Display light color detection system

Abstract

The invention discloses a display light color detection system, and relates to the technical field of low-brightness display light color detection, and the system comprises a dynamic modulation module which is configured to carry out the high-frequency pulse modulation of a dark-state signal of a detection object, and is used for enabling the dark-state signal to carry a modulation component of a characteristic frequency; the time sequence division acquisition unit comprises a high-speed optical shutter assembly and a detector, and the high-speed optical shutter assembly synchronously switches the on-off state of an optical path according to the characteristic frequency; according to the invention, through high-frequency pulse modulation of dark state signals, time-sequence-divided acquisition of noise and signals and time-frequency domain combined decoupling, traditional hardware limitation is broken through, ultra-low brightness signals and stray light interference are effectively separated, 0.01 cd / m < 2 >-level high-precision detection is realized, the detection efficiency is improved by more than 7 times, and the misjudgment rate is significantly reduced. The problems of missing detection of a novel display technology and the yield of an automatic production line are solved, and an active coding-time sequence calibration-cross-domain cooperation framework has remarkable innovativeness and industrial value.

Description

Technical Field

[0001] The present invention relates to the technical field of low-brightness display light color detection, and in particular to a display light color detection system. Background Art

[0002] In low-brightness display detection scenarios (such as mobile phone screen-off displays and MicroLED dark-state pixels), the inherent brightness of the detection object is extremely low, which leads to an extreme amplification of the contradiction between the detection signal and environmental interference, creating a hidden challenge that traditional technologies are difficult to overcome. The specific problem can be analyzed in depth from the following dimensions:

[0003] 1. The nature and characteristics of stray light interference:

[0004] 1.1 Complexity of Interference Sources: Even with traditional light shields (e.g., multi-layer physical shielding structures), two key types of stray light cannot be completely eliminated:

[0005] Equipment-inherent interference: The propagation path of parasitic light signals generated by the detection system's own calibration light source, diffuse reflection from indicator lights, and electromagnetic radiation coupling from circuit components highly overlaps with the detection light path, making it difficult to completely isolate them through hardware shielding.

[0006] Dynamic environmental interference: Reflections caused by people walking and scattered light from peripheral equipment (such as conveyor belt lighting and reflective surfaces of robotic arms) are less intense than the main light source. However, during dark-state detection, they overlap with the weak panel signal at the same frequency, causing the signal characteristics to be submerged by noise.

[0007] 1.2 Nonlinear effects of interference:

[0008] The superposition of stray light and the panel dark state signal is not a simple power superposition, but a photoelectron mixing effect on the photosensitive surface of the detector - the DC component of the ambient light continuously excites the background noise of the detector, while the pulse component of the panel dark state signal (if it is dynamic detection) is severely modulated by the noise baseline, causing the traditional time domain averaging or filtering algorithm to fail.

[0009] 2. Bottlenecks and mindset of existing technologies:

[0010] 2.1. Hardware-oriented technical limitations: Traditional detection solutions follow the linear thinking of "enhancing signals and suppressing noise":

[0011] Detector upgrade: Enhanced CCD or EM-CCD is used to improve sensitivity. However, due to the physical limitations of the detector (such as readout noise and dark current), the noise level increases nonlinearly with the sensitivity increase, and the improvement in signal-to-noise ratio is limited at extremely low brightness.

[0012] Static light-shielding design: Relying on hardware measures such as fixed light shields and matte coatings, it can only handle external strong light interference, and lacks the ability to dynamically suppress the device's internal stray light (such as the leakage path of the calibration light source) and diffuse stray light (such as multiple reflections of indirect light).

[0013] 2.2. Missing Dimensions in Signal Processing: Existing algorithms do not fully utilize the time / frequency domain differences between signals and noise:

[0014] Ambient stray light often manifests as a steady-state DC signal (or low-frequency noise), while the panel's dark-state signal (if there is dynamic drive during the detection process) may contain high-frequency modulation characteristics (such as PWM drive signals). However, traditional detection systems use static integration acquisition and do not perform time-frequency domain decoupling of the signal, resulting in the effective signal characteristics being masked by noise.

[0015] 3. Systemic impact in practical applications:

[0016] 3.1. Detection blind spots of new display technologies: The dark pixel brightness of self-luminous display technologies such as MicroLED and OLED has been reduced to 0.1cd / m 2 Below (some near-eye display scenes require 0.01cd / m 2 The luminous intensity is close to the detector noise floor. The traditional detection system fails to detect in this range, resulting in:

[0017] Missed defect detection: Defects such as dark spots and uneven brightness cannot be identified, directly affecting the visual experience of the end product (such as the "screen door effect" or black spot problem of AR / VR devices).

[0018] Yield misjudgment: Qualified products are misjudged as defective (passed inspection) or defective products are passed (missed inspection), resulting in material waste on the production line or batch quality accidents.

[0019] 3.2 The Irreplaceable Dilemma of Manual Visual Inspection: When relying on manual visual inspection to fix leaks, there are two core problems:

[0020] Low efficiency: Low-brightness inspection requires pixel-by-pixel observation in a darkroom environment. Single-screen inspection takes more than 1 minute, which cannot meet the cycle requirements of automated production lines (usually <10 seconds / piece).

[0021] Subjective error: The human eye has limited ability to distinguish low brightness differences (the threshold is about 0.5cd / m 2 ), and is easily affected by visual fatigue, resulting in a significant increase in the rate of false defect judgment.

[0022] In view of this, a display light color detection system is provided to overcome the above problems. Summary of the Invention

[0023] The object of the present invention is to provide a display light color detection system to solve the problems raised in the above background technology.

[0024] To solve the above technical problems, the present invention provides a display light color detection system, comprising:

[0025] Dynamic modulation module: configured to perform high-frequency pulse modulation on the dark-state signal of the detection object, so as to modulate the dark-state signal to carry a modulation component of a characteristic frequency;

[0026] Time-division acquisition unit: includes a high-speed optical shutter component and a detector. The high-speed optical shutter component synchronously switches the on-off state of the optical path according to the characteristic frequency, and is used to time-divisionally acquire the optical signals of the target signal period and the reference noise period within the pulse period of the dark state signal;

[0027] Cross-domain decoupling processing module: It is configured to separate dark state signals from stray light interference through the joint processing of time domain difference and frequency domain filtering based on the collected data of the target signal period and the reference noise period.

[0028] Furthermore, the dynamic modulation module includes a pulse width modulation drive circuit, the modulation frequency range of the PWM drive circuit is 1kHz-10MHz, and the modulation frequency is higher than the highest characteristic frequency of the ambient stray light.

[0029] Furthermore, the high-speed optical shutter assembly includes an electrically controlled shutter synchronized with the dynamic modulation module. The opening time of the electrically controlled shutter corresponds to the pulse peak period of the dark state signal, and the closing time corresponds to the pulse cutoff period of the dark state signal, which are used to define the target signal period and the reference noise period respectively.

[0030] Furthermore, the cross-domain decoupling processing module includes:

[0031] Time domain difference unit: configured to perform real-time differential calculations on the collected data during the target signal period and the reference noise period to eliminate the DC bias of steady-state stray light;

[0032] Frequency domain filtering unit: configured to perform bandpass filtering on the differential signal, retain the modulation component of the characteristic frequency, and filter out low-frequency noise.

[0033] Furthermore, it also includes a non-coaxial optical path isolation structure;

[0034] The outgoing optical path of the dynamic modulation module and the receiving optical path of the detector are arranged at a preset angle, and a wavelength-matched narrowband filter is set in the optical path to isolate the parasitic optical signal generated by the detection system.

[0035] Furthermore, the preset angle is greater than the critical angle of diffuse reflection of stray light, the passband wavelength of the narrowband filter matches the luminescence wavelength of the dark state signal, and the bandwidth is less than 10 nm.

[0036] Furthermore, the acquisition frame rate of the detector corresponds to the modulation frequency of the dynamic modulation module in an integer multiple, and is used for synchronous sampling of target signal periods and reference noise periods at equal intervals.

[0037] Furthermore, the cross-domain decoupling processing module calibrates the phase of the modulation component of the characteristic frequency to eliminate signal distortion caused by optical path delay and circuit noise.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. Coaxial Isolation and Wavelength Screening of Endogenous Interference in Equipment: Through the collaborative design of a non-coaxial optical path isolation structure (preset angle > critical angle for diffuse reflection, such as 45° in the embodiment) and a narrowband filter (bandwidth < 10nm, such as 5nm), the coaxial propagation path of endogenous stray light, such as leakage from the calibration light source and diffuse reflection from LEDs on the circuit board, is blocked, filtering out parasitic light signals of non-target wavelengths (e.g., attenuating 625nm red light interference by over 95% in the embodiment). Compared to traditional static light shields that can only handle strong external light, this structure has a targeted suppression effect on coaxially propagating endogenous interference in equipment, resolving the problem of hardware shielding being unable to isolate parasitic signals in the same optical path.

[0040] 2. Time domain difference cancellation of environmental dynamic interference: Using the synchronous control of the high-speed optical shutter component and the dynamic modulation module (such as the 2μs signal period / 18μs noise period in the embodiment), the mixed light containing the signal and the pure noise light are collected in the same detection cycle, and the mixed light containing the signal and the pure noise light are collected in the same detection cycle through time domain difference operation. Real-time DC offset subtraction of steady-state stray light (residual <0.1%). This method converts dynamic environmental interference (such as conveyor belt scattered light and robotic arm reflections) into a measurable reference baseline, leveraging its steady-state characteristics to achieve precise noise subtraction, avoiding the problem of dynamic noise and weak signal co-frequency superposition in traditional methods.

[0041] 3. Cross-domain decoupling of nonlinear effects: To address the persistent excitation of the detector's background noise by the DC component of stray light, high-frequency pulse modulation (such as a 50kHz PWM signal) is used to impart characteristic frequency components to the dark-state signal. Combining time-domain differencing (to eliminate the noise baseline) with frequency-domain filtering (to extract the 50kHz high-frequency signal), this method converts signal distortion caused by the "photoelectron mixing effect" into separable time-frequency domain differences. The measured signal-to-noise ratio has increased from 1:3 to 15:1, overcoming the failure bottleneck of traditional algorithms caused by noise baseline modulation.

[0042] 4. Reverse breakthrough of hardware-oriented limitations: Abandoning the linear thinking of "enhancing detector sensitivity", through active signal coding (high-frequency pulse modulation) and optical path timing control (time-sharing acquisition), the effective signal is separated from the noise floor without relying on high-noise detectors such as EM-CCD. For example, in the embodiment, ordinary scientific-grade CMOS detectors (quantum efficiency 75% @ 450nm) are used to achieve 0.01cd / m 2 It achieves super-high detection accuracy, breaking the physical limit of “sensitivity improvement accompanied by nonlinear noise growth”, and reduces detection costs by 60% compared with traditional solutions.

[0043] 5. Joint decoupling of signal processing in the time and frequency domains: Traditional static integration acquisition fails to exploit the time / frequency domain differences between signal and noise. This method uses dynamic modulation to impart high-frequency characteristics (e.g., 50kHz) to the signal and implements joint sampling in the time and frequency domains via synchronized optical gates. Time-domain differential filtering removes DC noise, while frequency-domain filtering isolates low-frequency interference (0-10kHz), preserving the high-frequency components of the target signal. This method achieves signal separation without complex algorithms, achieving a threefold improvement in efficiency compared to traditional full-band FFT processing, while avoiding the computational complexity of algorithms like blind source separation.

[0044] 6. Ultra-low-cost, high-precision detection of new display technologies: 0.01cd / m2 for MicroLED and OLED dark-state pixels 2 To meet the requirements of level detection, the minimum detectable brightness is reduced from 0.1cd / m2 in traditional system through the “modulation coding-timing sampling-cross-domain decoupling” architecture. 2 Increased to 0.01cd / m 2 The dark spot missed detection rate has been reduced from 12% to 0.3%, and the false positive rate has been reduced from 8% to 0.5%. For example, in AR / VR near-eye display panel inspection, it can accurately identify uneven brightness defects that cause the "screen door effect," avoiding batch quality accidents.

[0045] 7. Efficiency and standardization of automated production lines: Single-screen inspection time is shortened to 8 seconds (traditional manual visual inspection takes 1 minute), and no darkroom environment is required. Dynamic modulation and optical path control enable full-environment adaptive detection. The detection process is completed by hardware timing and algorithm, eliminating the subjective error of the human eye (the human eye resolution threshold is 0.5cd / m 2 ), meeting the production line cycle requirement of <10 seconds / piece and significantly improving yield control efficiency.

[0046] 8. Inverse Signal Modulation: From "Passive Noise Reduction" to "Active Encoding": Traditional technologies rely on detector or light-shielding hardware upgrades. This invention takes the opposite approach, using high-frequency PWM modulation to impart unique frequency domain characteristics (e.g., 50kHz) to weak signals, completely separating them from low-frequency noise in the frequency domain. This approach of "converting the signal into the 'opposite' of noise" breaks the stereotype that inherent signal strength cannot be changed, achieving a "dimensionality reduction" attack on noise.

[0047] 9. Time-sequential noise utilization: Converting noise into a reference baseline: Utilizing the steady-state characteristics of ambient stray light, we proactively create a comparison between "signal periods" and "noise periods," converting them into a real-time measurable reference baseline (e.g., the average of 45 noise sampling points in this example). This method overturns the conventional wisdom that "noise must be completely eliminated" by directly offsetting the noise floor through differential calculations, eliminating the need for complex noise modeling or post-compensation, achieving the counterintuitive effect of "countering noise with noise."

[0048] 10. Cross-domain collaborative architecture: Three-dimensional coupling of optics, electronics, and algorithms: The deep collaboration of non-coaxial optical paths (optics), high-frequency modulation drive (electronics), and time-frequency domain decoupling algorithms (signal processing) forms a three-dimensional noise reduction system of "hardware isolation-signal encoding-digital separation." For example, optical path angle design and filter selection provide the physical basis for signal modulation, and time series sampling provides accurate data for algorithm processing; all three are indispensable. This interdisciplinary design cannot be derived by non-technical personnel through the optimization of a single link, and requires a breakthrough in the conventional thinking of "independent design of hardware and algorithms."

[0049] This invention overcomes the industry challenge of "extreme signal-noise imbalance" in low-brightness display detection by leveraging the technical approach of "dynamic modulation to impart signal characteristics - time-sharing acquisition to separate time-frequency differences - and cross-domain decoupling and combined noise suppression" without relying on high-end detectors or complex algorithms. Its core principle is to leverage the inherent differences in the properties of signal and noise (time-domain periodicity, frequency-domain distribution, and optical path propagation characteristics) to transform disadvantages into advantages through systematic design, achieving simultaneous breakthroughs in detection accuracy, efficiency, and cost, providing a revolutionary solution for mass production testing of emerging display technologies such as MicroLED. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a schematic diagram of a display light color detection system according to the present invention. DETAILED DESCRIPTION

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0052] See also Figure 1 , the present invention provides a technical solution:

[0053] See Figure 1 As shown, an embodiment of a display light color detection system:

[0054] Application scenarios:

[0055] This embodiment is applied to the dark pixel detection scenario of MicroLED display panel, and the detection object is 0.05cd / m 2 The goal is to detect dark pixel defects (brightness abnormality less than 0.03cd / m 2 ).

[0056] Specific implementation steps:

[0057] 1. High-frequency pulse coding of dynamic modulation module:

[0058] Hardware implementation: A PWM drive circuit (model: TIDRV8834) is used to apply a 50kHz pulse width modulation signal with a duty cycle of 10% to the MicroLED pixels, so that the dark state signal in the time domain presents a high-frequency pulse with a period of 20μs and a high level duration of 2μs (characteristic frequency 50kHz).

[0059] Control logic: The modulation frequency is configured to 50kHz through registers (which is 10kHz higher than the highest characteristic frequency of ambient stray light. Spectrum analysis shows that the ambient dynamic interference is mainly concentrated in the range of 0-5kHz).

[0060] The dark state signal with extremely low inherent brightness is converted into a pulse signal with a characteristic frequency of 50kHz, so that it is significantly different from the steady-state stray light (DC / low frequency) in the frequency domain, solving the problem of "signal and noise superposition at the same frequency".

[0061] Avoid relying on increased detector sensitivity, enhance signal detectability through active signal encoding, and break through the physical limitations of hardware.

[0062] 2. Synchronous shutter control of the time-series acquisition unit:

[0063] High-speed optical shutter assembly: An electronically controlled shutter (model: Copal N3, response time ≤ 1 μs) is used, which is synchronized with the PWM drive circuit via an FPGA (Xilinx Spartan-6).

[0064] Target signal period: The shutter is open during the PWM high level period (2μs) to collect a mixed optical signal containing dark state signal and stray light (detector model: scientific-grade CMOS, Teledyne FLIR Grayline, quantum efficiency 75% @ 450nm).

[0065] Reference noise period: The shutter is closed during the PWM low-level period (18 μs), and only stray light (including internal leakage light from the device and dynamic scattered light from the environment) is collected.

[0066] Sampling timing: The detector frame rate was set to 1 MHz (a 20-fold integer multiple of the 50 kHz modulation frequency) to ensure that 100 data points were collected at equal intervals within each modulation cycle.

[0067] Through time domain segmentation, "signal effective period" and "noise reference period" are actively created, and shutter synchronization is used to directly separate pulse signals and steady-state noise, solving the problem that traditional static integration acquisition cannot distinguish time domain differences.

[0068] Dynamic environmental interference (such as glare from people walking) appears as the same DC bias in the reference period and the signal period, and can be completely offset through subsequent differential operations without relying on complex noise modeling.

[0069] 3. Hardware design of non-coaxial optical path isolation structure:

[0070] Optical path layout:

[0071] The modulated dark-state signal outgoing optical path forms a 45° angle with the detector receiving optical path (greater than the critical angle of 30° for diffuse reflection of stray light, as verified by Zemax optical simulation), blocking the coaxial propagation path of stray light generated within the device (such as leakage light from the calibration light source returning along the original optical path).

[0072] A narrowband filter (model: Semrock FF01-450 / 5-25, passband wavelength 450nm±2.5nm, bandwidth 5nm) was installed at the front end of the detector to match the MicroLED emission wavelength and filter out endogenous interference from non-target wavelengths such as the circuit board LED (625nm red light).

[0073] The combination of physical isolation and wavelength screening effectively suppresses the device's inherent stray light (for example, the leakage of the calibration light source is attenuated by more than 95%), breaking through the limitation of traditional static light shields that cannot handle coaxial parasitic light.

[0074] The non-coaxial design does not increase the complexity of the optical path and only achieves intrinsic interference isolation through angle adjustment and filter selection. The cost is 60% lower than the traditional multi-layer shielding structure.

[0075] 4. Signal separation of cross-domain decoupling processing modules:

[0076] 4.1、Time Domain Difference Unit:

[0077] Processing logic: Perform time domain mean difference on the collected data of the target signal period and the reference noise period to eliminate the DC bias component of the steady-state stray light. The details are as follows:

[0078] Target signal period: During the high-frequency pulse high level period output by the dynamic modulation module (2 μs in this embodiment), the detector collects 5 photocurrent signal samples (denoted as S1, S2, ..., S5) at a frame rate of 1 MHz.

[0079] ), whose mean is:

[0080]

[0081] The mean value includes the superposition signal of the high-frequency pulse component of the dark state signal and the steady-state stray light.

[0082] Reference noise period: corresponds to the period of low-level high-frequency pulse output by the dynamic modulation module (18 μs in this embodiment), the detector collects 45 photocurrent signal samples at the same frame rate (denoted as n1, n2, ..., n 45 ), whose mean is:

[0083]

[0084] This mean value only includes steady-state stray light (DC / low-frequency noise such as internal leakage light from the device and dynamic scattered light from the environment).

[0085] Real-time differential operation:

[0086]

[0087] Through the above calculation, the DC bias of the steady-state stray light is deducted in real time, leaving only the target signal component with high-frequency modulation characteristics (actual measurement shows that the residual rate of the DC component of the ambient dynamic scattered light is less than 0.1%).

[0088] Parameter explanation:

[0089] S i : the photocurrent amplitude at the i-th sampling point during the target signal period (unit: nA), corresponding to the mixed response of the dark state signal pulse and stray light;

[0090] n j: The photocurrent amplitude at the jth sampling point during the reference noise period (unit: nA), which only reflects the noise floor of stray light;

[0091] The number of sampling points is set based on the integer multiple relationship between the detector frame rate (1 MHz) and the modulation period (20 μs), ensuring that the number of sampling points in the signal period and the noise period is an integer division of the modulation frequency (in this embodiment, the signal period accounts for 10% of the modulation period, corresponding to 5 sampling points; the noise period accounts for 90%, corresponding to 45 sampling points);

[0092] is the sampling mean of the target signal period, including the mixed response of dark state signal and stray light, in nA. In this embodiment, the value is calculated based on the 5-point mean:

[0093] is the sampling mean value of the reference noise period, including only the stray light noise floor, in nA. In this embodiment, the value is calculated based on the 45-point mean.

[0094] S diff is the target signal component after time domain difference, unit is nA.

[0095] 4.2, Frequency domain filtering unit:

[0096] Processing logic: time domain differential signal S diff Perform frequency domain feature extraction. The specific steps are as follows:

[0097] Fast Fourier Transform (FFT): Converts the time domain signal into a frequency domain distribution, showing the spectrum difference between the 50kHz characteristic frequency component of the dark state signal and the low-frequency noise;

[0098] Bandpass filtering: An 8th-order IIR bandpass filter (center frequency 50kHz, passband range 45kHz-55kHz) is used to retain the high-frequency signal components that match the modulation frequency and filter out low-frequency noise of 0-10kHz (including device circuit noise, low-frequency fluctuations of ambient diffuse light, etc.).

[0099] Parameter design basis:

[0100] Modulation frequency 50kHz: higher than the highest characteristic frequency of environmental dynamic interference (measured at 10kHz by spectrum analysis), ensuring complete separation of target signal and noise in the frequency domain;

[0101] ±5kHz filter bandwidth: takes into account both signal integrity and noise suppression capabilities, and can filter out more than 98% of low-frequency noise components in actual measurements.

[0102] 4.3 Phase calibration unit:

[0103] Processing logic: Using the synchronous clock signal of the dynamic modulation module as a reference, the filtered high-frequency signal is subjected to phase distortion correction:

[0104] Phase offset detection: Calculates the phase difference between the detection signal and the reference clock using a cross-correlation algorithm to identify phase offsets caused by optical path delay (approximately 5 ns, due to non-coaxial optical path length differences) and circuit noise (up to ±5°).

[0105] Phase compensation operation: Based on the detected phase difference, the signal is adjusted in the opposite phase to restore the original amplitude accuracy of the dark state signal and avoid brightness detection errors caused by phase distortion.

[0106] Technical advantages:

[0107] Optical path delay compensation: Pre-calibrate the optical path length difference through geometric optical modeling, and write the phase calibration parameters into the FPGA control logic to achieve real-time dynamic compensation;

[0108] Circuit noise robustness: Even with a ±5° phase offset (corresponding to a signal amplitude attenuation of <1.5%), the detection error can still be controlled to 0.001cd / m through phase calibration. 2 Within.

[0109] Noise Floor Subtraction Using Time-Domain Difference: By separating the mean differences between signal and noise periods, this method directly eliminates the DC bias of steady-state stray light, addressing noise baseline modulation caused by the "photoelectron mixing effect" (e.g., nonlinear interference from ambient light continuously stimulating the detector's noise floor). Measurements show that this process improves the signal-to-noise ratio from 1:3 with traditional methods to 15:1, significantly improving the detectability of weak signals.

[0110] Frequency-domain filtering for feature signal extraction: Utilizing the frequency domain differences between the target signal and noise (high-frequency modulation vs. low-frequency noise), bandpass filtering accurately extracts the 50kHz feature component, avoiding the blind amplification of full-band signals by traditional algorithms (e.g., the nonlinear noise growth problem in enhanced detectors). This operation reduces reliance on detector sensitivity and improves detection accuracy by more than 10 times under the same hardware conditions.

[0111] Phase-calibrated signal integrity restoration: By using a synchronized clock reference and phase correlation algorithm, signal distortion caused by optical path delay and circuit noise is corrected to ensure that the amplitude measurement error of the dark state signal is less than 5%, meeting the MicroLED dark state pixel standard of 0.01cd / m 2 Level 1 detection accuracy requirements.

[0112] 5. Defect determination and system integration:

[0113] Threshold comparison: Convert the decoupled signal amplitude into a brightness value (calibrated so that 1mV corresponds to 0.01cd / m2 ), set the dark spot defect threshold to 0.03cd / m 2 , when the detection value is less than the threshold, it is judged as a defect.

[0114] Production line adaptation: The inspection system is integrated into the automated assembly line, with a single-screen inspection time of 8 seconds (meeting the cycle requirement of <10 seconds per piece), replacing the manual visual inspection time of 1 minute per screen, increasing efficiency by 7.5 times.

[0115] Example verification data table:

[0116]

[0117]

[0118] Summarize:

[0119] Stray light interference suppression:

[0120] The device's endogenous stray light is isolated by a non-coaxial optical path (45° angle) and a narrowband filter (5nm bandwidth). Combined with time-series acquisition (2μs / 18μs time segmentation), the environmental dynamic interference is dynamically offset to solve the coaxial propagation and same-frequency superposition problems of two types of key stray light.

[0121] Nonlinear effect cracking:

[0122] The dark state signal is actively modulated into a 50kHz high-frequency pulse. Time domain difference (S_diff operation) is used to eliminate the steady-state noise baseline, and frequency domain filtering (bandpass 50kHz) is used to extract characteristic components to avoid signal distortion caused by photoelectron mixing, thus breaking through the failure bottleneck of traditional algorithms.

[0123] Collaborative innovation of hardware and algorithms:

[0124] Abandoning the linear thinking of "simply improving the sensitivity of the detector", signal encoding and timing control are realized through PWM modulation and high-speed optical gate synchronization, completing noise separation at the hardware level and reducing dependence on high-end detectors.

[0125] Adaptation of new display technologies:

[0126] At 0.05cd / m 2 In ultra-low brightness detection, the detection accuracy reaches 0.01cd / m 2 , solve the problems of missed detection and misjudgment of MicroLED dark pixels, and avoid the "screen door effect" and batch quality accidents of AR / VR devices.

[0127] Automated production lines improve efficiency:

[0128] The single-screen inspection time is 8 seconds, which meets the assembly line rhythm requirements and replaces the subjectivity and inefficiency of manual visual inspection. The misjudgment rate is reduced from 8% to 0.5%, significantly improving the yield control capability.

[0129] This embodiment achieves efficient signal-noise separation in extremely low-brightness scenarios through the unconventional technical path of "dynamic modulation-time-sharing acquisition-cross-domain decoupling" without relying on advanced detectors or complex algorithms. Its core innovations (active signal coding, timing noise reference, and optical path angle modulation) can be deduced by non-technical personnel through conventional design, and have significant creativity and industrial application value.

Claims

1. A display light color detection system, characterized in that: include: Dynamic modulation module: configured to perform high-frequency pulse modulation on the dark-state signal of the detection object, so as to modulate the dark-state signal to carry a modulation component of a characteristic frequency; Time-division acquisition unit: includes a high-speed optical shutter component and a detector. The high-speed optical shutter component synchronously switches the on-off state of the optical path according to the characteristic frequency, and is used to time-divisionally acquire the optical signals of the target signal period and the reference noise period within the pulse period of the dark state signal; Cross-domain decoupling processing module: It is configured to separate dark state signals from stray light interference through the joint processing of time domain difference and frequency domain filtering based on the collected data of the target signal period and the reference noise period.

2. A display light color detection system according to claim 1, characterized in that: The dynamic modulation module includes a pulse width modulation drive circuit. The modulation frequency range of the PWM drive circuit is 1kHz-10MHz, and the modulation frequency is higher than the highest characteristic frequency of the ambient stray light.

3. The display light color detection system according to claim 1, wherein: The high-speed optical shutter assembly includes an electrically controlled shutter synchronized with the dynamic modulation module. The opening time of the electrically controlled shutter corresponds to the pulse peak period of the dark state signal, and the closing time corresponds to the pulse cutoff period of the dark state signal, which are used to define the target signal period and the reference noise period respectively.

4. A display light color detection system according to claim 1, characterized in that: The cross-domain decoupling processing module includes: Time domain difference unit: configured to perform real-time differential calculations on the collected data during the target signal period and the reference noise period to eliminate the DC bias of steady-state stray light; Frequency domain filtering unit: configured to perform bandpass filtering on the differential signal, retain the modulation component of the characteristic frequency, and filter out low-frequency noise.

5. A display light color detection system according to claim 1, characterized in that: It also includes a non-coaxial optical path isolation structure; The outgoing optical path of the dynamic modulation module and the receiving optical path of the detector are arranged at a preset angle, and a wavelength-matched narrowband filter is set in the optical path to isolate the parasitic optical signal generated by the detection system.

6. A display light color detection system as claimed in claim 5, characterized in that: The preset angle is greater than the diffuse reflection critical angle of stray light, the passband wavelength of the narrowband filter matches the luminescence wavelength of the dark state signal, and the bandwidth is less than 10nm.

7. The display light color detection system according to claim 1, wherein: The acquisition frame rate of the detector corresponds to the modulation frequency of the dynamic modulation module in an integer multiple, and is used for synchronous sampling of target signal periods and reference noise periods at equal intervals.

8. The display light color detection system according to claim 1, wherein: The cross-domain decoupling processing module calibrates the phase of the modulation component of the characteristic frequency and eliminates signal distortion caused by optical path delay and circuit noise.

Images