, scanning speed 50mm / s

[0082] Adopt a spiral scanning path, point spacing 0.2mm

[0083] Synchronously trigger the camera to take pictures (exposure time 0.5ms)

[0084] 3. Feature fusion processing:

[0085] The laser point cloud data is denoised by the RANSAC algorithm

[0086] The image data is used to extract edges by the Canny operator (threshold 0.1 - 0.3)

[0087] Establish a multi-dimensional feature vector: V = [x, y, z, gray gradient, radius of curvature, reflection intensity]

[0088] Phase 3: Intelligent Positioning and Correction

[0089] 1. Deviation calculation:

[0090] Centroid offset: ΔX = Σ(x_i - x_ref) / N

[0091] Angle deviation: Δθ = arctan((y_max - y_min) / (x_max - x_min))

[0092] Using Mahalanobis distance to determine outliers (threshold 3σ)

[0093] 2. Motion compensation:

[0094] Generate an S-shaped acceleration and deceleration curve:

[0095] a(t) = a_max·sin 2 (πt / 2T)

[0096] (a_max = 0.5g, T = 0.8s)

[0097] Drive the six-axis platform to perform compensatory motion

[0098] 3. Verification and iteration

[0099] Use the three-point verification method to measure the actual pose

[0100] If |ΔX| > 0.03mm, start secondary compensation: New compensation amount = original compensation amount × 0.7 + current deviation × 0.3.

[0101] In this embodiment, the working wavelengths of the phase-type laser ranging unit include a dual-band of 850nm and 1550nm, the modulation frequency range is adjustable from 10MHz to 100MHz, the multi-modal data fusion algorithm uses weighted Kalman filtering, the weight coefficient is dynamically adjusted according to the real-time signal-to-noise ratio, and the surface reflectance adaptive adjustment unit includes a programmable LED array and can output 32-level illuminance adjustment in the range of 0-1000lux.

[0102] Embodiment 2

[0103] Based on the debugging and positioning method, this embodiment proposes a debugging and positioning system, which includes:

[0104] A distributed multi-spectral optoelectronic sensor array, including 3 groups of phase-type laser ranging units and 2 groups of visible light imaging units;

[0105] A dynamic compensation module, integrating a temperature compensation circuit and a surface reflectance adaptive adjustment unit;

[0106] An intelligent control unit, equipped with a multi-modal data fusion algorithm and a motion trajectory planning module;

[0107] A six-degree-of-freedom precision motion platform, having XYZ-axis linear guide rails and a rotation adjustment mechanism.

[0108] The system includes a sensing module, arranged in a five-sided annular array (top + four sides). The distance between adjacent sensors is d = √2 × the component feature size. The sampling frequency is 100 kHz for the laser ranging unit and 60 fps for the imaging unit. Its synchronization mechanism: realizes μs-level synchronization using the IEEE 1588 Precision Time Protocol;

[0109] The signal processing flow includes: raw signal acquisition, wavelet denoising (Daubechies9 basis function), feature extraction (extracting edge gradient and curvature features), coordinate transformation (establishing the workpiece coordinate system), and error compensation calculation. An improved PID control is adopted: Kp = 1.2, Ki = 0.05, Kd = 0.3. The path planning includes an S-shaped acceleration and deceleration curve, with a maximum acceleration of 0.5g.

[0110] Taking the IC chip packaging as an example:

[0111] Component size: 10mm × 10mm × 0.5mm

[0112] Sensor configuration: 2 groups of lasers + 1 group of cameras on the top, and 1 group of lasers on each of the four sides

[0113] Positioning process:

[0114] The rough positioning takes 0.8 s, with an accuracy of ±0.2 mm

[0115] The fine positioning takes 2.5 s, with a final accuracy of ±0.03 mm

[0116] Temperature change compensation: When the ambient temperature fluctuates by ±5°C, the system automatically compensates for an error of 0.6 μm.

[0117] Through the collaborative innovation of multi-physical field coupling measurement, intelligent compensation algorithms, and a hard real-time control system, the present invention significantly improves the positioning accuracy and reliability. It acquires the three-dimensional position information of the component through a multi-spectral optoelectronic sensor array and realizes sub-millimeter positioning accuracy in combination with an adaptive calibration algorithm. The system includes a distributed sensor module, a dynamic compensation module, and an intelligent feedback control unit, and adopts the fusion technology of phase laser ranging and image edge recognition to solve the problems of large cumulative errors and poor flexible adaptability existing in traditional mechanical positioning.

[0118] It can be understood that the present invention is described through some embodiments. Those skilled in the art know that without departing from the spirit and scope of the present invention, various changes or equivalent replacements can be made to these features and embodiments. Additionally, under the teaching of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application belong to the scope protected by the present invention.

Claims

1. A method for debugging and positioning a single component based on a photoelectric sensor, characterized in that: The steps include: S1. Multi-spectral data synchronous acquisition: The three-dimensional point cloud data of the component is obtained through the phase laser ranging unit (wavelength 850nm / 1550nm switchable), and the surface texture features are captured in conjunction with the visible light imaging unit (frame rate ≥60fps), and the PTP precision time protocol is used to achieve μs-level data synchronization; S2. Adaptive parameter optimization: Measure the surface reflectivity R. When R < 15%, switch to 1550nm long wavelength mode and increase the fill light intensity by 30%. When specular reflection is detected, automatically adjust the laser incident angle θ = arcsin (0.7R); S3. Multi-physics error compensation: Establish temperature-displacement compensation model: ΔL=α(T-T0)L0+β(T-T0) 2 (α＝2.3×10 -6 / ℃，β＝1.7×10 -9 / ℃ 2 )； Execute mechanical backlash compensation: δ=γ·v 2 ·e^(-kv), γ=0.12, k=0.05; S4. Iterative positioning correction: The posture deviation is calculated based on the Levenberg-Marquardt algorithm, and a quintic polynomial motion trajectory is generated for smooth correction. After positioning is completed, a three-point verification measurement is performed. If the deviation > threshold, secondary compensation is triggered.

2. The method for debugging and positioning a single component based on a photoelectric sensor according to claim 1, characterized in that: Based on debugging positioning system, including: Distributed multi-spectral photoelectric sensor array, including 3 sets of phase laser ranging units and 2 sets of visible light imaging units; Dynamic compensation module, integrating temperature compensation circuit and surface reflectivity adaptive adjustment unit; Intelligent control unit, equipped with multi-modal data fusion algorithm and motion trajectory planning module; Six-degree-of-freedom precision motion platform with XYZ axis linear guides and rotation adjustment mechanism.

3. The method for debugging and positioning a single component based on a photoelectric sensor according to claim 2, characterized in that: The working wavelength of the phase laser ranging unit includes dual bands of 850nm and 1550nm, and the modulation frequency range is adjustable from 10MHz to 100MHz.

4. The method for debugging and positioning a single component based on a photoelectric sensor according to claim 2, characterized in that ,The multimodal data fusion algorithm adopts weighted Kalman filtering, and the weight coefficient is dynamically adjusted ,according to the real-time signal-to-noise ratio.

5. The method for debugging and positioning a single component based on a photoelectric sensor according to claim 2, characterized in that: The surface reflectivity adaptive adjustment unit includes a programmable LED array, which can output 32 levels of illumination adjustment within the range of 0-1000 lux.

6. The method for debugging and positioning a single component based on a photoelectric sensor according to claim 2, characterized in that: The S1 data synchronous collection specifically includes: The benchmark is established and the CAD model of the loading component is used to generate the theoretical coordinate system; Arrange the sensor array, with 2 sets of laser rangefinders and 1 set of wide-angle cameras on the top, and 1 set of tilted laser sensors on each of the four sides (the tilt angle is adjustable from 15° to 45°); Perform spatial calibration, use a standard calibration ball (diameter 50mm±1μm), fit the sensor coordinate system by the least squares method, and the residual is <5μm; Data acquisition and processing consists of coarse scanning mode, fine scanning mode and feature fusion processing.

7. The method for debugging and positioning a single component based on a photoelectric sensor according to claim 6, characterized in that: The sampling density of the three-dimensional point cloud is dynamically adjusted according to the positioning stage. The point spacing in the coarse positioning stage is 2mm, and the point spacing in the fine positioning stage is 0.2mm. The db9 wavelet basis is used to perform multi-scale denoising on the original point cloud data.

8. The method for debugging and positioning a single component based on a photoelectric sensor according to claim 7, characterized in that: The process of S4 is intelligent positioning correction, which specifically includes: Deviation calculation, center of mass offset: ΔX = Σ(x_i-x_ref) / N; angular deviation: Δθ = arctan((y_max-y_min) / (x_max-x_min)); Motion compensation, generating S-shaped acceleration and deceleration curves, driving the six-axis platform to perform compensation motion; Verify the iteration and use the three-point verification method to measure the actual posture. If |ΔX|>0.03mm, start the secondary compensation: new compensation amount = original compensation amount × 0.7 + current deviation × 0.3.

Images