Gamma correction method and system based on genetic algorithm

Abstract

The invention provides a Gamma correction method and system based on a genetic algorithm, and the method comprises the steps: S1, collecting individual parameters, and generating a gray-scale curve; s2, preliminarily optimizing the gray scale curve to obtain a preliminary optimization result; and S3, performing iterative optimization through a genetic algorithm to obtain a final optimization result. The invention provides a multi-objective collaborative optimization framework based on a genetic algorithm, carries out global optimization on digital driving parameters, and aims to synchronously improve the linear response precision of a gray-scale curve through an optimization algorithm.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of image processing, and particularly relates to a Gamma correction method and system based on a genetic algorithm. BACKGROUND

[0002] The driving current of an OLED device is not only dependent on the drain current, but also affected by transient current pulses caused by its equivalent capacitance, which specifically manifests as abnormal current spikes in the opening / closing stage of the device. In a sub-field driving mode, the actual brightness of a single frame of image is represented by the time-domain integration of all the instantaneous brightness of sub-fields in the frame period.

[0003] In traditional digital sub-field driving, a fractional sub-field has a blanking period after being turned on because its effective light-emitting time is less than that of a reference sub-field. This feature causes any subsequent sub-field displayed by the fractional sub-field to necessarily undergo a switching transient process and induce additional current pulses. This mechanism makes the cumulative brightness generated by the continuous opening of multiple fractional sub-fields higher than the brightness output of a single integer sub-field independently turned on. The compensation mechanism based on DTSED prolongs the device life by adjusting the sub-field switching time length, but significantly increases the switching action frequency in unit time. The intensification of switching transients increases the charging / discharging frequency of the equivalent capacitance, thereby introducing more current pulses and amplifying the brightness deviation.

[0004] At present, as a transformation technology in the field of image signal processing, the core of Gamma correction is to correct the response mismatch between the photoelectric conversion characteristics of the display system and the visual perception characteristics of the human eye. However, when the picture brightness experiences a local drop to the gray scale interval where the original value is restored, forming an invalid gray scale region, it directly leads to the attenuation of the available sampling data quantity of Gamma correction. The high linearity of the ideal gray scale curve enables Gamma correction to be implemented through a low-complexity value searching algorithm, significantly reducing the time cost and implementation difficulty. Moreover, the gray scale nonlinearity of a silicon-based OLED micro display significantly restricts its display accuracy.

[0005] Patent document CN111681607A discloses a Gamma adjustment method and system based on a genetic algorithm. This scheme does not use a multi-objective collaborative optimization framework and does not implement global optimization on digital driving parameters, and cannot simultaneously improve the linear response accuracy of the gray scale curve.

[0006] This problem needs to be solved urgently. SUMMARY

[0007] In view of the defects in the prior art, the purpose of the present application is to provide a Gamma correction method and system based on a genetic algorithm.

[0008] According to the Gamma correction method and system based on a genetic algorithm provided by the present application, the method comprises the following steps:

[0009] Step S1: collecting individual parameters, and generating a gray scale curve;

[0010] Step S2: preliminarily optimizing the gray scale curve, and obtaining a preliminary optimization result;

[0011] Step S3: iteratively optimizing by a genetic algorithm, and obtaining a final optimization result.

[0012] Preferably, in the step S1, the individual parameters include field weight, field sequence, lighting probability and compensation number; and the gray scale curve is generated by a brightness imaging model according to the individual parameters.

[0013] Preferably, in the step S2, the gray scale curve is preliminarily optimized by an adaptive function, and a mathematical expression of the adaptive function is:

[0014] minF(y1,y2)=f(s1,s2,...,s N ,w1,w2,...,w n ,v1,v2,v3,m1...,m M )

[0015] Wherein, minF(y1,y2) represents the adaptive function, y1 represents a root mean square difference, and y2 represents an invalid gray point number; L represents a function; s1, s2,..., s n all are corresponding field sequences; w1, w2,..., w n all are corresponding field weights, n represents a number of decimal fields; v1, v2, v3 all are compensation numbers; the compensation number is 0-2; m1..., m M all are lighting probabilities, wherein M is a number of fields; and N represents a number of objective functions.

[0016] Preferably, in the step S3, the genetic algorithm comprises:

[0017] Step A1: initializing a population as the preliminary optimization result, and setting Gen=1; Gen represents a generation;

[0018] Step A2: calculating a crowding degree of individuals in a current parent population P t , and then judging whether the crowding degree meets a preset threshold value of a task; if yes, a child population Q is generated; and if no, crossover and mutation are performed, and the step A2 is re-executed;

[0019] Step A3: obtaining a population R based on the child population Q and the parent population P t ;

[0020] Step A4: calculating a crowding degree of individuals in the population R, and generating a new parent population P t+1 ;

[0021] Step A5: Determine whether Gen is greater than the preset maximum generation threshold. If yes, output the solution set as the final optimization result; if no, re-execute step A2, increment the current value of Gen by one, and update the current parent population to the parent population obtained in step A4.

[0022] Preferably, the mathematical expression for the congestion degree is:

[0023]

[0024] Among them, i distance That is, id represents the full name of the sorting operation, m represents the number of sorting operations, and M represents the total number of sorting operations, Δm. distance 'Indicates the normalized result of the crowding distance increment;

[0025] The Δm distance The mathematical expression for ' is:

[0026]

[0027] Where, Δm distance Indicates the increase in crowding distance. Let N and N represent the maximum and minimum values ​​of the objective function, respectively, where N represents the number of objective functions, i.e., the population size.

[0028] The Δm distance The mathematical expression is:

[0029]

[0030] in, Let j+1 be the objective function. Let j represent the (j-1)th objective function, where j is the ordinal number of the individual.

[0031] A Gamma correction system based on a genetic algorithm provided by the present invention includes:

[0032] Module M1: Collects individual parameters and generates grayscale curves;

[0033] Module M2: Performs preliminary optimization of the grayscale curve to obtain preliminary optimization results;

[0034] Module M3: The final optimization result is obtained through iterative optimization using a genetic algorithm.

[0035] Preferably, in module M1, the individual parameters include: field weight, field order, illumination probability, and compensation count; based on the individual parameters, a grayscale curve is generated using a brightness imaging model.

[0036] Preferably, in module M2, the grayscale curve is initially optimized using an adaptive function, the mathematical expression of which is:

[0037] minF(y1,y2)=f(s1,s2,...,s2,w1,w2,...,w n ,v1,v2,v3,m1...,m M )

[0038] Where minF(y1,y2) represents the adaptive function, y1 represents the root mean square error, y2 represents the number of invalid grayscale points; f represents the function; s1,s2,...,s N All are corresponding field sequences; w1, w2, ..., w n All values ​​represent the corresponding field weights, where n represents the number of fractional fields; v1, v2, and v3 represent the number of compensation cycles; the number of compensation cycles ranges from 0 to 2; m1...,m M Both are the lighting probabilities, where M is the number of fields; N represents the number of objective functions.

[0039] Preferably, in module M3, the genetic algorithm includes:

[0040] Module A1: Initialize the population as the initial optimization result, let Gen = 1; Gen represents the generation;

[0041] Module A2: Calculate the current parent population P t The crowding of individuals in the population is determined, and it is then determined whether the crowding meets the preset threshold of the task. If the result is yes, a progeny population Q is generated. If the result is no, crossover and mutation are performed, and module A2 is triggered again.

[0042] Module A3: Based on offspring population Q and parent population P t The population R is obtained;

[0043] Module A4: Calculate the crowding density of individuals in population R and generate a new parent population P. t+1 ;

[0044] Module A5: Determine whether Gen is greater than the preset maximum generation threshold. If yes, output the solution set as the final optimization result; if no, re-trigger module A2, increment the current Gen value by one, and update the current parent population to the parent population obtained by module A4.

[0045] Preferably, the mathematical expression for the congestion degree is:

[0046]

[0047] Among them, i distanceThat is, id represents the full name of the sorting operation, m represents the number of sorting operations, and M represents the total number of sorting operations, Δm. distance ' represents the normalized result of the crowding distance increment.

[0048] The Δm distance The mathematical expression for ' is:

[0049]

[0050] Where, Δm distance Indicates the increase in crowding distance. Let N and N represent the maximum and minimum values ​​of the objective function, respectively, where N represents the number of objective functions, i.e., the population size.

[0051] The Δm distance The mathematical expression is:

[0052]

[0053] in, Let j+1 be the objective function. Let j represent the (j-1)th objective function, where j is the ordinal number of the individual.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] 1. This invention proposes a multi-objective collaborative optimization framework based on genetic algorithms to perform global optimization of digital driving parameters, aiming to improve the linear response accuracy of grayscale curves simultaneously through optimization algorithms.

[0056] 2. This invention proposes a crowding index to quantify the spatial distribution density of individuals within the same Pareto level.

[0057] 3. The genetic algorithm provided by this invention has a population containing multiple individuals, whose encoding consists of field sequence, field weight, lighting probability, and compensation number. By quantifying the spatial distribution density of individuals in the same Pareto level, the risk of probabilistically getting trapped in local optima is avoided. Attached Figure Description

[0058] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0059] Figure 1 This is a schematic diagram of the Pareto sorting method provided by the present invention;

[0060] Figure 2 This is a schematic diagram of the genetic algorithm provided by the present invention;

[0061] Figure 3This is a schematic diagram of the Pareto optimal frontier results provided by the present invention. Detailed Implementation

[0062] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0063] Gray-scale curve optimization based on a brightness model is essentially a nonlinear bi-objective optimization problem. The core challenge lies in the fact that optimizing the brightness of a specific gray level can cause nonlinear perturbations in the brightness of other gray levels. This system-level optimization requirement exceeds the capabilities of traditional single-point optimization; therefore, this invention employs a multi-objective optimization method. Furthermore, this invention uses a genetic algorithm as the global optimization method.

[0064] The genetic algorithm simulates the replication, crossover, and mutation operations of parent chromosome sets to achieve intergenerational transmission of superior genes. Starting with a randomly generated initial population, the algorithm iteratively cultivates high-fitness offspring individuals through selection, crossover, and mutation operations. In this continuous iteration process, the population individuals continuously evolve towards a better solution until the fitness values ​​of all individuals reach a set threshold, thus approximating the global optimum of the problem.

[0065] However, genetic algorithms still have inherent limitations:

[0066] 1. It consumes a lot of computing resources and iteration time, especially in high-dimensional solution space optimization scenarios.

[0067] 2. There is a risk of probabilistically getting trapped in local optima, and when the population diversity is insufficient, trap convergence is likely to occur.

[0068] The improved genetic algorithm includes: First, fast non-dominated sorting. This invention hierarchically classifies the solution set based on the dominance relationships between solutions, dividing all solutions into different Parent levels. Solutions within the same level are non-dominated, and solutions with lower levels are closer to the optimal front.

[0069] Specifically, suppose an individual i in population P is dominated by n(i) other individuals and simultaneously dominates S(i) individuals. The specific process is as follows:

[0070] Store all individuals with n(i) = 0 in set F1.

[0071] ② Initialize set H, traverse all individuals j in the population, execute n(j)-1, if n(j)-1=0, then execute H=HU{j}.

[0072] ③Assign all individuals in F1 a Pareto rank of 1, and then perform iterative ranking operations on the set H: sequentially assign increasing Pareto ranks to the individuals in H until all individuals in the population have completed hierarchical partitioning.

[0073] ④After fast non-dominated sorting, the Pareto front is divided into different ranks. As Figure 1 shown, in the case of two-objective optimization, the solution set is clearly divided into three Pareto ranks. Among them, Rank1 corresponds to the optimal Pareto front. All Pareto optimal solutions within each rank form the non-dominated solution set of that level, and the smaller the level value, the higher the superiority of the solution. Rank2 and Rank3 correspond to the sub-optimal Pareto and sub-sub-optimal Pareto ranks respectively.

[0074] In addition, the genetic algorithm maintains population diversity through a sharing function. However, a sharing radius parameter needs to be set. For this reason, a crowding degree index is proposed to quantify the spatial distribution density of individuals in the same Pareto rank.

[0075] Let the Pareto rank of individual i be irank and the crowding degree be id, and the Pareto rank of individual j be jrank and the crowding degree be jd. Then the comparison rules are as follows:

[0076] If irank < jrank, then individual i dominates individual j, and select i.

[0077] If irank = jrank, then judge by the crowding degree: when id < jd, select j.

[0078] If irank > jrank, then individual j dominates individual i, and select j.

[0079] The so-called crowding degree, the mathematical expression is:

[0080]

[0081] Among them, i distance That is, the full name of id, representing the crowding degree; m represents the number of sorting operations; M represents the total number of sorting operations, and Δm distance ' represents the normalized result of the crowding distance increment;

[0082] The so-called Δm distance The mathematical expression of' is:

[0083]

[0084] Among them, Δm distance represents the crowding distance increment, Let N and N represent the maximum and minimum values ​​of the objective function, respectively, where N represents the number of objective functions, i.e., the population size.

[0085] The Δm distance The mathematical expression is:

[0086]

[0087] in, Let j+1 be the objective function. Let j represent the (j-1)th objective function, where j is the ordinal number of the individual.

[0088] In summary, the process of a genetic algorithm includes:

[0089] Step 1: Initialize the population by setting Gen = 1;

[0090] Step 2: Calculate the current parent population P t The crowding of individuals in the population is determined, and it is then determined whether the crowding meets the preset threshold of the task. If the result is yes, a progeny population Q is generated. If the result is no, crossover and mutation are performed, and step 2 is executed again.

[0091] Step 3: Based on the offspring population Q and the parent population P t The population R is obtained;

[0092] Step 4: Calculate the crowding density of individuals in population R and generate a new parent population P. t+1 ;

[0093] Step 5: Determine if Gen is greater than the preset maximum generation threshold. If yes, output the solution set; if no, repeat step 2, increment the current value of Gen by one, and update the current parent population to the parent population obtained in step 4.

[0094] The nonlinearity of grayscale curves can be improved by the solution set.

[0095] Specifically, in grayscale curve optimization, the population of the genetic algorithm contains multiple individuals, and its encoding consists of field sequence, field weight, lighting probability, and compensation number.

[0096] The field sequence is a lighting arrangement sequence, using permutation encoding (PE). The field weights are a set of weights for fractional fields, satisfying the constraint that smaller weights are no greater than larger weights, and are represented as discrete values ​​using real-valued encoding (RE). The lighting probability is the probability of each field being turned on under different parameters, using real-valued encoding. The compensation count represents the number of times the device performs lifetime compensation, with a value between 0 and 2, and is also using real-valued encoding.

[0097] In this embodiment, the parameter system can achieve a 120Hz frame rate display with 65536 gray levels. The field sequence is guaranteed to be valid through permutation encoding, and the field weights are optimized for discrete values ​​through real number encoding.

[0098] The grayscale-luminance mapping is established based on a luminance model. The fitness function takes individual parameters as input, generates an actual grayscale curve by traversing the entire grayscale, and calculates the root mean square difference between the actual curve and the ideal curve and the number of invalid grayscale points as evaluation indicators. After substituting the individual parameters into the model, the function generates a curve by traversing the entire grayscale, and then uses the root mean square difference and invalid grayscale points to quantify the optimization objective, outputting it to a multi-objective genetic algorithm for iterative optimization.

[0099] In other words, the Gamma correction method based on a genetic algorithm provided by the present invention includes:

[0100] Step S1: Collect individual parameters and generate grayscale curves;

[0101] Step S2: Perform preliminary optimization of the grayscale curve to obtain preliminary optimization results;

[0102] Step S3: Iteratively optimize using a genetic algorithm to obtain the final optimization result.

[0103] In step S1, the individual parameters include: field weight, field order, illumination probability, and number of compensations; based on the individual parameters, a grayscale curve is generated using a brightness imaging model.

[0104] In step S2, the grayscale curve is initially optimized using an adaptive function, the mathematical expression of which is:

[0105] minF(y1,y2)=f(s1,s2,...,s N ,w1,w2,...,w n ,v1,v2,v3,m1...,m M )

[0106] Where minF(y1,y2) represents the adaptive function, y1 represents the root mean square error, y2 represents the number of invalid grayscale points; f represents the function; s1,s2,...,s n All are corresponding field sequences; w1, w2, ..., w n All values ​​represent the corresponding field weights, where n represents the number of fractional fields; v1, v2, and v3 represent the number of compensation cycles; the number of compensation cycles ranges from 0 to 2; m1...,m M Both are the lighting probabilities, where M is the number of fields; N represents the number of objective functions.

[0107] The results of the single-objective optimization of the root mean square difference in this embodiment are shown in Table 3.5, and the results of the single-objective optimization of the number of invalid gray points are shown in Table 3.6.

[0108] Table 3.5

[0109]

[0110] Table 3.6

[0111]

[0112] Data analysis in Tables 3.5 and 3.6 shows that while single-objective optimization strategies can directly improve specific indicators, they can lead to unsatisfactory results for the other objective, making it difficult to obtain the global optimum. Experimental data indicates that the parameter configuration with a population size of 60 and 1000 iterations significantly outperforms schemes with small populations and low iteration counts. A bi-objective optimization scheme was chosen to optimize the root mean square and invalid grayscale numbers. The results of the bi-objective optimization are shown in Table 3.7, where the optimal front formed by the non-dominated solution set is shown in the table. Figure 3 As shown in Table 3.7. Results of the genetic algorithm with root mean square error and invalid grayscale number as dual objectives.

[0113] Table 3.7: Results of Dual-Objective Optimization

[0114]

[0115] Of the seven Pareto optimal solutions obtained from the dual-objective optimization, the third solution was selected as the optimal equilibrium solution after objective normalization and calculation of the distance from each solution to the origin: distance value 0.203, root mean square error 4.36 cd / m. 2 There were 2793 invalid grayscale points, resulting in improvements of 85.5% and 94.82%, respectively. The optimal solution parameter configuration is detailed in Tables 3.8 and 3.9. The optimization results, verified by quantitative indicators, demonstrate the significant optimization effect of the dual-objective strategy on the original grayscale curve.

[0116] Table 3.8: Field Sequence Allocation

[0117]

[0118] Table 3.9: Subfield Weight Configuration

[0119]

[0120] Specifically, the first and third rows of Table 3.8 show the field activation order, and the second and fourth rows show the field numbers, where 1-10 represent decimal subfields and 11-25 represent integer subfields. Table 3.9 shows the weight distribution of the decimal subfields, where the maximum weight is 10239 and the minimum weight is 0, indicating a completely off state. The optimal driving parameter space parameters were input into the model for simulation, yielding grayscale curves. The results show that under the optimal parameter driving, the nonlinear characteristics of the grayscale curves are significantly improved, effectively verifying the ability of the optimization scheme to adjust the brightness grayscale mapping relationship.

[0121] The present invention also provides a Gamma correction system based on a genetic algorithm. The Gamma correction system based on a genetic algorithm can be implemented by executing the process steps of the Gamma correction method based on a genetic algorithm. That is, those skilled in the art can understand the Gamma correction method based on a genetic algorithm as a preferred embodiment of the Gamma correction system based on a genetic algorithm.

[0122] A Gamma correction system based on a genetic algorithm provided by the present invention includes:

[0123] Module M1: Collects individual parameters and generates grayscale curves;

[0124] Module M2: Performs preliminary optimization of the grayscale curve to obtain preliminary optimization results;

[0125] Module M3: The final optimization result is obtained through iterative optimization using a genetic algorithm.

[0126] Specifically, in module M1, the individual parameters include: field weight, field order, illumination probability, and number of compensations; based on the individual parameters, a grayscale curve is generated through a brightness imaging model.

[0127] Specifically, in module M2, the grayscale curve is initially optimized using an adaptive function, the mathematical expression of which is:

[0128] minF(y1,y2)=f(s1,s2,...,s N ,w1,w2,...,w n ,v1,v2,v3,m1...,m M )

[0129] Where minF(y1,y2) represents the adaptive function, y1 represents the root mean square error, y2 represents the number of invalid grayscale points; f represents the function; s1,s2,...,s N All are corresponding field sequences; w1, w2, ..., w nAll values ​​represent the corresponding field weights, where n represents the number of fractional fields; v1, v2, and v3 represent the number of compensation cycles; the number of compensation cycles ranges from 0 to 2; m1...,m M Both are the lighting probabilities, where M is the number of fields; N represents the number of objective functions.

[0130] Specifically, in module M3, the genetic algorithm includes:

[0131] Module A1: Initialize the population as the initial optimization result, let Gen = 1; Gen represents the generation;

[0132] Module A2: Calculate the current parent population P t The crowding of individuals in the population is determined, and it is then determined whether the crowding meets the preset threshold of the task. If the result is yes, a progeny population Q is generated. If the result is no, crossover and mutation are performed, and module A2 is triggered again.

[0133] Module A3: Based on offspring population Q and parent population P t The population R is obtained;

[0134] Module A4: Calculate the crowding density of individuals in population R and generate a new parent population P. t+1 ;

[0135] Module A5: Determine whether Gen is greater than the preset maximum generation threshold. If yes, output the solution set as the final optimization result; if no, re-trigger module A2, increment the current Gen value by one, and update the current parent population to the parent population obtained by module A4.

[0136] Specifically, the mathematical expression for the congestion degree is:

[0137]

[0138] Among them, i distance That is, id represents the full name of the sorting operation, m represents the number of sorting operations, and M represents the total number of sorting operations, Δm. distance 'Indicates the normalized result of the crowding distance increment;

[0139] The Δm distance The mathematical expression for ' is:

[0140]

[0141] Where, Δm distance Indicates the increase in crowding distance. Let N and N represent the maximum and minimum values ​​of the objective function, respectively, where N represents the number of objective functions, i.e., the population size.

[0142] The Δm distanceThe mathematical expression is:

[0143]

[0144] in, Let j+1 be the objective function. Let j represent the (j-1)th objective function, where j is the ordinal number of the individual.

[0145] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0146] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A Gamma correction method based on a genetic algorithm, characterized in that, include: Step S1: Collect individual parameters and generate grayscale curves; Step S2: Perform preliminary optimization of the grayscale curve to obtain preliminary optimization results; Step S3: Iteratively optimize using a genetic algorithm to obtain the final optimization result.

2. The Gamma correction method based on genetic algorithm according to claim 1, characterized in that, In step S1, the individual parameters include: field weight, field order, illumination probability, and number of compensations; based on the individual parameters, a grayscale curve is generated using a brightness imaging model.

3. The Gamma correction method based on genetic algorithm according to claim 1, characterized in that, In step S2, the grayscale curve is initially optimized using an adaptive function, the mathematical expression of which is: minF(y1,y2)=f(s1,s2,...,s N ,w1,w2,...,w n ,v1,v2,v3,m1...,m M ) Where minF(y1,y2) represents the adaptive function, y1 represents the root mean square error, y2 represents the number of invalid grayscale points; f represents the function; s1,s2,...,s N All are corresponding field sequences; w1, w2, ..., w n All values ​​represent the corresponding field weights, where n represents the number of fractional fields; v1, v2, and v3 represent the number of compensation cycles; the number of compensation cycles ranges from 0 to 2; m1...,m M Both are the lighting probabilities, where M is the number of fields; N represents the number of objective functions.

4. The Gamma correction method based on genetic algorithm according to claim 1, characterized in that, In step S3, the genetic algorithm includes: Step A1: Initialize the population as the initial optimization result, let Gen = 1; Gen represents the generation; Step A2: Calculate the current parent population P t The crowding degree of individuals in the population is determined, and it is then determined whether the crowding degree meets the preset threshold of the task. If the result is yes, a progeny population Q is generated. If the result is no, crossover and mutation are performed, and step A2 is executed again. Step A3: Based on the offspring population Q and the parent population P t The population R is obtained; Step A4: Calculate the crowding density of individuals in population R and generate a new parent population P. t+1 ; Step A5: Determine whether Gen is greater than the preset maximum generation threshold. If yes, output the solution set as the final optimization result; if no, re-execute step A2, increment the current value of Gen by one, and update the current parent population to the parent population obtained in step A4.

5. The Gamma correction method based on genetic algorithm according to claim 4, characterized in that, The mathematical expression for the congestion level is: Among them, i distance That is, id represents the full name of the sorting operation, m represents the number of sorting operations, and M represents the total number of sorting operations, Δm. distance ' represents the normalized result of the crowding distance increment. The Δm distance The mathematical expression for ' is: Where, Δm distance Indicates the increase in crowding distance. Let N and N represent the maximum and minimum values ​​of the objective function, respectively, where N represents the number of objective functions, i.e., the population size. The Δm distance The mathematical expression is: in, Let j+1 be the objective function. Let j represent the (j-1)th objective function, where j is the ordinal number of the individual.

6. A Gamma correction system based on a genetic algorithm, characterized in that, include: Module M1: Collects individual parameters and generates grayscale curves; Module M2: Performs preliminary optimization of the grayscale curve to obtain preliminary optimization results; Module M3: The final optimization result is obtained through iterative optimization using a genetic algorithm.

7. The Gamma correction system based on genetic algorithm according to claim 6, characterized in that, In module M1, the individual parameters include: field weight, field order, illumination probability, and number of compensations; based on the individual parameters, a grayscale curve is generated through a brightness imaging model.

8. The Gamma correction system based on genetic algorithm according to claim 6, characterized in that, In module M2, the grayscale curve is initially optimized using an adaptive function, the mathematical expression of which is: minF(y1,y2)=f(s1,s2,...,s N ,w1,w2,...,w n ,v1,v2,v3,m1...,m M ) Where minF(y1,y2) represents the adaptive function, y1 represents the root mean square error, y2 represents the number of invalid grayscale points; f represents the function; s1,s2,...,s N All are corresponding field sequences; w1, w2, ..., w n All values ​​represent the corresponding field weights, where n represents the number of fractional fields; v1, v2, and v3 represent the number of compensation cycles; the number of compensation cycles ranges from 0 to 2; m1...,m M Both are the lighting probabilities, where M is the number of fields; N represents the number of objective functions.

9. The Gamma correction system based on genetic algorithm according to claim 6, characterized in that, In module M3, the genetic algorithm includes: Module A1: Initialize the population as the initial optimization result, let Gen = 1; Gen represents the generation; Module A2: Calculate the current parent population P t The crowding of individuals in the population is determined, and it is then determined whether the crowding meets the preset threshold of the task. If the result is yes, a progeny population Q is generated. If the result is no, crossover and mutation are performed, and module A2 is triggered again. Module A3: Based on offspring population Q and parent population P t The population R is obtained; Module A4: Calculate the crowding density of individuals in population R and generate a new parent population P. t+1 ; Module A5: Determine whether Gen is greater than the preset maximum generation threshold. If yes, output the solution set as the final optimization result; if no, re-trigger module A2, increment the current Gen value by one, and update the current parent population to the parent population obtained by module A4.

10. The Gamma correction system based on genetic algorithm according to claim 9, characterized in that, The mathematical expression for the congestion level is: Among them, i distance That is, id represents the full name of the sorting operation, m represents the number of sorting operations, and M represents the total number of sorting operations, Δm. distance 'Indicates the normalized result of the crowding distance increment; The Δm distance The mathematical expression for ' is: Where, Δm distance Indicates the increase in crowding distance. Let N and N represent the maximum and minimum values ​​of the objective function, respectively, where N represents the number of objective functions, i.e., the population size. The Δm distance The mathematical expression is: in, Let j+1 be the objective function. Let j represent the (j-1)th objective function, where j is the ordinal number of the individual.

Images