Method, system, device, and storage medium for adaptive control of cycle time and accuracy in quality inspection of notebook computer components.

By constructing a brightness convergence correlation diagram and dynamically adjusting the detection parameters, the problem of balancing accuracy and timing in keyboard backlight module testing was solved, achieving efficient and accurate quality testing.

CN123113318APending Publication Date: 2026-11-17苏州炳荣精密五金制品有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611172328.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-04
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the current quality testing of keyboard backlight modules, the backlight illumination waiting time and the number of brightness sampling times lack adaptive optimization, making it difficult to balance testing accuracy and testing cycle time, thus affecting the accuracy and stability of the test results.

Method used

By constructing a brightness convergence correlation graph, key convergence segments in the backlight component brightness establishment process are identified, and the backlight illumination waiting time and brightness acquisition count are dynamically adjusted. Combined with the brightness observation trajectory, camera trigger time and image acquisition sequence are generated to achieve adaptive optimization of detection accuracy and detection cycle.

Benefits of technology

It improves the brightness stability, uniformity, and the completeness and accuracy of detecting edge light leakage defects in backlight components, enhances the stability and consistency of detection results, improves the utilization rate of detection resources, and shortens the detection time.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113318A_ABST
    Figure CN123113318A_ABST
Patent Text Reader

Abstract

This invention discloses a method, system, device, and storage medium for adaptive control of the pace and accuracy of quality inspection of notebook computer components. It relates to the field of adaptive control technology for component quality inspection. The method includes: acquiring a continuous brightness change sequence during the historical backlight assembly lighting process, the initial backlight lighting waiting time, and the number of initial brightness acquisitions to form a first backlight monitoring control pair; generating several second backlight monitoring control pairs through time axis displacement and scaling perturbations, and converting them into a camera trigger time sequence and an image acquisition execution sequence; applying each control pair to backlight assembly quality inspection to obtain inspection results and inspection time; performing iterative training of the strategy to obtain pace and accuracy reward scores; constructing a pace-accuracy response surface and selecting the optimal second backlight monitoring control pair for actual inspection, thus solving the problem of difficulty in simultaneously achieving inspection accuracy and inspection pace in the existing keyboard backlight module quality inspection process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of adaptive control technology for component quality inspection, and more specifically, to a method, system, device, and storage medium for adaptive control of the cycle time and accuracy of quality inspection of notebook computer components. Background Technology

[0002] With the increasing demand for thinner and lighter laptops and higher-quality displays, keyboard backlight modules have become an important component of mid-to-high-end laptops. After the keyboard backlight module is assembled, it is usually necessary to test the backlight assembly by powering it on. Backlight images are captured using an industrial camera, and the uniformity of key brightness, brightness stability, and edge light leakage are analyzed to determine the assembly status of the light guide plate, the LED mounting status, and the integrity of the backlight structure, thereby completing the quality inspection of the keyboard backlight module.

[0003] In the current keyboard backlight module assembly and testing process, the backlight component does not immediately reach a stable light-emitting state after being powered on. Instead, it undergoes a process from initial illumination to a gradual increase in brightness before reaching a stable output. Therefore, the testing system typically needs to set a backlight illumination waiting time and then perform brightness image acquisition after the waiting period. Simultaneously, to obtain information about the brightness change process, the number of brightness acquisitions also needs to be set, forming a brightness change sequence through multiple acquisitions. The backlight illumination waiting time determines the start time of the test, and the number of brightness acquisitions determines the detection coverage; both jointly affect the test results and the test cycle time.

[0004] However, there is no fixed optimal value for the backlight activation wait time and the number of brightness samplings. When the backlight activation wait time and the number of brightness samplings change, the coverage of the brightness establishment process by the detection system also changes. Too high or too low a wait time, or too many or too few samplings, may cause problems such as missing detection information, wasted detection resources, or a decrease in detection cycle time, thereby affecting the accuracy and stability of the keyboard backlight module quality detection results.

[0005] When the backlight illumination wait time is short and the number of brightness samplings is low, the detection process often begins before the backlight brightness has fully established itself, and the sampling coverage is limited. The brightness information acquired at this time mainly focuses on the rapid brightness change phase, making it difficult to cover the subsequent brightness convergence process. This can easily lead to some transient brightness fluctuations being misjudged as brightness anomalies. Furthermore, due to insufficient sampling, some key brightness change characteristics cannot be fully recorded, reducing the accuracy of keyboard backlight module quality detection.

[0006] When the backlight illumination wait time is long and the number of brightness samplings is low, the detection process usually avoids the brightness setup phase, and the sampling content mainly focuses on the brightness stabilization phase. Although this can reduce the interference caused by transient brightness fluctuations, some early brightness anomalies caused by LED welding abnormalities, light guide structure misalignment, etc., cannot be identified because the brightness setup process is not observed. At the same time, the low number of samplings makes it difficult to continuously track the brightness evolution process, affecting the ability to detect abnormal states.

[0007] When the backlight illumination wait time is short and the number of brightness acquisitions is high, the detection system can cover a relatively long brightness change process. However, a large number of acquisitions occur during the phase of rapid brightness change. Because the brightness changes frequently during this phase, continuous acquisition will generate a large amount of dynamic image data, which not only increases the image processing burden but also introduces more transient change information. This requires the detection system to process a large amount of data that is not directly related to the final stable state, thereby reducing detection efficiency.

[0008] When the backlight illumination wait time is long and the number of brightness acquisitions is high, the detection system typically continues to perform a large number of image acquisitions even after the brightness has stabilized. Although this provides abundant brightness information, the differences between subsequent acquired images are small, which easily leads to repeated observations. This results in increased detection resource consumption, prolonged image processing time, and a decrease in the overall detection cycle time, ultimately affecting the detection efficiency of the production line.

[0009] Most existing keyboard backlight module assembly and testing technologies use fixed waiting time and fixed number of samplings for quality inspection. They lack an adaptive optimization mechanism for the synergistic relationship between backlight lighting waiting time and brightness sampling number. They cannot establish a pattern to dynamically adjust the testing strategy based on the actual brightness of the backlight components. This easily leads to problems in balancing testing accuracy and testing cycle time, resulting in defects such as low utilization of testing resources, decreased testing efficiency, and insufficient stability of quality testing results.

[0010] To address the above problems, this invention proposes a solution. Summary of the Invention

[0011] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method, system, device, and storage medium for adaptive control of the quality inspection cycle and accuracy of notebook computer components. By adaptively and collaboratively optimizing the backlight illumination waiting time and the number of brightness acquisitions, and establishing a joint evaluation mechanism for inspection accuracy and inspection cycle, the present invention solves the problem that it is difficult to balance inspection accuracy and inspection cycle in the existing keyboard backlight module quality inspection process.

[0012] To achieve the above objectives, the present invention provides the following technical solution: An adaptive control method for the cycle time and accuracy of quality inspection of notebook computer components.

[0013] An adaptive control system for the cycle time and accuracy of quality inspection of laptop components includes a convergence modeling module, a temporal perturbation module, an observation orchestration module, a brightness inspection execution module, an award and training module, and a superior control selection module. The convergence modeling module constructs a brightness convergence correlation graph based on the continuous brightness change sequence of the historical backlight component lighting process, analyzes and determines the initial backlight lighting waiting time and the number of initial brightness acquisitions, and constructs the first backlight monitoring and control pair. The temporal perturbation module performs displacement and scaling perturbations on the time axis of the first backlight monitoring and control pair to generate several second backlight monitoring and control pairs. The observation orchestration module constructs the brightness observation trajectory corresponding to the backlight monitoring and control pair and converts it into a camera image. The system includes a trigger time sequence and an image acquisition execution sequence; a brightness detection execution module, which converts each second backlight monitoring control pair and applies it to the brightness image acquisition process after the backlight component is powered on, and completes the backlight component quality detection; a reward and training module, which extracts features from the quality detection process of each second backlight monitoring control pair, uses the quality detection results and detection time as rewards, and obtains the corresponding beat and accuracy reward scores through strategy iterative training; and an optimal control selection module, which uses the backlight monitoring control pair as the surface coordinates and the beat and accuracy reward scores as the surface height to map and fit a beat accuracy response surface, and selects the best second backlight monitoring control pair for application to backlight component quality detection.

[0014] An electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform an adaptive control method for quality inspection cycle and accuracy of notebook computer components.

[0015] A computer-readable storage medium storing a computer program that, when executed by a processor, implements an adaptive control method for the cycle time and accuracy of quality inspection of notebook computer components.

[0016] The technical effects and advantages of the present invention regarding the adaptive control method, system, device, and storage medium for quality inspection cycle time and accuracy of notebook computer components are as follows: 1. This invention constructs a brightness convergence correlation graph based on a continuous brightness change sequence during the historical backlight component lighting process. It extracts the initial backlight lighting waiting time and the initial brightness acquisition count from the brightness convergence pattern. Furthermore, it generates several second backlight monitoring and control pairs through time axis displacement and scaling perturbations. These are then combined with the brightness convergence correlation graph to construct the corresponding brightness observation trajectory, thus realizing the conversion between the backlight monitoring and control pair's camera trigger time sequence and the image acquisition execution sequence. Compared to existing technologies that use fixed waiting times and fixed acquisition counts, this invention dynamically organizes the detection sequence based on the actual brightness establishment process of the backlight component. This makes the image acquisition process closer to the critical stage of brightness change, avoiding the problems of insufficient waiting time leading to incomplete brightness establishment and excessive waiting time causing the effective observation interval to be skipped. It also avoids information loss due to too few acquisition counts and redundant acquisition due to too many acquisition counts. Therefore, it improves the completeness and accuracy of detecting backlight component brightness stability, brightness uniformity, and edge light leakage defects.

[0017] 2. This invention applies different second backlight monitoring control pairs to the backlight component quality inspection process, extracts corresponding inspection process features, and constructs a reward signal using the quality inspection results and inspection time. Through iterative strategy training, it obtains the beat rate and accuracy reward score, and further establishes a beat-accuracy response surface between the backlight monitoring control pair and the beat rate and accuracy reward score, achieving joint optimization of inspection accuracy and inspection beat rate. Compared to existing technologies that focus only on single-objective optimization methods that focus only on inspection accuracy or only on inspection efficiency, this invention can simultaneously characterize the comprehensive impact of backlight illumination waiting time and brightness acquisition count on inspection quality and inspection time. Under the premise of meeting the actual inspection cycle and equipment acquisition load requirements, it selects the optimal second backlight monitoring control pair, thereby achieving an adaptive balance between inspection accuracy and inspection beat rate, improving inspection resource utilization, shortening inspection time, and enhancing the stability and consistency of backlight component quality inspection results. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the adaptive control method for the quality inspection cycle and accuracy of notebook computer components according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1, Figure 1The present invention provides an adaptive control method for the cycle time and accuracy of quality inspection of notebook computer components, comprising the following steps: S1. Construct a brightness convergence correlation diagram based on the continuous brightness change sequence of the historical backlight component lighting process, and analyze and determine the initial backlight lighting waiting time and the number of initial brightness acquisitions to construct the first backlight monitoring and control pair.

[0021] In this embodiment, the step of constructing a brightness convergence correlation graph based on the continuous brightness change sequence of the historical backlight component lighting process, and analyzing and determining the initial backlight lighting waiting time and the number of initial brightness acquisitions to construct the first backlight monitoring and control pair, specifically involves: Collect the continuous brightness change sequence of the historical backlight component from the start of power-on to the end of brightness stabilization, and divide it into several continuous time windows according to a preset time interval to obtain the brightness distribution state corresponding to each time window. The brightness of the keys in each brightness distribution state is classified into levels, and the corresponding key brightness level sequence is generated. Count the number of keys that maintain the same brightness level in the key brightness level sequence corresponding to adjacent time windows, and form a level maintenance sequence according to time order; Based on the grade preservation sequence, the correlation between time windows is constructed, adjacent time windows are used as correlation nodes, and the corresponding grade preservation number is used as the correlation strength between nodes to generate a brightness convergence correlation graph. In the brightness convergence correlation graph, identify the correlation node chains with continuously increasing correlation strength, and determine the time segments corresponding to each correlation node chain as candidate brightness convergence segments. The cumulative value of correlation strength within each brightness convergence candidate segment is calculated, and the brightness convergence candidate segment with the largest cumulative value is determined as the target brightness convergence segment. Extract the starting time position corresponding to the target brightness convergence segment, and determine the starting time position corresponding to the target brightness convergence segment as the initial backlight illumination waiting time; The number of time windows covered by the target brightness convergence segment is counted, and the number of time windows covered by the target brightness convergence segment is determined as the initial number of brightness acquisitions; The initial backlight illumination waiting time and the number of initial brightness sampling times are combined to form the first backlight monitoring and control pair.

[0022] In this embodiment, the continuous brightness change sequence is a time-series data set showing the continuous change in brightness of each key area over time from the moment the backlight assembly is powered on until its brightness stabilizes. The continuous brightness change sequence not only reflects the overall brightness increase process but also the differences in brightness establishment speed between different keys. For example, some keys may reach stable brightness earlier due to differences in their light guide structures, while others may experience a brightness lag. Therefore, the continuous brightness change sequence can comprehensively characterize the dynamic change process of the backlight assembly from startup to stability.

[0023] In this embodiment, the preset time interval is a uniform time division length used when segmenting a continuous brightness change sequence. This preset time interval is determined based on the sampling frequency of the backlight component's illumination process, ensuring that adjacent time windows can reflect the brightness change trend while avoiding the loss of brightness change details due to excessively large spans of individual time windows. For example, when the backlight component's brightness change process lasts for hundreds of milliseconds, it can be continuously divided according to a fixed millisecond time length to form a continuous time window.

[0024] In this embodiment, the brightness distribution state refers to the brightness distribution of all key areas within the corresponding time window. When acquiring the brightness distribution state, the brightness values ​​of all key areas within the current time window are summarized, and a brightness distribution structure is established according to the key positions, thereby reflecting the spatial distribution characteristics of the brightness of the entire keyboard backlight area within the time window.

[0025] In this embodiment, the key brightness level sequence is a ranking arrangement based on the brightness of each key. Specifically, all key brightness levels are divided into multiple levels according to a unified standard, and the corresponding levels are recorded in order of key position, thus forming the key brightness level sequence. For example, keys with higher brightness are assigned to higher-level areas, and keys with lower brightness are assigned to lower-level areas. Using level representation instead of direct brightness value representation reduces the impact of instantaneous brightness fluctuations on subsequent analysis.

[0026] In this embodiment, the brightness level maintenance sequence is a sequence formed by the number of keys whose brightness level remains unchanged between adjacent time windows, arranged chronologically. Specifically, the brightness level changes of corresponding keys in adjacent time windows are compared one by one, the number of keys whose brightness level did not change is counted, and these are arranged sequentially according to time. As the number of keys maintaining a stable brightness level gradually increases, it indicates that more and more keys are approaching a stable state.

[0027] In this embodiment, the associated nodes are time window representation units in the brightness convergence correlation graph. Each associated node corresponds to the brightness distribution state of a time window, and the nodes are connected in chronological order, thereby transforming the discrete time window into an association structure that can reflect the brightness evolution process.

[0028] In this embodiment, the association strength is a characterization value representing the degree to which the brightness state of adjacent associated nodes remains consistent. The association strength is derived from the number of keys that maintain the same brightness level between corresponding time windows. When the number of keys maintaining the same brightness level is large, it indicates that the brightness change between the two time windows is small, and the corresponding association strength is high; when the number of keys maintaining the same brightness level is small, it indicates that the brightness is still in a rapid change phase, and the corresponding association strength is low.

[0029] In this embodiment, the brightness convergence correlation graph is a temporal correlation structure constructed using correlation nodes and correlation strengths. This correlation graph is used to describe the gradual transition of the backlight component's brightness from a rapidly changing stage to a stable stage, and key time segments in the brightness convergence process are identified by observing the trend of correlation strength changes.

[0030] In this embodiment, the associated node chain is a set of nodes connected sequentially in the brightness convergence correlation graph. The associated node chain reflects the continuous evolution trajectory during the brightness change process. When the correlation strength in the associated node chain continues to increase, it indicates that the brightness state is maintained at a continuously improving level, and the backlight component gradually enters a stable state.

[0031] In this embodiment, the brightness convergence candidate segment is the time range corresponding to the associated node chain. Specifically, it is the continuous time segment covered during the period when the association strength continuously increases, and this time segment characterizes the convergence process of the backlight component transitioning from an unstable state to a stable state.

[0032] In this embodiment, the brightness level division method is based on a preset brightness level scale. The brightness level scale is a discrete division system that is uniformly applicable to all key areas. This system divides the brightness values ​​in the brightness distribution state of the key into a fixed number of level intervals in order from low to high. Each level interval corresponds to a clear brightness range boundary, and the same key corresponds to only one brightness level identifier within the same time window. This ensures that the brightness state between different time windows and between different keys has a consistent alignment standard, which is used to support the stable calculation process of the subsequent level maintenance quantity and correlation strength.

[0033] In this embodiment, the starting point of the time calculation is consistent with the starting point of the continuous brightness change sequence. That is, the moment when the backlight component is connected to the detection power supply and receives the power-on trigger signal is taken as the unified zero point. This zero point corresponds to the starting boundary of the first time window in the continuous brightness change sequence. All subsequent time windows are expanded in sequence according to a preset time interval based on this zero point, thereby ensuring that all related nodes in the brightness convergence correlation graph are constructed based on the same time reference, avoiding convergence segment deviation caused by time drift between different test batches.

[0034] In this embodiment, the determination of brightness convergence candidate segments is based on the structurally stable change characteristics of continuous time windows. Specifically, when the correlation strength between multiple adjacent time windows shows a continuously progressively increasing state, and the number of brightness-consistent keys within this continuous time range does not show a significant decline compared to the previous time window, the set of continuous time windows is marked as a candidate convergence segment. The continuously progressively increasing state is manifested in the number of brightness-consistent keys between adjacent windows gradually increasing or remaining at or above the level of the previous window in the continuous window sequence. This eliminates false convergence intervals caused by single-point fluctuations or local anomalies, enabling the candidate segments to stably reflect the true process of the backlight component transitioning from the changing stage to the stable stage. In this embodiment, the cumulative correlation strength value is the overall representation result formed by accumulating all correlation strengths within the brightness convergence candidate segment in chronological order. The larger the cumulative correlation strength value, the more keys with consistent brightness within the corresponding time segment, and the higher the degree of brightness convergence. Therefore, it can be used to identify the most representative brightness convergence segment.

[0035] In this embodiment, the target brightness convergence segment is the brightness convergence candidate segment with the largest cumulative correlation intensity value. This segment corresponds to the time range during which the backlight component transitions from dynamic changes to a stable state most fully, and can simultaneously reflect the brightness establishment process and the brightness stabilization process. Therefore, it is selected as the reference segment for subsequent observation and control.

[0036] In this embodiment, the initial backlight illumination waiting time is the starting time position corresponding to the target brightness convergence segment. This time position represents the length of time elapsed from the backlight component being powered on to entering the target brightness convergence segment, thus providing an initial waiting reference for subsequent brightness detection.

[0037] In this embodiment, the initial number of brightness acquisitions is the number of time windows covered by the target brightness convergence segment. In specific implementation, the number of time windows contained within the target brightness convergence segment is used as a representation of the brightness observation coverage, thereby determining the number of initial image acquisitions required in the subsequent brightness detection stage.

[0038] In this embodiment, the first backlight monitoring and control unit is a control unit formed by combining the initial backlight illumination waiting time and the initial brightness sampling count. The initial backlight illumination waiting time is used to determine the start time of detection, and the initial brightness sampling count is used to determine the detection observation coverage. Together, they determine the cycle time and detection accuracy in the subsequent backlight quality detection process.

[0039] It should be noted that the brightness convergence correlation diagram can not only be used to identify the time when the backlight component enters a stable state, but also to reflect the differences in the brightness establishment process between different batches of backlight components. When the backlight component structure changes or the performance of the light guide component changes, the correlation intensity distribution in the brightness convergence correlation diagram will also change synchronously, thus providing a unified data foundation for subsequent monitoring and control optimization.

[0040] It should be noted that the target brightness convergence segment does not directly correspond to the time period after the brightness has completely stabilized, but rather to the time period during which the brightness gradually approaches a stable state and the degree of stability continuously increases. By selecting this time segment as the control basis, effective change information during the brightness establishment process and information from the stable phase can be preserved simultaneously, thereby improving the adaptability of the subsequent control optimization process.

[0041] It should be noted that this embodiment identifies key convergence segments in the backlight assembly brightness establishment process by constructing a brightness convergence correlation diagram, and uses the target brightness convergence segment to determine the initial backlight lighting waiting time and the initial brightness sampling number, avoiding the direct use of fixed waiting time or fixed sampling number for detection, so that the initial monitoring and control parameters can match the actual backlight brightness establishment process, providing a reliable initial control basis for subsequent cycle time and accuracy co-optimization.

[0042] S2, perform time axis displacement and scaling disturbance on the first backlight monitoring and control pair to generate several second backlight monitoring and control pairs.

[0043] In this embodiment, the process of performing time-axis displacement and scaling disturbances on the first backlight monitoring and control pair to generate several second backlight monitoring and control pairs specifically involves: Extract the target brightness convergence segment corresponding to the first backlight monitoring and control pair, and obtain the start time position and end time position corresponding to the target brightness convergence segment; Construct a convergence segment time axis based on the start and end time positions of the target brightness convergence segment, and generate several time offsets according to a preset time step. Each time offset is superimposed onto the start and end time positions of the target brightness convergence segment, and a time axis displacement perturbation is performed on the target brightness convergence segment to generate several displacement convergence segments. The center position of each displacement convergence segment is obtained respectively, and the center position of each segment is used to replace the original center position of the target brightness convergence segment to generate several candidate observation center positions. Based on the center position of each displacement convergence segment, the time span of the displacement convergence segment is expanded and compressed according to a preset scaling ratio to generate several scaling convergence segments. The number of time windows covered by each expansion and contraction segment is counted, and the corresponding candidate observation coverage length is formed; Convert the time position corresponding to each candidate observation center location into the second backlight illumination waiting time; Convert the number of time windows corresponding to the coverage length of each candidate observation into the number of second brightness acquisitions; By combining the second backlight illumination waiting time and the second brightness acquisition number according to the one-to-one correspondence, several second backlight monitoring and control pairs are generated.

[0044] In this embodiment, the target brightness convergence segment is the time segment with the highest degree of brightness convergence determined in step S1. This time segment corresponds to the stage where the brightness of the backlight component transitions most fully from dynamic change to a stable state. Its starting time position indicates the time point when the brightness begins to enter the stable convergence process, and its ending time position indicates the time point when the brightness convergence process ends and approaches a stable state. Therefore, the target brightness convergence segment can serve as the reference segment for subsequent observation and control range adjustment.

[0045] In this embodiment, the convergence segment time axis is a time reference coordinate system established with the start and end times of the target brightness convergence segment as boundaries. This time axis is used to describe the positional relationship of the target brightness convergence segment throughout the lighting process and to provide a unified time reference standard for subsequent displacement and scaling perturbations.

[0046] In this embodiment, the preset time step is a uniform unit of movement used when performing time movement on the convergence segment time axis. Specifically, the time step is determined based on the temporal resolution during the historical backlight component illumination process, ensuring that each displacement covers different brightness change stages. For example, the time step can correspond to one or more consecutive time windows, thereby guaranteeing that the generated candidate observation schemes can cover different time positions before and after the target brightness convergence segment.

[0047] In this embodiment, the time offset is a time movement amount generated based on a preset time step. Specifically, using the current position of the target brightness convergence segment as a reference, multiple time positions are extended forward or backward according to the preset time step, forming multiple different time offsets. Different time offsets correspond to different observation start times, thus enabling the subsequently generated control scheme to cover various situations such as early observation, synchronous observation, and delayed observation.

[0048] In this embodiment, the time axis displacement perturbation is a process that keeps the time span of the target brightness convergence segment unchanged, only changing its position on the time axis. In specific implementation, the time offset is applied simultaneously to the start and end time positions of the target brightness convergence segment, causing the entire segment to move along the time axis as a whole, while the duration within the segment remains unchanged, thereby forming observation candidate segments at different time positions.

[0049] In this embodiment, the displacement convergence segment is a new time segment formed after the target brightness convergence segment is perturbed by the time axis displacement. Each displacement convergence segment has the same duration as the target brightness convergence segment, but corresponds to a different time position, thus reflecting the difference in the effect of performing brightness observations under different waiting conditions.

[0050] In this embodiment, the center position of the segment is the midpoint between the start and end times of the corresponding time segment. This center position is used to characterize the representative time position of the entire observation segment during the lighting process, enabling observation segments with different time spans to be compared and managed using a unified position representation method.

[0051] In this embodiment, the candidate observation center position is the set of time positions formed by the center positions of each displacement convergence segment. Different candidate observation center positions correspond to different brightness observation opportunities, which can be used to characterize the positional changes of the observation process throughout the entire backlight illumination process.

[0052] In this embodiment, the preset scaling ratio is the adjustment ratio used when expanding or compressing the duration of the displacement convergence segment. In specific implementation, based on the current duration of the displacement convergence segment, segment expansion or compression is performed according to multiple different scaling ratios, so that the generated observation segments include both observation schemes with large coverage and observation schemes with small coverage, thereby forming control combinations under different observation densities.

[0053] In this embodiment, the time span is the duration range covered between the start and end times of the corresponding convergence segment. A larger time span indicates that the observation process covers more brightness change stages; a smaller time span indicates that the observation process is more concentrated on local brightness change stages.

[0054] In this embodiment, the expansion / contraction convergence segment is a new time segment formed by expanding or compressing the displacement convergence segment over time. When expansion is performed, the observation segment covers more brightness changes; when compression is performed, the observation segment focuses on brightness changes within a shorter time range, thus forming candidate schemes with different observation coverage ranges.

[0055] In this embodiment, the candidate observation coverage length is the observation length characterization result corresponding to the time window range covered by the scaling convergence segment. In specific implementation, the observation coverage length is determined according to the number of time windows covered by the scaling convergence segment; the more time windows covered, the larger the corresponding observation coverage length; the fewer time windows covered, the smaller the corresponding observation coverage length.

[0056] In this embodiment, the second backlight illumination waiting time is the waiting time formed by the time position conversion corresponding to the candidate observation center position. In specific implementation, the moment when the backlight component is powered on is taken as the starting point, and the candidate observation center position is taken as the target time position. The corresponding time length between the two is the second backlight illumination waiting time, which is used to determine the timing when the brightness observation is started in the actual detection process.

[0057] In this embodiment, the second brightness acquisition count is the acquisition count converted from the number of time windows corresponding to the candidate observation coverage length. In specific implementation, the number of brightness acquisition actions generated corresponds to the number of consecutive time windows covered by the expansion and contraction convergence segment, thereby ensuring that the observation coverage range and the acquisition count are consistent.

[0058] In this embodiment, the second backlight monitoring and control pair is a control unit formed by combining the second backlight illumination waiting time and the second brightness acquisition count. Each second backlight monitoring and control pair corresponds to a specific brightness observation scheme, wherein the second backlight illumination waiting time is used to control the detection start timing, and the second brightness acquisition count is used to control the detection coverage area. Together, they determine the detection cycle and detection accuracy in the subsequent quality detection process.

[0059] It should be noted that the time axis displacement perturbation is mainly used to change the timing of intervention in the brightness observation process during the backlight illumination process, so that the observation scheme can cover different time regions such as the rapid brightness change stage, the brightness transition stage, and the brightness stabilization stage, thereby improving the adaptability of the subsequent optimization process to different detection scenarios.

[0060] It should be noted that the scaling perturbation is mainly used to change the observation coverage, rather than the observation starting position. By simultaneously retaining both the extended and compressed observation schemes, different detection strategies, ranging from few to many samples, can be covered, providing a richer candidate control space for subsequent optimization of the balance between cycle time and accuracy.

[0061] It should be noted that this embodiment generates a second set of backlight monitoring and control pairs covering different observation times and different observation ranges by performing time axis displacement perturbation and scaling perturbation on the basis of the target brightness convergence section. This allows the subsequent detection process to form a sufficient strategy exploration space in two dimensions: waiting time and number of acquisitions, thereby providing diversified candidate control schemes for the coordinated optimization of timing and accuracy.

[0062] S3 constructs the corresponding brightness observation trajectory for the backlight monitoring and control pair, and transforms it into a camera trigger time sequence and an image acquisition execution sequence.

[0063] In this embodiment, the construction of the backlight monitoring and control pair corresponding to the brightness observation trajectory, and its transformation into a camera trigger time sequence and an image acquisition execution sequence, specifically involves: Extract the waiting time for the second backlight to turn on and the number of times the second brightness is collected for each second backlight monitoring and control pair. The initial brightness observation trajectory is constructed by taking the time position corresponding to the second backlight illumination waiting time as the observation start node and the number of time windows corresponding to the second brightness acquisition times as the number of observation coverage nodes. Based on the time window covered by the initial brightness observation trajectory, the associated nodes of the corresponding time window in the brightness convergence correlation graph are extracted, and the observation node chain is formed in chronological order. The number of levels maintained by each associated node in the observation node chain is counted, and the proportion of the number of levels maintained by each associated node to the total number of levels maintained by the observation node chain is calculated to form a node maintenance proportion sequence. The time interval between adjacent observation nodes in the observation node chain is redistributed based on the node retention ratio sequence. The time interval between adjacent observation nodes is shortened for associated nodes with a higher retention ratio, and the time interval between adjacent observation nodes is extended for associated nodes with a lower retention ratio. The time position of each observation node is updated according to the reallocated time interval to generate an optimized brightness observation trajectory. Extract the time position corresponding to each observation node along the optimized brightness observation trajectory, and form a set of observation times in chronological order; Convert each time location in the observation time set into a camera trigger time to generate a camera trigger time sequence; The image acquisition execution sequence is generated according to the sequential order of the camera trigger time sequence, thereby realizing the conversion of the second backlight monitoring and control towards the actual detection execution action.

[0064] In this embodiment, the second backlight illumination waiting time is a time control quantity in the second backlight monitoring and control pair, representing the waiting time experienced by the backlight component from power-on to the start of brightness observation. Different second backlight illumination waiting times correspond to different observation intervention timings. When the waiting time is short, the observation process covers more of the brightness establishment phase; when the waiting time is long, the observation process is closer to the brightness stabilization phase. Therefore, it directly determines the starting position of the brightness observation trajectory in the entire illumination process.

[0065] In this embodiment, the second brightness acquisition count is the observation control quantity in the second backlight monitoring and control pair, which represents the number of brightness image acquisition actions that need to be performed in one backlight detection process. The more second brightness acquisition counts, the more complete the brightness change process is covered; the fewer second brightness acquisition counts, the more concentrated the observation process is, and therefore it determines the time range covered by the brightness observation trajectory.

[0066] In this embodiment, the observation starting node is the first brightness observation position determined by the second backlight illumination waiting time. Specifically, the moment the backlight component is powered on is taken as the time starting point, the corresponding time window is determined according to the second backlight illumination waiting time, and the associated node corresponding to this time window is taken as the observation starting node, thereby determining the starting position of the entire observation process.

[0067] In this embodiment, the number of observation coverage nodes is the total number of observation nodes determined by the number of second brightness acquisitions. In specific implementation, starting from the observation start node, a corresponding number of consecutive time windows are selected sequentially according to time order, so that the observation process covers the same number of observation nodes as the number of second brightness acquisitions.

[0068] In this embodiment, the initial brightness observation trajectory is an observation path established based on the observation starting node and the number of observation coverage nodes. This trajectory describes the coverage area and observation sequence of the brightness observation process in the brightness convergence correlation graph, and is used to characterize the original observation scheme corresponding to the second backlight monitoring and control pair.

[0069] In this embodiment, the observation node chain is a continuous set of nodes formed in chronological order by the associated nodes covered by the initial brightness observation trajectory. Each node in the observation node chain corresponds to the brightness distribution state of a time window, and the entire node chain reflects the brightness evolution process covered by the current observation scheme.

[0070] In this embodiment, the level maintenance number refers to the number of keys that maintain the same brightness level between the corresponding associated node and its preceding associated node. Specifically, the changes in key brightness levels are compared one by one in adjacent time windows, and the number of keys whose brightness levels remain unchanged is counted. A larger level maintenance number indicates smaller brightness changes at the corresponding time position, and the brightness state is closer to convergence.

[0071] In this embodiment, the node retention percentage sequence represents the relative distribution of the retention quantity of each associated node at its corresponding level within the entire observation node chain. Specifically, the retention quantities of all levels in the observation node chain are aggregated, and then the proportion of each associated node's retention quantity within the overall sequence is determined. This is then arranged chronologically to form the node retention percentage sequence. A higher node retention percentage indicates a stronger brightness convergence characteristic at the corresponding time position, while a lower node retention percentage indicates that there are still significant brightness changes at the corresponding time position.

[0072] In this embodiment, the time interval between adjacent observation nodes is the time span between the corresponding time positions of two adjacent observation nodes in the observation node chain. This time interval is used to determine the time distance between two consecutive image acquisition actions, thereby determining the distribution of observation resources throughout the brightness change process.

[0073] In this embodiment, the reallocation of time intervals is a process of adjusting the temporal distribution within the observation node chain based on the node retention rate sequence. Specifically, more observation opportunities are allocated around associated nodes with higher retention rates, reducing the temporal distance between adjacent observation nodes; conversely, fewer observation opportunities are allocated around associated nodes with lower retention rates, increasing the temporal distance between adjacent observation nodes. This method concentrates observation resources in areas with higher brightness convergence.

[0074] In this embodiment, the optimized brightness observation trajectory is the new observation trajectory formed after the time interval is redistributed. Compared with the initial brightness observation trajectory, the optimized brightness observation trajectory still maintains the same observation starting position and number of observations, but the distribution of observation nodes on the time axis is adjusted, thereby forming an observation layout that is more in line with the brightness convergence characteristics.

[0075] In this embodiment, the observation time set is a time set formed by the time positions of all observation nodes in the optimized brightness observation trajectory in chronological order. The observation time set is used to record all time positions of the actual image acquisition action and serves as the basis data for subsequent detection execution control.

[0076] In this embodiment, the camera trigger time is the camera control command execution time formed by the specific time position transformation within the observation time set. Specifically, when the system reaches the corresponding time position, it sends a trigger signal to the industrial camera, enabling the industrial camera to complete the brightness image acquisition for the corresponding time.

[0077] In this embodiment, the camera trigger time sequence is a control sequence formed by arranging all camera trigger times in chronological order. This sequence is used to guide the industrial camera to perform the acquisition task in a predetermined time sequence throughout the entire backlight illumination process.

[0078] In this embodiment, the image acquisition execution sequence is a set of image acquisition action sequences generated based on the camera trigger time sequence. The image acquisition execution sequence not only records the execution order of the acquisition actions, but also records the time position corresponding to each acquisition action, thereby realizing the conversion of the detection strategy into device execution actions.

[0079] It should be noted that the node-preservation ratio sequence not only reflects the degree of brightness convergence at different time locations, but also reflects the value distribution of observed resources during brightness changes. Therefore, this embodiment uses the node-preservation ratio sequence to adjust the distribution of observed nodes, so that a limited number of acquisitions prioritizes coverage of time regions with more brightness convergence characteristics.

[0080] It should be noted that optimizing the brightness observation trajectory does not change the corresponding waiting time and number of acquisitions for the second backlight monitoring control pair. It only changes the distribution of each acquisition action on the time axis under the same number of acquisitions, thereby ensuring that the subsequent evaluation process can simultaneously reflect the impact of the observation timing and observation distribution on the detection results.

[0081] It should be noted that this embodiment transforms the abstract second backlight monitoring control pair into a specific executable camera trigger time sequence and image acquisition execution sequence by constructing an initial brightness observation trajectory, an observation node chain, and optimizing the brightness observation trajectory. This achieves a direct mapping from the detection strategy to the device action, enabling the subsequent quality detection process to truly reflect the differences in pace and accuracy under different combinations of waiting time and number of acquisitions.

[0082] S4 converts and applies each of the second backlight monitoring and control pairs to the brightness image acquisition process after the backlight component is powered on, and completes the backlight component quality inspection.

[0083] In this embodiment, the process of converting and applying each of the second backlight monitoring and control pairs to the brightness image acquisition process after the backlight assembly is powered on, and completing the backlight assembly quality detection, specifically involves: Obtain the camera trigger time sequence and image acquisition execution sequence corresponding to each second backlight monitoring and control pair; Control the backlight assembly to connect to the detection power supply and execute the power-on lighting, establishing the corresponding brightness change process; The camera is controlled to perform image acquisition at the corresponding trigger time according to the camera trigger time sequence, and the corresponding brightness image set is obtained according to the image acquisition execution sequence. The brightness image set is sorted chronologically according to the order of image acquisition execution sequence to generate a continuous brightness image sequence. Brightness extraction is performed on each key region in a continuous brightness image sequence to generate a key brightness variation sequence; Based on the key brightness change sequence, the brightness change amplitude of each key during continuous acquisition is statistically analyzed to generate key brightness stability results; Based on the continuous brightness image sequence, the brightness difference between adjacent keys is statistically analyzed to generate the brightness uniformity result of the keys; The brightness distribution of the key edge region is extracted from the continuous brightness image sequence to generate edge light leakage detection results; By integrating the key brightness stability results, key brightness uniformity results, and edge light leakage detection results, the backlight component quality detection results corresponding to the second backlight monitoring and control pair are generated. Record the detection time of the corresponding second backlight monitoring and control from the power-on of the backlight component to the acquisition of the backlight component quality detection result.

[0084] In this embodiment, the camera trigger time sequence is the camera control timing set generated in step S3. This sequence records each specified acquisition time position after the backlight component is powered on, and is used to control the industrial camera to perform image acquisition actions at the corresponding time positions, thereby ensuring that different second backlight monitoring and control pairs complete the acquisition of brightness information according to their respective observation strategies.

[0085] In this embodiment, the image acquisition execution sequence is the order in which acquisition actions are executed, established based on the camera trigger time sequence. The image acquisition execution sequence not only includes the order in which each image acquisition action is executed, but also records the brightness change stage at which each acquisition action occurs, thereby ensuring that the brightness images acquired subsequently can reflect the complete brightness evolution process.

[0086] In this embodiment, the detection power supply is a standard test power supply device used to drive the backlight assembly to light up. In specific implementation, the detection power supply provides a stable voltage and current to the backlight assembly according to predetermined power supply conditions, enabling the backlight assembly to enter a normal lighting state, thereby establishing a brightness change process consistent with the actual production testing environment.

[0087] In this embodiment, the brightness change process is a dynamic brightness evolution process formed from the moment the backlight component is powered on until the brightness tends to stabilize. During this process, the brightness of each key area will go through multiple stages such as brightness establishment, brightness transition, and brightness stabilization. Therefore, the brightness change process can reflect the true light-emitting characteristics of the backlight component.

[0088] In this embodiment, the brightness image set is an image set consisting of all brightness images acquired by the industrial camera based on the camera's trigger time sequence. Each image in the brightness image set corresponds to a specific time position in the brightness change process, so the entire image set can completely record the state of the backlight component's brightness change over time.

[0089] In this embodiment, the continuous brightness image sequence is a set of brightness images arranged in the order of the image acquisition execution sequence. The continuous brightness image sequence maintains the temporal continuity of the brightness change process, enabling subsequent analysis to accurately track the evolution trajectory of brightness in each key region.

[0090] In this embodiment, the key area refers to the light-emitting area corresponding to each key position in the keyboard backlight assembly. Specifically, the detection area corresponding to each key is pre-divided according to the keyboard layout, and the same area is consistently used for brightness analysis in subsequent image processing to ensure the comparability of data from different acquisition times.

[0091] In this embodiment, brightness extraction is a process of obtaining brightness information of each key area from a continuous brightness image sequence. Specifically, the luminous intensity distribution of each key area is read and converted into a key brightness record at the corresponding time position, thereby establishing a data basis for the change of key brightness over time.

[0092] In this embodiment, the key brightness change sequence is a record of brightness changes in each key area during continuous acquisition. This sequence records the brightness state of each key at each acquisition moment in chronological order, thus reflecting the changing patterns of the key brightness establishment and stabilization processes.

[0093] In this embodiment, the brightness variation range represents the brightness change of the same key during continuous data acquisition. Specifically, by comparing the brightness state of the key at different acquisition times, the brightness variation range of that key within the entire detection cycle is formed. A larger brightness variation range indicates lower brightness stability of the key, while a smaller range indicates more stable brightness.

[0094] In this embodiment, the key brightness stability result is a stability evaluation result based on the brightness variation range of all keys. In specific implementation, the brightness variation of all keys is summarized, and a detection result reflecting the overall brightness stability of the backlight assembly is formed, which is used to characterize the stable light emission capability of the backlight assembly during the lighting process.

[0095] In this embodiment, the brightness difference between adjacent keys refers to the difference in brightness between two adjacent key areas at the same sampling time. Since the light guide structure and light emission distribution of the backlight assembly affect the brightness consistency between different keys, the brightness difference between adjacent keys can reflect the degree of uniformity in the brightness distribution of the backlight assembly.

[0096] In this embodiment, the key brightness uniformity result is a uniformity evaluation result based on the brightness differences of all adjacent keys. In specific implementation, the brightness differences between all adjacent keys in the keyboard area are summarized and analyzed to form a detection result reflecting the overall balance of brightness distribution.

[0097] In this embodiment, the key edge region is the detection area formed by the keycap edge, the light guide structure edge, and the transition area between adjacent keys. This area is prone to light leakage, light guide misalignment, and abnormal light emission caused by structural assembly errors, and is therefore designated as a key area for light leakage detection.

[0098] In this embodiment, the brightness distribution in the edge region refers to the spatial distribution of brightness within the key edge region. By analyzing the changes in brightness in the edge region at different locations, it is possible to identify whether there are abnormal light concentrations or diffusions.

[0099] In this embodiment, the edge light leakage detection result is based on the light leakage detection result formed by the brightness distribution of the edge region. In specific implementation, when a brightness concentration area that significantly deviates from the normal light emission distribution appears in the edge region, it is identified as a light leakage phenomenon, and the corresponding light leakage detection result is recorded.

[0100] In this embodiment, the backlight assembly quality inspection result is a comprehensive inspection result formed by fusing the key brightness stability result, the key brightness uniformity result, and the edge light leakage detection result. This result is used to reflect the overall quality status of the backlight assembly obtained by the current second backlight monitoring and control system.

[0101] In this embodiment, the detection time is the total detection time from the moment the backlight component is connected to the detection power supply until the corresponding backlight component quality detection result is obtained. The detection time includes not only the backlight lighting waiting stage, but also the image acquisition stage and the quality analysis stage, thus reflecting the detection cycle time level corresponding to the current detection scheme.

[0102] It should be noted that the continuous brightness image sequence is not only used to obtain the final quality inspection result, but also retains the dynamic change information during the brightness establishment process of the backlight component. Compared with the detection method that only acquires a single stable image, this embodiment can identify more potential quality defects by utilizing the temporal information during the brightness change process.

[0103] It should be noted that the key brightness stability results, key brightness uniformity results, and edge light leakage detection results correspond to the three detection dimensions of backlight component quality: temporal stability, spatial consistency, and structural integrity, respectively. By retaining the detection results of all three dimensions simultaneously, misjudgments caused by a single detection indicator can be avoided.

[0104] It should be noted that this embodiment applies the observation strategy corresponding to the second backlight monitoring control to the brightness image acquisition process after the backlight component is powered on, and simultaneously acquires the backlight component quality detection results and detection time, so that the subsequent cycle and accuracy optimization process can establish an evaluation basis based on real detection data, thereby accurately reflecting the influence of different waiting time and acquisition number combinations on detection efficiency and detection quality.

[0105] S5 extracts features from the quality detection process of each second backlight monitoring and control pair, uses the quality detection results and detection time as rewards, and obtains the corresponding beat and accuracy reward scores through strategy iterative training.

[0106] In this embodiment, the quality detection process for each second backlight monitoring and control pair involves feature extraction. The quality detection results and detection time are used as rewards, and corresponding beat and accuracy reward scores are obtained through iterative strategy training. Specifically: Read the continuous brightness image sequence, key brightness stability result, key brightness uniformity result, edge light leakage detection result, backlight component quality detection result, and detection time corresponding to each second backlight monitoring and control pair respectively; Based on the continuous brightness image sequence, the brightness change features of the key positions at each acquisition time are extracted, and a brightness convergence feature group is constructed according to the brightness difference between adjacent acquisition times. Based on the key brightness stability results, the distribution characteristics of brightness variation amplitude of each key are extracted to form a stability feature group; Based on the results of key brightness uniformity, the distribution features of brightness differences between keys are extracted to form a uniformity feature group; Based on the edge light leakage detection results, the abnormal brightness distribution characteristics of the edge region are extracted to form a light leakage feature group; Based on the detection time consumption, the time consumption features from power-on lighting to the completion of quality detection are extracted to form a cycle time feature group; The brightness convergence feature group, stability feature group, uniformity feature group, light leakage feature group and rhythm feature group are combined in a fixed order to form a detection process feature vector that corresponds one-to-one with each second backlight monitoring and control pair. The detection process feature vector is used as the state of the reinforcement learning environment, the corresponding second backlight monitoring and control pair is used as the action input, and the backlight component quality detection result and detection time are used to jointly construct the reward signal to generate the backlight detection reinforcement learning training environment. Based on each second backlight monitoring and control, multiple rounds of strategy iteration training are performed on the corresponding backlight detection process, so that the reinforcement learning model learns the influence relationship between the backlight lighting waiting time and the number of brightness acquisitions on the detection quality and detection cycle, and obtains the cycle accuracy reward generation strategy. Input the corresponding detection process feature vector of each second backlight monitoring and control pair into the beat accuracy reward generation strategy, and output the corresponding beat and accuracy reward score; The corresponding beat and accuracy bonus points for each second backlight monitoring and control pair are summarized to form a bonus distribution sequence.

[0107] In this embodiment, the continuous brightness image sequence is a set of brightness images arranged in the order of image acquisition in step S4. This sequence completely preserves the brightness evolution process of the backlight component from the start of illumination to the end of detection. Different acquisition times correspond to different brightness states, thus reflecting the dynamic changes in the brightness establishment and stabilization process of the backlight component.

[0108] In this embodiment, the key brightness change feature is the characterization result of the key brightness change pattern extracted from a continuous brightness image sequence. In specific implementation, the brightness state of the same key at different acquisition times is read sequentially along time, and feature information reflecting the brightness change trend, the continuity of brightness change, and the degree of concentration of brightness change is formed, which is used to describe the dynamic characteristics of the backlight component brightness establishment process.

[0109] In this embodiment, the brightness convergence feature set is a set of features formed based on the characteristics of key brightness changes. Specifically, by observing the brightness changes between adjacent acquisition times, the process of gradually weakening brightness changes and gradually converging brightness states is recorded, thereby forming a feature set used to characterize the brightness convergence process. The brightness convergence feature set can reflect the degree of coverage of the current observation scheme on the brightness establishment process.

[0110] In this embodiment, the stability feature set is a set of features formed based on the key brightness stability results. Specifically, the distribution of brightness variation amplitude of each key across the entire keyboard area is analyzed to form feature information characterizing the degree of brightness stability. The stability feature set reflects the backlight assembly's luminous stability during continuous illumination.

[0111] In this embodiment, the key brightness variation amplitude distribution feature refers to the distribution of the brightness variation amplitude of each key in the keyboard area. This feature not only reflects the magnitude of the brightness variation amplitude but also whether the brightness variation is concentrated in a local area, thus describing the overall stability of the backlight assembly.

[0112] In this embodiment, the uniformity feature set is a set of features formed based on the brightness uniformity results of the key positions. Specifically, the brightness differences between different key positions are summarized to form feature information characterizing the consistency of the spatial distribution of brightness. The uniformity feature set can reflect the degree of balance between the light guide structure and the light-emitting structure of the backlight assembly.

[0113] In this embodiment, the key brightness difference distribution feature refers to the distribution of brightness differences between adjacent keys across the entire keyboard area. This feature describes whether the brightness differences are uniformly distributed across the entire keyboard area or concentrated in localized areas, thus providing a basis for uniformity evaluation.

[0114] In this embodiment, the light leakage feature set is a feature set formed based on the edge light leakage detection results. In specific implementation, the brightness anomalies appearing in the edge region are sorted out and feature information is formed to characterize the distribution of light leakage location, the distribution of light leakage range, and the degree of light leakage concentration.

[0115] In this embodiment, the edge region brightness anomaly distribution characteristic refers to the spatial distribution of edge light leakage in the keyboard edge region. When the brightness anomaly is concentrated in a small number of areas, it indicates that there may be a local assembly defect; when the brightness anomaly covers a large area, it indicates that there may be an overall light guide defect. Therefore, this characteristic can reflect the severity of the light leakage defect.

[0116] In this embodiment, the cycle time feature set is a set of features formed based on the detection time. Specifically, the time consumption from the moment the backlight assembly is powered on to the completion of quality detection is compiled and used to form feature information characterizing the detection efficiency. The cycle time feature set reflects the degree of influence of the current detection scheme on the detection cycle time.

[0117] In this embodiment, the time consumption feature represents the time occupancy at each stage of the detection process. Specifically, the time consumption for the backlight waiting stage, image acquisition stage, and quality analysis stage is recorded to form feature information that reflects the resource occupancy of the detection process.

[0118] In this embodiment, the detection process feature vector is a data representation result formed by combining the brightness convergence feature group, stability feature group, uniformity feature group, light leakage feature group, and beat feature group in a uniform arrangement order. Each detection process feature vector corresponds to a second backlight monitoring control pair, used to fully characterize the detection process features generated under that control pair.

[0119] In this embodiment, the reinforcement learning environment state is an environmental description information composed of feature vectors from the detection process. This environment state reflects the performance of the detection process under the current detection strategy, enabling the reinforcement learning model to learn corresponding optimization rules based on different detection process features.

[0120] In this embodiment, the action input is the corresponding second backlight monitoring and control pair. In specific implementation, each action includes two control contents: the second backlight illumination waiting time and the number of second brightness acquisitions. The reinforcement learning model learns the influence relationship between the two control contents on the detection result and the detection cycle by continuously trying different actions.

[0121] In this embodiment, the feedback signal is the feedback information formed by the backlight component quality detection result and the detection time. In specific implementation, when the detection quality is high and the detection time is short, the corresponding feedback signal is superior; when the detection quality decreases or the detection time increases, the corresponding feedback signal is weaker, thereby guiding the reinforcement learning model to find a balance between quality and timing.

[0122] In this embodiment, the backlight detection reinforcement learning training environment is a training scenario jointly constructed by the environmental state, action input, and reward signal. The reinforcement learning model continuously tries different second backlight monitoring control pairs in this training environment and adjusts its strategy based on the reward signal, thereby gradually learning the optimal control law.

[0123] In this embodiment, multi-round policy iteration training is a training process in which the reinforcement learning model continuously performs policy selection, result feedback, and policy updates. Specifically, in each round of training, the model selects the corresponding second backlight monitoring control pair for learning and adjusts the subsequent control policy according to the feedback signal, so that the control policy is gradually optimized towards high-quality detection and low detection time.

[0124] In this embodiment, the beat accuracy reward generation strategy is a reward generation rule formed after multiple rounds of policy iteration training of the reinforcement learning model. This strategy can determine the comprehensive performance of the current detection scheme in terms of detection accuracy and detection beat based on the feature vector of the detection process.

[0125] In this embodiment, the clock speed and accuracy reward are a comprehensive evaluation result output by the clock speed and accuracy reward generation strategy. The clock speed and accuracy reward score is used to characterize the comprehensive balance between detection quality and detection efficiency of the corresponding second backlight monitoring control pair. The higher the reward score, the better the corresponding control pair can balance detection accuracy and detection clock speed.

[0126] In this embodiment, the reward distribution sequence is a reward set formed by the corresponding beat and accuracy reward scores of all second backlight monitoring control pairs according to the control pair index order. The reward distribution sequence is used to reflect the overall evaluation situation corresponding to different combinations of waiting time and number of acquisitions, and provides basic data for subsequent selection of the optimal control pair.

[0127] It should be noted that the brightness convergence feature group, stability feature group, uniformity feature group, and light leakage feature group correspond to the dynamic convergence capability, stable light emission capability, balanced light emission capability, and structural integrity capability in the quality inspection of backlight components, respectively, while the beat feature group corresponds to the detection efficiency capability. Therefore, the feature vector of the detection process can simultaneously cover information in both the quality dimension and the efficiency dimension.

[0128] It should be noted that reinforcement learning models do not directly evaluate the waiting time or the number of collections themselves, but rather learn the patterns of change in detection results caused by changes in waiting time and the number of collections through the feature vectors of the detection process. Therefore, they can identify the real impact of different control strategies on the quality detection process.

[0129] It should be noted that this embodiment constructs a backlight detection reinforcement learning training environment, enabling the second backlight monitoring and control pair to establish a feedback closed-loop relationship with the actual detection results. It also utilizes the clock speed and accuracy reward scores to uniformly characterize the comprehensive performance between detection quality and detection efficiency, thereby providing a reliable evaluation basis for the subsequent selection of the optimal backlight monitoring and control pair.

[0130] S6 uses the backlight monitoring and control pair as the surface coordinates and the beat and accuracy reward score as the surface height to map and fit the construction of the beat accuracy response surface, and selects the best second backlight monitoring and control pair and applies it to the backlight component quality detection.

[0131] In this embodiment, the step of using the backlight monitoring and control pair as the surface coordinates and the beat and accuracy reward score as the surface height to map and fit a beat accuracy response surface, and then selecting the best second backlight monitoring and control pair and applying it to the backlight component quality detection, specifically involves: Read the waiting time for the second backlight to turn on, the number of times the second brightness is collected, and the clock speed and accuracy bonus points for each second backlight monitoring and control pair. Using the second backlight illumination waiting time as the first coordinate value, the second brightness sampling number as the second coordinate value, and the corresponding beat and accuracy bonus score as the height value, a three-dimensional coordinate mapping relationship between the second backlight monitoring and control pair and the beat and accuracy bonus score is established. A discrete reward point set is generated based on the three-dimensional coordinate mapping relationship, and the discrete reward point set is arranged according to the numerical order of the second backlight illumination waiting time and the second brightness sampling number. The discrete reward point set is subjected to surface fitting processing to establish a continuous response relationship between the second backlight lighting waiting time, the number of second brightness acquisitions, the beat and the accuracy reward score, and generate a beat accuracy response surface. The fitting points in the beat accuracy response surface are traversed and read, and the second backlight illumination waiting time, the number of second brightness acquisitions, and the beat and accuracy bonus points corresponding to each fitting point are obtained. The fitted points are sorted according to the beat and accuracy bonus scores to generate a bonus ranking sequence; Extract the corresponding second backlight monitoring control pair sequentially along the reward sorting sequence, and establish a candidate control pair set; Read the waiting time for the second backlight to turn on and the number of times the second brightness is collected for each second backlight monitoring control pair in the candidate control pair set; The second backlight monitoring and control pair whose second backlight illumination waiting time exceeds the preset detection cycle range is screened out; The second backlight monitoring and control pair whose second brightness sampling number exceeds the preset sampling load range is filtered out; Among the remaining second backlight monitoring and control pairs, the second backlight monitoring and control pair with the highest beat and accuracy bonus score is selected as the best second backlight monitoring and control pair; The optimal second backlight monitoring control pair is converted into the corresponding camera trigger time sequence and image acquisition execution sequence, and applied to the backlight component quality inspection process.

[0132] In this embodiment, the second backlight illumination waiting time is the waiting time between the backlight component being powered on and the start of acquiring the first brightness image. This time is used to control the specific position where the detection process intervenes in the brightness establishment process. When the waiting time is short, the acquisition process can cover more brightness establishment stage information; when the waiting time is long, the acquisition process covers more brightness stabilization stage information. Therefore, this parameter is used to characterize the detection start timing.

[0133] In this embodiment, the second brightness acquisition count is the total number of brightness image acquisitions performed during a single backlight component detection process. This count determines the density of brightness change observations. Fewer acquisition counts shorten the detection time; more acquisition counts yield more complete brightness change information, thus this parameter characterizes the detection coverage.

[0134] In this embodiment, the beat and accuracy reward are a comprehensive evaluation result generated after reinforcement learning training in step S5. This reward score reflects the comprehensive performance of the corresponding second backlight monitoring control pair in terms of both detection quality and detection efficiency. When the detection result is complete and the detection time is low, the corresponding reward score is high; when the detection quality decreases or the detection time increases, the corresponding reward score decreases.

[0135] In this embodiment, the three-dimensional coordinate mapping relationship is the correspondence established between the second backlight illumination waiting time, the number of second brightness acquisitions, and the cycle time and accuracy bonus score. The second backlight illumination waiting time corresponds to the first coordinate axis, the number of second brightness acquisitions corresponds to the second coordinate axis, and the cycle time and accuracy bonus score correspond to the height axis, thus ensuring that each second backlight monitoring and control pair corresponds to a unique spatial position.

[0136] In this embodiment, the discrete reward point set is a data set formed by the corresponding three-dimensional coordinate points of all second backlight monitoring and control pairs. Each discrete reward point represents a beat and accuracy reward score corresponding to a combination of waiting time and number of acquisitions. Multiple discrete reward points together constitute the basis for subsequent response surface generation.

[0137] In this embodiment, the discrete reward point set is arranged as a point set structure after organizing the discrete reward points according to the second backlight illumination waiting time and the second brightness sampling number. In specific implementation, the points are first arranged in order of waiting time, and then in order of sampling number, so that adjacent points have a clear parameter adjacency relationship.

[0138] In this embodiment, the surface fitting process is a process of constructing a continuous reward change trend based on a set of discrete reward points. In specific implementation, the change trend between adjacent discrete reward points is used as the connection basis to continuously fill in the positions between discrete points that have not been actually tested, thereby forming a complete beat accuracy response surface.

[0139] In this embodiment, the continuous response relationship is the continuous change pattern of the beat and accuracy bonus score when the second backlight illumination waiting time and the number of brightness sampling changes. This continuous response relationship not only retains the actual test results, but also reflects the reward change trend in areas not directly tested.

[0140] In this embodiment, the beat accuracy response surface is a three-dimensional response surface generated based on continuous response relationships. This response surface is used to demonstrate the overall impact of changes in waiting time and number of acquisitions on beat rate and accuracy bonus score, making the superiority or inferiority relationship between different control strategies readily apparent.

[0141] In this embodiment, each fitting point is a sampling position in the beat accuracy response surface. Each fitting point corresponds to a waiting time, a number of samplings, and a beat and accuracy bonus score. Therefore, the fitting point can represent the local control scheme in the response surface.

[0142] In this embodiment, the reward ranking sequence is a ranking result based on the order of reward scores for beat and accuracy. In specific implementation, the reward scores corresponding to all fitted points are arranged in descending order to form the ranking basis for optimizing the control scheme.

[0143] In this embodiment, the candidate control pair set is a set of control schemes extracted based on the reward ranking sequence. The second backlight monitoring control pairs in this set all have high reward performance and are used for subsequent constraint screening and optimal scheme determination.

[0144] In this embodiment, the preset inspection cycle range is the allowable inspection time range for the current backlight component quality inspection station. This range is determined based on the production line cycle time requirements. For example, if the inspection station stipulates that the inspection time for a single piece must not exceed the corresponding production cycle time, then the allowable waiting time range is determined by this cycle time requirement, thereby avoiding the inspection process from affecting overall production efficiency.

[0145] In this embodiment, the preset acquisition load range is the range of image acquisitions allowed by the current visual inspection device. This range is determined based on the camera's buffering capacity, image processing capacity, and the inspection server's processing capacity. When the number of acquisitions exceeds the device's capacity, it may lead to image backlog or processing delays; therefore, this range is used to limit abnormal acquisition schemes.

[0146] In this embodiment, the optimal second backlight monitoring control pair is the one that achieves the highest cycle time and accuracy bonus score under the constraints of detection cycle and acquisition load. This control pair simultaneously considers both detection quality and detection efficiency requirements, and is therefore chosen as the final execution scheme.

[0147] In this embodiment, the camera trigger timing sequence and image acquisition execution sequence are the actual execution action sequences corresponding to the optimal second backlight monitoring control pair. By converting the optimal control pair into specific execution actions, it can be directly applied to the actual backlight component detection process.

[0148] It should be noted that surface fitting is not used to generate new control parameters, but rather to establish the overall trend of change between waiting time, number of samplings and reward score, thereby avoiding judgment based solely on discrete test results and improving the stability of the optimal control scheme selection.

[0149] It should be noted that during the candidate control pair screening process, both detection cycle constraints and acquisition load constraints are introduced to ensure that the screening results not only meet the reward score requirements but also the actual production line operation requirements, thereby guaranteeing that the final control scheme has the ability to be implemented in practice.

[0150] It should be noted that this embodiment constructs a beat accuracy response surface, which maps the backlight lighting waiting time, brightness acquisition times, beat and accuracy bonus scores to the same analysis space. Combined with detection cycle constraints and acquisition load constraints, the optimal control scheme is selected, so that the best second backlight monitoring control pair can reduce detection time consumption while ensuring the backlight defect detection capability, and achieve synergistic optimization of detection accuracy and detection beat.

[0151] Example 2: This invention also includes an adaptive control system for the quality inspection cycle and accuracy of notebook computer components, comprising a convergence modeling module, a time-domain perturbation module, an observation orchestration module, a brightness inspection execution module, a reward and evaluation training module, and a superior control screening module. The convergence modeling module is used to construct a brightness convergence correlation graph based on the continuous brightness change sequence of the historical backlight component lighting process, and analyze and determine the initial backlight lighting waiting time and the initial brightness acquisition count to construct a first backlight monitoring and control pair. The time-domain perturbation module is used to perform displacement and scaling perturbations on the time axis of the first backlight monitoring and control pair to generate several second backlight monitoring and control pairs. The observation orchestration module is used to construct the brightness observation trajectory corresponding to the backlight monitoring and control pair. The system transforms the data into a camera trigger time sequence and an image acquisition execution sequence. The brightness detection execution module converts each second backlight monitoring control pair and applies it to the brightness image acquisition process after the backlight component is powered on, and completes the backlight component quality detection. The reward and training module extracts features from the quality detection process of each second backlight monitoring control pair, uses the quality detection results and detection time as rewards, and obtains the corresponding beat and accuracy reward scores through strategy iterative training. The optimal control selection module maps and fits the beat and accuracy response surface using the backlight monitoring control pair as the surface coordinates and the beat and accuracy reward scores as the surface height, and selects the best second backlight monitoring control pair and applies it to the backlight component quality detection.

[0152] The present invention also includes an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform an adaptive control method for quality inspection cycle and accuracy of notebook computer components.

[0153] The present invention also includes a computer-readable storage medium storing a computer program that, when executed by a processor, implements an adaptive control method for the quality inspection cycle and accuracy of notebook computer components.

[0154] In the embodiments provided by this invention, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0155] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0156] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0157] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0158] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0159] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0160] In the embodiments provided in this disclosure, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0161] It should be noted that, in this disclosure, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element limited by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0162] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. An adaptive control method for the cycle time and accuracy of quality inspection of notebook computer components, characterized in that, Includes the following steps: A brightness convergence correlation diagram is constructed based on the continuous brightness change sequence of the historical backlight component lighting process, and the initial backlight lighting waiting time and the number of initial brightness acquisitions are analyzed and determined to construct the first backlight monitoring and control pair. The first backlight monitoring and control pair is subjected to displacement and scaling disturbances on the time axis to generate several second backlight monitoring and control pairs; Construct the corresponding brightness observation trajectory for the backlight monitoring and control pair, and convert it into a camera trigger time sequence and an image acquisition execution sequence; Each second backlight monitoring and control pair is converted and applied to the brightness image acquisition process after the backlight component is powered on and lit, and the backlight component quality inspection is completed. Feature extraction is performed on the quality detection process of each second backlight monitoring and control pair. The quality detection results and detection time are used as rewards. The corresponding beat and accuracy reward scores are obtained through strategy iterative training. Using the backlight monitoring and control pair as the surface coordinates and the beat and accuracy reward score as the surface height, a beat accuracy response surface is mapped and fitted, and the best second backlight monitoring and control pair is selected and applied to the quality inspection of the backlight assembly.

2. The adaptive control method for quality inspection cycle time and accuracy of notebook computer components according to claim 1, characterized in that, The process involves constructing a brightness convergence correlation graph based on the continuous brightness change sequence of the historical backlight component lighting process, analyzing and determining the initial backlight lighting waiting time and the number of initial brightness acquisitions, and constructing the first backlight monitoring and control pair, specifically as follows: Collect the continuous brightness change sequence of the historical backlight component from the start of power-on to the end of brightness stabilization, and divide it into several continuous time windows according to a preset time interval to obtain the brightness distribution state corresponding to each time window. The brightness of the keys in each brightness distribution state is classified into levels, and the corresponding key brightness level sequence is generated. Count the number of keys that maintain the same brightness level in the key brightness level sequence corresponding to adjacent time windows, and form a level maintenance sequence according to time order; Based on the grade preservation sequence, the correlation between time windows is constructed, adjacent time windows are used as correlation nodes, and the corresponding grade preservation number is used as the correlation strength between nodes to generate a brightness convergence correlation graph. Extract the starting time position corresponding to the target brightness convergence segment, and determine the starting time position corresponding to the target brightness convergence segment as the initial backlight illumination waiting time; The number of time windows covered by the target brightness convergence segment is counted, and the number of time windows covered by the target brightness convergence segment is determined as the initial number of brightness acquisitions; The initial backlight illumination waiting time and the number of initial brightness sampling times are combined to form the first backlight monitoring and control pair.

3. The adaptive control method for quality inspection cycle time and accuracy of notebook computer components according to claim 2, characterized in that, The first backlight monitoring and control pair is subjected to time axis displacement and scaling disturbances to generate several second backlight monitoring and control pairs, specifically as follows: Extract the target brightness convergence segment corresponding to the first backlight monitoring and control pair, and obtain the start time position and end time position corresponding to the target brightness convergence segment; Construct a convergence segment time axis based on the start and end time positions of the target brightness convergence segment, and generate several time offsets according to a preset time step. Each time offset is superimposed onto the start and end time positions of the target brightness convergence segment, and a time axis displacement perturbation is performed on the target brightness convergence segment to generate several displacement convergence segments. The center position of each displacement convergence segment is obtained respectively, and the center position of each segment is used to replace the original center position of the target brightness convergence segment to generate several candidate observation center positions. Based on the center position of each displacement convergence segment, the time span of the displacement convergence segment is expanded and compressed according to a preset scaling ratio to generate several scaling convergence segments. The number of time windows covered by each expansion and contraction segment is counted, and the corresponding candidate observation coverage length is formed; Convert the time position corresponding to each candidate observation center location into the second backlight illumination waiting time; Convert the number of time windows corresponding to the coverage length of each candidate observation into the number of second brightness acquisitions; By combining the second backlight illumination waiting time and the second brightness acquisition number according to the one-to-one correspondence, several second backlight monitoring and control pairs are generated.

4. The adaptive control method for quality inspection cycle time and accuracy of notebook computer components according to claim 3, characterized in that, The construction of the backlight monitoring and control system corresponds to the brightness observation trajectory, which is then converted into a camera trigger time sequence and an image acquisition execution sequence, specifically as follows: Extract the waiting time for the second backlight to turn on and the number of times the second brightness is collected for each second backlight monitoring and control pair. The initial brightness observation trajectory is constructed by taking the time position corresponding to the second backlight illumination waiting time as the observation start node and the number of time windows corresponding to the second brightness acquisition times as the number of observation coverage nodes. Based on the time window covered by the initial brightness observation trajectory, the associated nodes of the corresponding time window in the brightness convergence correlation graph are extracted, and the observation node chain is formed in chronological order. The number of levels maintained by each associated node in the observation node chain is counted, and the proportion of the number of levels maintained by each associated node to the total number of levels maintained by the observation node chain is calculated to form a node maintenance proportion sequence. The time interval between adjacent observation nodes in the observation node chain is redistributed based on the node retention ratio sequence. The time interval between adjacent observation nodes is shortened for associated nodes with a higher retention ratio, and the time interval between adjacent observation nodes is extended for associated nodes with a lower retention ratio. The time position of each observation node is updated according to the reallocated time interval to generate an optimized brightness observation trajectory. Extract the time position corresponding to each observation node along the optimized brightness observation trajectory, and form a set of observation times in chronological order; Convert each time location in the observation time set into a camera trigger time to generate a camera trigger time sequence; The image acquisition execution sequence is generated according to the sequential order of the camera trigger time sequence, thereby realizing the conversion of the second backlight monitoring and control towards the actual detection execution action.

5. The adaptive control method for quality inspection cycle time and accuracy of notebook computer components according to claim 4, characterized in that, The process of converting and applying the second backlight monitoring and control pairs to the brightness image acquisition process after the backlight assembly is powered on, and completing the backlight assembly quality inspection, specifically involves: Obtain the camera trigger time sequence and image acquisition execution sequence corresponding to each second backlight monitoring and control pair; Control the backlight assembly to connect to the detection power supply and execute the power-on lighting, establishing the corresponding brightness change process; The camera is controlled to perform image acquisition at the corresponding trigger time according to the camera trigger time sequence, and the corresponding brightness image set is obtained according to the image acquisition execution sequence. The brightness image set is sorted chronologically according to the order of image acquisition execution sequence to generate a continuous brightness image sequence. Brightness extraction is performed on each key region in a continuous brightness image sequence to generate a key brightness variation sequence; Based on the key brightness change sequence, the brightness change amplitude of each key during continuous acquisition is statistically analyzed to generate key brightness stability results; Based on the continuous brightness image sequence, the brightness difference between adjacent keys is statistically analyzed to generate the brightness uniformity result of the keys; The brightness distribution of the key edge region is extracted from the continuous brightness image sequence to generate edge light leakage detection results; By integrating the key brightness stability results, key brightness uniformity results, and edge light leakage detection results, the backlight component quality detection results corresponding to the second backlight monitoring and control pair are generated. Record the detection time of the corresponding second backlight monitoring and control from the power-on of the backlight component to the acquisition of the backlight component quality detection result.

6. The adaptive control method for quality inspection cycle time and accuracy of notebook computer components according to claim 5, characterized in that, The process of quality detection for each second backlight monitoring and control pair involves feature extraction, using the quality detection results and detection time as rewards, and obtaining corresponding beat and accuracy reward scores through strategy iterative training. Specifically: Read the continuous brightness image sequence, key brightness stability result, key brightness uniformity result, edge light leakage detection result, backlight component quality detection result, and detection time corresponding to each second backlight monitoring and control pair respectively; Based on the continuous brightness image sequence, the brightness change features of the key positions at each acquisition time are extracted, and a brightness convergence feature group is constructed according to the brightness difference between adjacent acquisition times. Based on the key brightness stability results, the distribution characteristics of brightness variation amplitude of each key are extracted to form a stability feature group; Based on the results of key brightness uniformity, the distribution features of brightness differences between keys are extracted to form a uniformity feature group; Based on the edge light leakage detection results, the abnormal brightness distribution characteristics of the edge region are extracted to form a light leakage feature group; Based on the detection time consumption, the time consumption features from power-on lighting to the completion of quality detection are extracted to form a cycle time feature group; The brightness convergence feature group, stability feature group, uniformity feature group, light leakage feature group and rhythm feature group are combined in a fixed order to form a detection process feature vector that corresponds one-to-one with each second backlight monitoring and control pair. The detection process feature vector is used as the state of the reinforcement learning environment, the corresponding second backlight monitoring and control pair is used as the action input, and the backlight component quality detection result and detection time are used to jointly construct the reward signal to generate the backlight detection reinforcement learning training environment. Based on each second backlight monitoring and control, multiple rounds of strategy iteration training are performed on the corresponding backlight detection process, so that the reinforcement learning model learns the influence relationship between the backlight lighting waiting time and the number of brightness acquisitions on the detection quality and detection cycle, and obtains the cycle accuracy reward generation strategy. Input the corresponding detection process feature vector of each second backlight monitoring and control pair into the beat accuracy reward generation strategy, and output the corresponding beat and accuracy reward score; The corresponding beat and accuracy bonus points for each second backlight monitoring and control pair are summarized to form a bonus distribution sequence.

7. The adaptive control method for quality inspection cycle time and accuracy of notebook computer components according to claim 6, characterized in that, The process involves using the backlight monitoring and control pair as the surface coordinates, and the beat and accuracy reward scores as the surface height. A beat-accuracy response surface is mapped and fitted, and the optimal second backlight monitoring and control pair is selected and applied to backlight component quality inspection. Specifically: Read the waiting time for the second backlight to turn on, the number of times the second brightness is collected, and the clock speed and accuracy bonus points for each second backlight monitoring and control pair. Using the second backlight illumination waiting time as the first coordinate value, the second brightness sampling number as the second coordinate value, and the corresponding beat and accuracy bonus score as the height value, a three-dimensional coordinate mapping relationship between the second backlight monitoring and control pair and the beat and accuracy bonus score is established. A discrete reward point set is generated based on the three-dimensional coordinate mapping relationship, and the discrete reward point set is arranged according to the numerical order of the second backlight illumination waiting time and the second brightness sampling number. The discrete reward point set is subjected to surface fitting processing to establish a continuous response relationship between the second backlight lighting waiting time, the number of second brightness acquisitions, the beat and the accuracy reward score, and generate a beat accuracy response surface. The fitting points in the beat accuracy response surface are traversed and read, and the second backlight illumination waiting time, the number of second brightness acquisitions, and the beat and accuracy bonus points corresponding to each fitting point are obtained. The fitted points are sorted according to the beat and accuracy bonus scores to generate a bonus ranking sequence; Extract the corresponding second backlight monitoring control pair sequentially along the reward sorting sequence, and establish a candidate control pair set; Read the waiting time for the second backlight to turn on and the number of times the second brightness is collected for each second backlight monitoring control pair in the candidate control pair set; The second backlight monitoring and control pair whose second backlight illumination waiting time exceeds the preset detection cycle range is screened out; The second backlight monitoring and control pair whose second brightness sampling number exceeds the preset sampling load range is filtered out; Among the remaining second backlight monitoring and control pairs, the second backlight monitoring and control pair with the highest beat and accuracy bonus score is selected as the best second backlight monitoring and control pair; The optimal second backlight monitoring control pair is converted into the corresponding camera trigger time sequence and image acquisition execution sequence, and applied to the backlight component quality inspection process.

8. A system using the adaptive control method for quality inspection cycle time and accuracy of notebook computer components as described in any one of claims 1-7, characterized in that, It includes a convergence modeling module, a time-domain perturbation module, an observation arrangement module, a bright detection execution module, a reward and evaluation training module, and a superior control screening module; The convergence modeling module is used to construct a brightness convergence correlation graph based on the continuous brightness change sequence of the historical backlight component lighting process, and analyze and determine the initial backlight lighting waiting time and the number of initial brightness acquisitions to construct the first backlight monitoring and control pair. The time-domain perturbation module is used to perform displacement and scaling perturbations on the time axis of the first backlight monitoring and control pair to generate several second backlight monitoring and control pairs. The observation orchestration module is used to construct the corresponding brightness observation trajectory for the backlight monitoring and control pair, and convert it into a camera trigger time sequence and an image acquisition execution sequence; The brightness inspection execution module is used to convert and apply each second backlight monitoring control pair to the brightness image acquisition process after the backlight component is powered on and lit, and to complete the quality inspection of the backlight component. The reward and evaluation training module is used to extract features from the quality detection process of each second backlight monitoring and control pair, and uses the quality detection results and detection time as rewards. Through strategy iterative training, the corresponding beat and accuracy reward points are obtained. The optimal control screening module is used to map and fit a beat accuracy response surface using the backlight monitoring control pair as the surface coordinates and the beat and accuracy reward score as the surface height. It then selects the best second backlight monitoring control pair and applies it to the quality inspection of the backlight components.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the adaptive control method for quality inspection cycle and accuracy of notebook computer components as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the adaptive control method for quality inspection cycle and accuracy of notebook computer components as described in any one of claims 1 to 7.