Embedded nanometer light source intelligent control system based on self-learning

Abstract

The invention relates to the technical field of light source intelligent control, in particular to an embedded nanometer light source intelligent control system based on self-learning. The light-emitting unit comprises a plurality of light-emitting mechanisms embedded in the base body; the instruction output unit is used for controlling each light-emitting mechanism to adjust lighting parameters according to the test instruction; the acquisition unit is used for periodically acquiring image information of the surface of the substrate; the storage unit is used for storing expected image information corresponding to each acquisition period; the verification unit is used for listing the acquired image information and the corresponding expected image information under the same time node into a single image group, and comparing illumination parameters of each feature point in the image group and a corresponding preset feature point; and the analysis unit is used for determining whether the operation of the light emitting unit is qualified or not according to the comparison result of the verification unit. According to the embedded nano light source system, intelligent self-learning control of the embedded nano light source system is realized by improving the accuracy of light source control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control of light sources, and particularly relates to an embedded nano light source intelligent control system based on self-learning. BACKGROUND

[0002] As a new type of precision lighting technology, the embedded nano light source has been widely used in precision manufacturing, biological medicine, high-end display, scientific research and detection and other fields with high requirements for light precision, stability and dynamic adjustment ability due to its small size, low energy consumption, high luminous efficiency, high spectral purity and strong controllability. However, the control system of the existing embedded nano light source still has many technical limitations in practical application, mainly in the aspects of insufficient control precision and dynamic adaptability, fuzzy fault diagnosis and single processing method, lack of self-learning and closed-loop optimization mechanism, and poor coordination and weak verification mechanism of each module.

[0003] Chinese patent application No. CN102878497A discloses a high-output nano-oxygen LED intelligent control air conditioner lamp, which is installed on a ceiling and comprises an LED light source module, a grid lamp disc, a ballast, and an anodic electrochemical aluminum reflector. The reflector is arranged on the lamp disc base, and the LED light source module is an LED nano photocatalyst. The lamp disc is U-shaped and is formed by hydraulic forming. The lamp disc is coated with nano silver and sodium ions. The lamp disc is provided with a plurality of air outlets on both sides. The application can promote the upgrading of the air conditioner lamp market, and is conducive to the realization of the national green, environmental protection, energy saving and emission reduction goals. The lamp disc is provided with an intelligent control system, so that the application can independently complete the automatic control of the control target when the environmental conditions change.

[0004] However, the prior art still has the following problems:

[0005] The control precision and dynamic adaptability of the nano light source are insufficient, the fault diagnosis is fuzzy and the processing method is single, the self-learning and closed-loop optimization mechanism is lacking, and the coordination of each module is poor and the verification mechanism is weak. SUMMARY

[0006] Therefore, the present application provides an embedded nano light source intelligent control system based on self-learning to overcome the problems in the prior art, such as insufficient control precision and dynamic adaptability of the nano light source, fuzzy fault diagnosis and single processing method, lack of self-learning and closed-loop optimization mechanism, and poor coordination and weak verification mechanism of each module.

[0007] To achieve the above purpose, the present application provides an embedded nano light source intelligent control system based on self-learning. It comprises:

[0008] a base body;

[0009] A light-emitting unit is connected with the base body and comprises a plurality of light-emitting mechanisms embedded in the base body;

[0010] An instruction output unit is connected with each of the light-emitting mechanisms, configured to receive a test instruction and control each light-emitting mechanism to adjust the lighting parameter according to the test instruction;

[0011] A collection unit is configured to periodically collect image information of the base body surface, extract a plurality of feature points corresponding to the light-emitting mechanisms from each of the image information, and obtain actual lighting parameters of each feature point;

[0012] A storage unit is connected with the collection unit and configured to store expected image information corresponding to each collection period, each of the expected image information comprising a plurality of preset feature points and corresponding preset lighting parameters;

[0013] A verification unit is connected with the storage unit and the collection unit respectively, configured to list the collected image information and the corresponding expected image information at the same time node as a single image group, compare the lighting parameters of each feature point in the image group with the corresponding preset feature points, and screen each image information according to the comparison result;

[0014] An analysis unit is connected with the verification unit, configured to determine whether the operation of the light-emitting unit is qualified according to the comparison result of the verification unit, and determine the fault cause based on the average abnormality proportion when it is unqualified, to generate a corresponding processing mode;

[0015] A control unit is connected with the analysis unit and the instruction output unit respectively, configured to control the instruction output unit to output relevant instructions in the subsequent actual use process according to the processing mode generated by the analysis unit, so as to adjust the operation parameter of the light-emitting unit to a target value.

[0016] Further, the process of determining whether the operation of the light-emitting unit is qualified by the analysis unit comprises:

[0017] Calculating a matching rate, the matching rate being a ratio of the number of image information matched with the corresponding expected image information to the total number of image information;

[0018] For a single image information, if the ratio of the number of matched feature points to the total number of feature points reaches a preset value, it is determined that the image information is matched with the corresponding expected image information;

[0019] For a single feature point, if the similarity between its actual lighting parameter and the preset lighting parameter of the corresponding preset feature point is higher than a preset similarity, it is determined that the feature point is a matched feature point;

[0020] If the matching rate is higher than a preset matching rate threshold, it is determined that the operation of the light-emitting unit is qualified.

[0021] If the matching rate is lower than or equal to the preset matching rate threshold, it is determined that the light emitting unit is unqualified, and the average number of abnormal feature points in each image information is counted, and the unqualified reason is determined based on the average abnormal proportion.

[0022] Further, the average abnormal proportion is the average value of the abnormal proportions in the image information; wherein the abnormal proportion of a single image information is the ratio of the number of unmatched feature points to the total number of feature points in the image information.

[0023] Further, the process of determining the unqualified reason based on the average abnormal proportion includes:

[0024] If the average abnormal proportion is greater than the preset abnormal proportion threshold, it is determined that the unqualified reason is in the instruction output unit, and the lighting color of the light emitting mechanism is adjusted preferentially, and if the unqualified reason is still not qualified after adjusting the color, the lighting brightness is further adjusted;

[0025] If the average abnormal proportion is less than or equal to the preset abnormal proportion threshold, it is determined that the unqualified reason is in a single light emitting mechanism, and the line connected to the light emitting mechanism is repaired.

[0026] Further, the analysis unit is further used for:

[0027] Calculate the average color difference of each image information, and the color difference is the difference between the actual color value of the feature point and the preset color value of the corresponding preset feature point;

[0028] Calculate the variance of each average color difference, and determine whether the color deviation is caused by instruction input delay according to the variance:

[0029] If the variance is less than the preset variance threshold, it is determined that the color deviation is caused by instruction delay, and a processing mode for correcting the instruction output time length of the instruction output unit is generated;

[0030] If the variance is greater than or equal to the preset variance threshold, it is determined that the color deviation is caused by the light emitting mechanism, and a processing mode for repairing the light emitting mechanism is generated.

[0031] Further, when the analysis unit generates the processing mode for correcting the instruction output time length, the instruction output delay is corrected according to the difference between the preset matching rate and the actual matching rate, and the correction amount of the instruction output delay is positively correlated with the difference.

[0032] Further, when adjusting the lighting brightness of the light emitting mechanism, the analysis unit adjusts the operating power of the light emitting mechanism according to the average color difference of each image information, and the adjustment amount of the operating power of the light emitting mechanism is positively correlated with the average color difference.

[0033] Further, in the process of adjusting the illumination brightness, the analysis unit is further configured to correct the brightness according to the number of feature points, and the correction amount of the brightness is positively correlated with the number of feature points.

[0034] Further, after completing the brightness adjustment, if the light-emitting unit is still unqualified, the analysis unit controls to repeat the adjustment of the brightness until a critical condition is reached, and the critical condition includes a maximum cumulative adjustment number or a critical brightness; when the critical condition is reached, a processing mode of shortening the distance between the acquisition unit and the light-emitting unit is generated.

[0035] Further, when the analysis unit generates the processing mode of shortening the distance, the shortening distance is determined according to the difference between the target brightness value and the critical brightness value, and the shortening amount of the distance is positively correlated with the difference; after the distance is shortened, the analysis unit re-detects, if the light-emitting unit is still unqualified, determines that the system is faulty, and generates a processing mode of issuing a maintenance notice.

[0036] Compared with the prior art, the beneficial effects of the present application are that the present application verifies from the most basic feature point parameter, gradually aggregates to the overall image, and then to the final matching rate statistics, and each level of determination has a clear quantitative standard, avoiding the roughness of directly determining the overall result, reducing the probability of misjudgment and omission, and when the scheme determines that it is unqualified, further statistics the average number of abnormal feature points in each image information, and analyzes the unqualified reason based on the average abnormal proportion, providing a basis for subsequent maintenance and debugging. From the similarity calculation of the feature point parameter, to the matching determination of the single image, to the statistics of the overall matching rate, all are completed based on clear numerical logic without artificial subjective judgment. This feature makes it easy to integrate into an automatic detection system to realize batch and rapid detection of the running state of the light-emitting unit, not only reduces the labor cost, but also avoids the inconsistency of the results caused by individual differences in human observation, ensuring the objectivity and stability of the determination.

[0037] Further, the present application directly locks the fault in the two core directions of systematic instruction output or individual light-emitting mechanism circuit by the average abnormal proportion, avoiding the inefficient mode of checking all components one by one in the traditional detection, greatly shortening the fault positioning time, and the order of preferentially adjusting the illumination color and then adjusting the brightness conforms to the influence law of the illumination parameter on the image feature point recognition, the color deviation may directly cause the distortion of the feature point color information (such as the red feature point is difficult to identify under the greenish illumination), and the brightness deviation more affects the definition of the feature point (overexposure or overdarkness), reducing the invalid adjustment steps and improving the one-time repair success rate.

[0038] Further, the present application judges the color deviation reason through the variance of the average color difference, and the color deviation caused by instruction input delay has systematicity (all light emitting mechanisms are affected by the same delay factor, the deviation is consistent, and the variance is small); and the deviation caused by light emitting mechanism failure has randomness (different mechanisms have different failure degrees and types, the deviation has no uniform rule, and the variance is large); the distinguishing method based on data fluctuation makes the determination result more accurate, avoids confusion between systematic problems and individual hardware problems, and directly points to the processing direction of correcting the instruction output time length (instruction delay) or repairing the light emitting mechanism (mechanism failure) according to the comparison between the variance and the preset threshold, so that subsequent operations do not deviate from the core problem and invalid processing is reduced; the variance analysis can directly distinguish whether the failure is caused by the instruction system or the hardware mechanism, avoids the tedious process of checking all possible links one by one, and greatly shortens the fault positioning time.

[0039] Further, the instruction output delay correction, light emitting power adjustment and brightness correction in the present application are quantitatively adjusted through the linear relationship of the multiplication of the proportion coefficient and the variable, which directly links the correction amount and the deviation degree, the greater the deviation, the stronger the adjustment, and the system parameters can be quickly adjusted to a reasonable range. In the brightness adjustment, not only the average color difference is considered, but also the number of feature points is introduced for secondary correction, so that the adjustment is more in line with the actual scene, and the overallness and accuracy of the correction are improved.

[0040] Further, the present application first adjusts the brightness repeatedly, then takes physical adjustment, and finally intervenes manually, which reduces unnecessary operation cost and equipment loss, provides clear boundaries for critical conditions, sets the maximum cumulative adjustment times and the critical brightness, and defines a reasonable boundary for brightness adjustment to prevent the system from falling into an infinite cycle of invalid adjustment, thereby protecting the light emitting unit and saving system resources; the distance shortening amount is positively correlated with the difference between the target brightness value and the critical brightness value, the greater the brightness gap, the greater the distance adjustment range, and the brightness deficiency can be quickly made up; after shortening the distance, the light emitting unit is re-detected for eligibility, and if it is still unqualified, it is determined that the system is faulty, thereby avoiding the situation that the problem is not solved after distance adjustment, and when all automatic adjustment means fail, a maintenance notice is sent in time to transfer the problem to manual processing, thereby ensuring the integrity of fault handling. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 The structure block diagram of the embedded nano light source intelligent control system based on self-learning of the present application;

[0042] Figure 2 The determination flowchart for determining whether the light emitting unit is qualified or not;

[0043] Figure 3 The determination flowchart for determining the reason why the light emitting unit is unqualified. DETAILED DESCRIPTION

[0044] In order to make the objects and advantages of the present application clearer, the following further describes the present application with reference to examples; it should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0045] It should be noted that the data in the present examples are obtained by comprehensive analysis and evaluation of historical data and corresponding historical determination results in the past 6 months before the present determination according to the present application. Those skilled in the art can understand that the determination method of the present application for a single parameter can be to select the value with the highest proportion as the preset standard parameter according to the data distribution, to use weighted summation to obtain the value as the preset standard parameter, to substitute each historical data into a specific formula and to obtain the value by using the formula as the preset standard parameter, or other selection methods, as long as the present application can clearly define different specific situations in the single determination process by the obtained value.

[0046] The preferred embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application and are not intended to limit the protection scope of the present application.

[0047] It should be noted that in the description of the present application, the terms "upper", "lower", "left", "right", "inner", "outer" and the like indicate the direction or positional relationship terms based on the direction or positional relationship shown in the drawings, which are only for the convenience of description and do not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present application.

[0048] In addition, it should also be noted that in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be direct connection, or indirect connection through an intermediate medium, or internal communication of two elements. Those skilled in the art can understand the specific meaning of the above terms in the present application according to the specific circumstances.

[0049] Please refer to Figure 1 As shown in the figure, it is a structure block diagram of the self-learning-based embedded nano light source intelligent control system of the present application.

[0050] The present application provides a self-learning-based embedded nano light source intelligent control system, which comprises:

[0051] a base body;

[0052] A light-emitting unit is connected with the base body and comprises a plurality of light-emitting mechanisms embedded in the base body;

[0053] An instruction output unit is connected with each of the light-emitting mechanisms, configured to receive a test instruction and control each light-emitting mechanism to adjust the lighting parameter according to the test instruction;

[0054] A collection unit is configured to periodically collect image information of the base body surface, extract a plurality of feature points corresponding to the light-emitting mechanisms from each of the image information, and obtain actual lighting parameters of each feature point;

[0055] A storage unit is connected with the collection unit and configured to store expected image information corresponding to each collection period, each of the expected image information comprising a plurality of preset feature points and corresponding preset lighting parameters;

[0056] A verification unit is connected with the storage unit and the collection unit respectively, configured to list the collected image information and the corresponding expected image information at the same time node as a single image group, compare the lighting parameters of each feature point and the corresponding preset feature point in the image group, and screen each image information according to the comparison result;

[0057] An analysis unit is connected with the verification unit, configured to determine whether the operation of the light-emitting unit is qualified according to the comparison result of the verification unit, and determine the fault cause based on the average abnormality proportion when it is unqualified, to generate a corresponding processing mode;

[0058] A control unit is connected with the analysis unit and the instruction output unit respectively, configured to control the instruction output unit to output relevant instructions in the subsequent actual use process according to the processing mode generated by the analysis unit, so as to adjust the operation parameter of the light-emitting unit to a target value.

[0059] The illumination parameters in the embodiments of the present application include, but are not limited to, "light intensity, color temperature, wavelength, light uniformity, light-emitting frequency, beam angle, and illumination duration". The periodic acquisition of the image information of the substrate surface includes triggering the acquisition action at a preset time interval (the period can be set based on the physical characteristics of the light-emitting mechanism (such as the startup stabilization time of a nano light source, the parameter adjustment response speed), for example, if the light-emitting mechanism needs 1 second from the adjustment instruction to the parameter stabilization, the acquisition period should be no less than the response time to avoid acquisition errors caused by unstable parameters), ensuring continuous and regular monitoring of the illumination state of the substrate surface; the acquisition range covers all illumination areas corresponding to the light-emitting mechanisms on the substrate surface, ensuring that the illumination effects of each light-emitting mechanism can be completely captured; during each acquisition process, the acquisition time stamp is recorded synchronously, so that the acquired image information and the expected image information at the corresponding time node in the storage unit form a precise matching time correlation; the image information is preprocessed (such as noise reduction and contrast enhancement) during the acquisition process.

[0060] Please refer to Figure 2 as shown in the determination flowchart for determining whether the light-emitting unit is qualified.

[0061] Specifically, the process of determining whether the light-emitting unit is qualified by the analysis unit includes:

[0062] calculating a matching rate, the matching rate being a ratio of the number of image information matched with the corresponding expected image information to the total number of image information;

[0063] For a single image information, if the ratio of the number of matched feature points to the total number of feature points reaches a preset value, it is determined that the image information is matched with the corresponding expected image information;

[0064] For a single feature point, if the similarity between its actual illumination parameter and the preset illumination parameter of the corresponding preset feature point is higher than a preset similarity, it is determined that the feature point is a matched feature point;

[0065] If the matching rate is higher than a preset matching rate threshold, it is determined that the light-emitting unit is qualified;

[0066] If the matching rate is lower than or equal to the preset matching rate threshold, it is determined that the light-emitting unit is unqualified, and the average number of abnormal feature points in each image information is counted to determine the unqualified reason based on the average abnormal proportion.

[0067] The preset value in the embodiment of the application is an average value of a ratio of a number of feature points matched with expected image information to a total number of feature points when a plurality of groups of light emitting units are in normal operation without failure, the preset similarity is an average value of a similarity between an actual lighting parameter of a feature point and a preset lighting parameter of a corresponding preset feature point when the plurality of groups of light emitting units are in normal operation without failure, the preset matching rate threshold is an average matching rate when the plurality of groups of light emitting units are in normal operation without failure, but the above values are not limited thereto, and a person skilled in the art can value according to actual conditions.

[0068] The application verifies from the most basic feature point parameter, gradually aggregates to the overall image, and then to the final matching rate statistics, and each level of judgment has a clear quantitative standard, avoiding the roughness of directly judging the overall result, reducing the probability of misjudgment and omission, and when the scheme determines that it is unqualified, further statistics of the average number of abnormal feature points in each image information are performed, and the unqualified reason is analyzed based on the average abnormal proportion, thereby providing a basis for subsequent maintenance and debugging. From the similarity calculation of the feature point parameter, to the matching judgment of the single image, to the statistics of the overall matching rate, all are completed based on clear numerical logic without artificial subjective judgment. This feature makes it easy to integrate into an automatic detection system to realize batch and rapid detection of the running state of the light emitting unit, not only reduces the labor cost, but also avoids the inconsistency of the results caused by individual differences in human observation, thereby ensuring the objectivity and stability of the judgment.

[0069] Specifically, the average abnormal proportion is an average value of abnormal proportions in each image information, and the abnormal proportion of a single image information is a ratio of a number of unmatched feature points to a total number of feature points in the image information.

[0070] Please refer to Figure 3 as shown in the drawing, which is a determination flowchart of a process of determining a running unqualified reason of a light emitting unit.

[0071] Specifically, the process of determining the unqualified reason by the analysis unit based on the average abnormal proportion includes:

[0072] If the average abnormal proportion is greater than a preset abnormal proportion threshold, it is determined that the unqualified reason is in the instruction output unit, and the lighting color of the light emitting mechanism is preferentially adjusted, and if the lighting color is still unqualified after adjustment, the lighting brightness is further adjusted;

[0073] If the average abnormal proportion is less than or equal to the preset abnormal proportion threshold, it is determined that the unqualified reason is in a single light emitting mechanism, and a line connected to the light emitting mechanism is maintained.

[0074] The preset abnormal proportion threshold in the embodiment of the application can be determined by the following method: two fault scenarios are simulated in a laboratory environment, and the average abnormal proportions of a large number of samples are recorded respectively; the average abnormal proportion data at this time is collected, denoted as set A, by simulating the fault of the instruction output unit (such as intentionally setting a wrong lighting color / luminance instruction); the average abnormal proportion data at this time is collected, denoted as set B, by simulating the fault of a single light emitting mechanism circuit (such as artificially disconnecting a certain mechanism circuit); the distribution characteristics of set A and set B are analyzed, and the critical gap between the two is found; for example, if the minimum value of set A is 30% and the maximum value of set B is 10%, the threshold can be initially set between 10%-30%, preferably 15%; but the above values are not limited thereto, and a person skilled in the art can value according to the actual situation.

[0075] In the embodiment of the application, the lighting color of the light emitting mechanism is adjusted first, and if it is still unqualified after adjusting the color, the lighting brightness is further adjusted, including presetting a plurality of standard color parameters (such as RGB values or color temperature ranges) corresponding to the optimal lighting scheme under different scenes (such as using cold white light to detect light-colored objects and using warm white light to detect dark-colored objects); when adjusting, the lighting color of the light emitting mechanism is adjusted in turn according to the preset color gradient (such as gradually approaching the standard color from the current color, or switching to the adjacent standard color gear); after each adjustment, the image information is re-collected and the average abnormal proportion is calculated, and if the proportion falls below the preset threshold (i.e. qualified), the adjustment is stopped; if the color adjustment is still unqualified, it means that the fault may be caused by the deviation of the brightness parameter (such as overexposure of the image caused by too high brightness, and blurring of the feature points caused by insufficient brightness, which will increase the number of mismatched feature points), based on the optimal color parameter that has been adjusted, the lighting intensity of the light emitting mechanism is adjusted according to the preset brightness gradient (starting from the current brightness value, gradually increasing or decreasing by 5% / 10%); after each brightness adjustment, the image is re-collected and the average abnormal proportion is calculated, and if the proportion falls below the threshold, the adjustment is stopped; if it is still unqualified after traversing all the preset brightness gears, other maintenance methods are combined;

[0076] The preset standard color parameters can be determined based on the inherent characteristics of the detected object. If the detected object is a color feature point, the standard parameters need to be set according to the inherent color of the feature point. For example, when detecting a red mark, the standard color needs to avoid color deviation with red (which can be set as neutral white light, RGB value about (255, 255, 255)), to avoid the absorption or reflection of the lighting color to the feature point color leading to recognition failure; the preset color gradient can be determined based on the sensitivity of feature point matching. The influence of color change on abnormal proportion is tested by experiment. For example, when the color parameter changes X value, if the abnormal proportion decreases by 10% or more, X is the effective gradient; if the abnormal proportion does not change significantly after the change of X value, it means that the gradient is too fine, and finally the smallest gradient that can significantly reduce the abnormal proportion by single adjustment is selected. For example: it is found by test that when the RGB channel is ±8 each time, the average change of abnormal proportion is 15% (effective), and the gradient is set to 8; the determination method of the preset brightness gradient is based on the influence of brightness change on the abnormal proportion, and the determination method is the same as that of the preset color gradient determination method, but the above values are not limited thereto, and those skilled in the art can value according to the actual situation.

[0077] The present application directly locks the fault in the two core directions of systematic instruction output or individual light emitting mechanism circuit by the average abnormal proportion, avoids the inefficient mode of checking all components one by one in the traditional detection, greatly shortens the fault positioning time, and adjusts the order of lighting color first and then brightness, which conforms to the influence law of lighting parameters on image feature point recognition. Color deviation may directly lead to distortion of feature point color information (such as red feature point under greenish lighting), and brightness deviation more affects the definition of feature points (overexposure or overdarkness), reduces the invalid adjustment steps, and improves the success rate of one-time repair.

[0078] Specifically, the analysis unit is further configured to:

[0079] calculate the average color difference of each image information, the color difference being the difference between the actual color value of the feature point and the preset color value of the corresponding preset feature point;

[0080] calculate the variance of each average color difference, and determine whether the color deviation is caused by instruction input delay according to the variance:

[0081] if the variance is less than a preset variance threshold, it is determined that the color deviation is caused by instruction delay, and a processing mode of correcting the instruction output time length of the instruction output unit is generated;

[0082] if the variance is greater than or equal to the preset variance threshold, it is determined that the color deviation is caused by the light emitting mechanism, and a processing mode of repairing the light emitting mechanism is generated.

[0083] The preset variance threshold in the embodiment of the application can simulate two fault scenarios through experimental simulation, collect sample data for statistical analysis, simulate an instruction input delay scenario, trigger systematic color deviation by artificially setting different degrees of instruction delay, collect variance data of at least 100 groups of average color difference to form a set C (representing the variance distribution of the instruction delay scenario), simulate a light emitting mechanism fault scenario, trigger non-systematic color deviation by artificially manufacturing different mechanism faults (such as randomly selecting 3-5 mechanisms for color decay processing, simulating poor line contact), and also collect variance data of at least 100 groups of average color difference to form a set D (representing the variance distribution of the light emitting mechanism fault scenario). Statistical analysis is performed on the set C and the set D, the maximum value of the set C (i.e., the maximum variance that may occur under the instruction delay scenario); the minimum value of the set D (i.e., the minimum variance that may occur under the light emitting mechanism fault scenario); and the critical gap of the distribution interval of the two is the preset variance threshold. For example, the maximum value of the set C is 5, the minimum value of the set D is 15, and the preset variance threshold is 10. However, the above values are not limited thereto, and a person skilled in the art can determine the values according to actual conditions.

[0084] The application determines the cause of color deviation based on the variance of average color difference. The color deviation caused by instruction input delay is systematic (all light emitting mechanisms are affected by the same delay factor, the deviation is consistent, and the variance is small); and the deviation caused by the failure of the light emitting mechanism itself is random (different mechanisms have different failure degrees and types, the deviation has no uniform rule, and the variance is large). This distinction based on data volatility makes the determination result more accurate and avoids confusing systematic problems with individual hardware problems. According to the comparison between the variance and the preset threshold, the direction of processing is directly pointed to the correction of the instruction output duration (instruction delay) or the repair of the light emitting mechanism (mechanism failure), which ensures that subsequent operations do not deviate from the core problem and reduces invalid processing. Through variance analysis, it can be directly determined whether the fault is caused by the instruction system or the hardware mechanism, avoiding the tedious process of checking all possible links one by one, and greatly shortening the fault positioning time.

[0085] Specifically, when the analysis unit generates the processing mode of correcting the instruction output duration, the instruction output delay is corrected according to the difference between the preset matching rate and the actual matching rate. The correction amount of the instruction output delay is positively correlated with the difference.

[0086] The correction amount of the instruction output delay in the embodiment of the present application is the product of the first proportional coefficient and the difference value, and the first proportional coefficient can be determined by the following method: fixing other conditions (such as illumination parameters and detection objects), artificially setting different degrees of instruction output delay, recording the actual matching rate under the corresponding scene; calculating the difference value under each delay, and recording the delay correction amount required to restore the actual matching rate to the preset matching rate threshold, taking the difference value as the independent variable and the delay correction amount as the dependent variable, establishing a scatter plot, and fitting the relationship between the two by linear regression, and the slope of the linear equation obtained by fitting is the first proportional coefficient, but the above value is not limited thereto, and a person skilled in the art can take the value according to the actual situation.

[0087] Specifically, when adjusting the illumination brightness of the light emitting mechanism, the analysis unit adjusts the operating power of the light emitting mechanism according to the average color difference of each image information, and the adjustment amount of the operating power of the light emitting mechanism is positively correlated with the average color difference.

[0088] The adjustment amount of the operating power of the light emitting mechanism in the embodiment of the present application is the product of the second proportional coefficient and the average color difference, and the determination method of the second proportional coefficient is the same as that of the first proportional coefficient, which will not be described here.

[0089] Specifically, in the process of adjusting the illumination brightness, the analysis unit is also used for correcting the brightness according to the number of feature points, and the correction amount of the brightness is positively correlated with the number of feature points.

[0090] The correction amount of the brightness in the embodiment of the present application is the product of the third proportional coefficient and the number of feature points, and the determination method of the third proportional coefficient is the same as that of the first proportional coefficient, which will not be described here.

[0091] The instruction output delay correction, light emitting power adjustment and brightness correction in the present application are all quantitatively adjusted through the linear relationship of the proportional coefficient and the variable, which directly correlates the correction amount with the deviation degree, and the greater the deviation, the stronger the adjustment, so that the system parameters can be quickly adjusted to a reasonable range. In the brightness adjustment, not only the average color difference is considered, but also the number of feature points is introduced for secondary correction, so that the adjustment is more in line with the actual scene, and the overall and accuracy of the correction are improved.

[0092] Specifically, after completing the brightness adjustment, if the light emitting unit is still unqualified, the analysis unit controls to repeatedly adjust the brightness until the critical condition is reached, and the critical condition includes the maximum cumulative adjustment times or the critical brightness; when the critical condition is reached, a processing mode for shortening the distance between the collection unit and the light emitting unit is generated.

[0093] The maximum cumulative adjustment times in the embodiment of the present application can be determined by the following method: based on experimental data and statistical analysis of actual adjustment effect, in a laboratory environment, typical scenarios (such as excessive high / low brightness, non-uniform brightness, etc.) of the light-emitting unit caused by brightness deviation are simulated, the light-emitting mechanism is adjusted multiple times (according to the preset brightness gradient, such as ±5% power each time), and the average abnormal proportion change after each adjustment is recorded; the correlation between the adjustment times and the average abnormal proportion decrease is determined by analyzing the experimental data; for example, if the abnormal proportion decreases significantly (such as ≥10% each time) after the first three adjustments; the decrease is <3% after the fourth adjustment (considered as weak adjustment effect); the abnormal proportion does not change or even increases after the fifth adjustment (considered as invalid adjustment), the maximum cumulative adjustment times can be set to 4-5 times (preferably 4 times); the critical brightness is the upper limit or lower limit of the brightness adjustment of the light-emitting mechanism, and beyond this range will cause damage to the light-emitting mechanism, failure of the lighting effect or not meeting the actual application requirements.

[0094] Specifically, when the analysis unit generates the processing mode of shortening the distance, the shortening distance is determined according to the difference between the target brightness value and the critical brightness value, and the shortening amount of the distance is positively correlated with the difference; the analysis unit re-detects after shortening the distance, and if the light-emitting unit is still unqualified, it is determined that the system is malfunctioning, and a maintenance notification is generated.

[0095] In the embodiment of the present application, the shortening amount of the distance is the product of the fourth proportional coefficient and the difference between the target brightness value and the critical brightness value, and the determination method of the fourth proportional coefficient is the same as that of the first proportional coefficient, which will not be repeated here.

[0096] The present application first adjusts the brightness repeatedly, then adjusts physically when invalid, and finally intervenes manually. This hierarchical processing mode reduces unnecessary operation cost and equipment loss, and the critical condition provides a clear boundary. The maximum cumulative adjustment times and the critical brightness set a reasonable boundary for brightness adjustment, prevent the system from being trapped in invalid adjustment in an infinite loop, protect the light-emitting unit, and save system resources; the distance shortening amount is positively correlated with the difference between the target brightness value and the critical brightness value, the larger the brightness gap, the larger the distance adjustment amplitude, which can quickly make up for the lack of brightness; the light-emitting unit is re-detected after shortening the distance, and if it is still unqualified, it is determined that the system is malfunctioning, which avoids the situation that the problem is not solved after distance adjustment; when all automatic adjustment means fail, a maintenance notification is sent in time, and the problem is handed over to manual processing, ensuring the integrity of fault handling.

[0097] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but it is easy for those skilled in the art to understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after the changes or replacements will all fall within the protection scope of the present application.

[0098] The above only describes the preferred embodiments of the present application and is not intended to limit the present application; for those skilled in the art, the present application can have various changes and variations, and any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A self-learning-based embedded nano-light source intelligent control system, characterized in that, include: Matrix; A light-emitting unit, which is connected to the substrate, includes several light-emitting mechanisms embedded in the substrate; An instruction output unit, which is connected to each of the light-emitting mechanisms, is used to receive test instructions and control each light-emitting mechanism to adjust the lighting parameters according to the test instructions; The acquisition unit is used to periodically acquire image information of the substrate surface, extract several feature points corresponding to the light-emitting mechanism from each image information, and obtain the actual illumination parameters of each feature point; A storage unit, connected to the acquisition unit, is used to store the expected image information corresponding to each acquisition cycle. Each expected image information includes several preset feature points and corresponding preset lighting parameters. The verification unit is connected to the storage unit and the acquisition unit respectively. It is used to list the acquired image information and the corresponding expected image information at the same time point into a single image group, compare the illumination parameters of each feature point in the image group with the corresponding preset feature point, and filter each image information according to the comparison result. An analysis unit, connected to the verification unit, is used to determine whether the operation of the light-emitting unit is qualified based on the comparison results of the verification unit, and to determine the cause of the failure based on the average abnormality ratio when it is unqualified, so as to generate a corresponding processing method. The control unit, which is connected to the analysis unit and the instruction output unit respectively, is used to control the instruction output unit to output relevant instructions in subsequent actual use according to the processing method generated by the analysis unit, so as to adjust the operating parameters of the light-emitting unit to the target value.

2. The intelligent control system for embedded nano-light sources based on self-learning according to claim 1, characterized in that, The process by which the analysis unit determines whether the light-emitting unit is operating correctly includes: Calculate the matching rate, which is the ratio of the number of image information that matches the corresponding expected image information to the total number of image information; For a single image, if the ratio of the number of matching feature points to the total number of feature points reaches a preset value, then the image is determined to match the corresponding expected image. For a single feature point, if the similarity between its actual lighting parameters and the preset lighting parameters of the corresponding preset feature point is higher than the preset similarity, then the feature point is determined to be a matching feature point. If the matching rate is higher than the preset matching rate threshold, the light-emitting unit is deemed to be operating successfully. If the matching rate is lower than or equal to the preset matching rate threshold, the light-emitting unit is determined to be unqualified, and the average number of abnormal feature points in each image information is counted. The reason for the unqualification is determined based on the average abnormality ratio.

3. The intelligent control system for embedded nano-light sources based on self-learning according to claim 2, characterized in that, The average anomaly percentage is the average of the anomaly percentages in each image information; wherein, the anomaly percentage of a single image information is the ratio of the number of mismatched feature points to the total number of feature points in that image information.

4. The intelligent control system for embedded nano-light sources based on self-learning according to claim 3, characterized in that, The process by which the analysis unit determines the cause of non-compliance based on the average percentage of anomalies includes: If the average abnormality rate is greater than the preset abnormality rate threshold, the reason for non-compliance is determined to be the instruction output unit. The unit first adjusts the lighting color of the light-emitting mechanism. If the light color adjustment still fails to meet the requirements, the lighting brightness is further adjusted. If the average abnormality percentage is less than or equal to the preset abnormality percentage threshold, the reason for non-compliance is determined to be a single light-emitting mechanism, and the line connecting the light-emitting mechanism is inspected.

5. The intelligent control system for embedded nano-light sources based on self-learning according to claim 4, characterized in that, The analysis unit is also used for: Calculate the average color difference of each image information, where the color difference is the difference between the actual chromaticity value of the feature point and the preset chromaticity value of the corresponding preset feature point; Calculate the variance of each of the aforementioned average color differences, and determine whether the color deviation is caused by instruction input delay based on the variance: If the variance is less than a preset variance threshold, it is determined that the color deviation is caused by instruction delay, and a processing method to correct the instruction output duration of the instruction output unit is generated. If the variance is greater than or equal to the preset variance threshold, it is determined that the color deviation is caused by the light-emitting mechanism, and a repair procedure for the light-emitting mechanism is generated.

6. The intelligent control system for embedded nano-light sources based on self-learning according to claim 5, characterized in that, When the analysis unit generates the processing method for the correction instruction output duration, it corrects the instruction output delay based on the difference between the preset matching rate and the actual matching rate. The amount of correction for the instruction output delay is positively correlated with the difference.

7. The intelligent control system for embedded nanolight sources based on self-learning according to claim 6, characterized in that, When the illumination brightness of the light-emitting mechanism is adjusted, the analysis unit adjusts the operating power of the light-emitting mechanism according to the average color difference of each image information. The adjustment amount of the operating power of the light-emitting mechanism is positively correlated with the average color difference.

8. The intelligent control system for embedded nano-light sources based on self-learning according to claim 7, characterized in that, During the process of adjusting the lighting brightness, the analysis unit is also used to correct the brightness based on the number of feature points, and the amount of brightness correction is positively correlated with the number of feature points.

9. The intelligent control system for embedded nano-light sources based on self-learning according to claim 8, characterized in that, After the analysis unit completes the brightness adjustment, if the light-emitting unit is still unqualified, it controls the repeated adjustment of brightness until a critical condition is reached. The critical condition includes the maximum cumulative number of adjustments or the critical brightness. When the critical condition is reached, a processing method to shorten the distance between the acquisition unit and the light-emitting unit is generated.

10. The intelligent control system for embedded nanolight sources based on self-learning according to claim 9, characterized in that, When the analysis unit generates the processing method for shortening the distance, it determines the shortening distance based on the difference between the target brightness value and the critical brightness value. The amount of distance shortening is positively correlated with the difference. After shortening the distance, the analysis unit re-detects. If the light-emitting unit is still unqualified, it determines that the system has malfunctioned and generates a processing method to issue a maintenance notice.

Images