Exposure data adjusting method and device and computer equipment

Abstract

The invention relates to an exposure data adjusting method and device and computer equipment. The method comprises the following steps: determining histogram data and image quality data of a current image; determining first exposure data of the current image according to the histogram data; and determining second exposure data of the current image according to the image quality data, fusing the first exposure data and the second exposure data to obtain expected exposure data; and adjusting the first exposure data according to the difference condition between the first exposure data and the expected exposure data. By adopting the technical scheme, the accuracy of the adjustment result of the first exposure data is improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to an exposure data adjustment method, device and computer equipment. Background Art

[0002] As the requirements for vehicle intelligence continue to increase, the usage rate of vehicle-mounted cameras in vehicles is also gradually increasing, providing convenience for vehicle assisted driving, safety monitoring, and driving records.

[0003] Existing vehicle-mounted cameras determine the exposure data of the corresponding image frame by continuously monitoring the image histogram information of each frame of image generated by the ISP (Image Signal Processor) based on the AEC (Automatic Exposure Control) algorithm, and reversely adjust the exposure data based on the difference between the exposure data and the expected exposure data.

[0004] However, the above method is limited by the limited information carried by the histogram, resulting in a problem of low accuracy in the determined exposure data adjustment result. Summary of the invention

[0005] Based on this, it is necessary to provide an exposure data adjustment method, apparatus and computer equipment that can improve the accuracy of exposure data adjustment results in response to the above technical problems.

[0006] In a first aspect, the present application provides an exposure data adjustment method, comprising:

[0007] Determining histogram data and image quality data of the current image;

[0008] Determining first exposure data of the current image according to the histogram data; and

[0009] determining second exposure data of the current image according to the image quality data;

[0010] fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0011] The first exposure data is adjusted according to a difference between the first exposure data and the expected exposure data.

[0012] In one embodiment, determining the image quality data of the current image includes: obtaining texture distribution features obtained by extracting features from the image to be processed; wherein the image to be processed includes the current image and a reference image; and determining the image quality data of the current image based on the difference between the texture distribution features of the current image and the texture distribution features of the reference image.

[0013] In one of the embodiments, the texture distribution characteristics of the image to be processed are determined in the following manner: extracting target texture characteristics of the image to be processed; based on a multivariate Gaussian model, performing distribution feature extraction on the target texture characteristics of the image to be processed to obtain the texture distribution characteristics of the image to be processed; wherein the multivariate Gaussian model is obtained by fitting the target texture characteristics extracted from at least one sample image.

[0014] In one of the embodiments, extracting the target texture features of the image to be processed includes: extracting the first texture features of the image to be processed based on a generalized Gaussian distribution; and extracting the second texture features of the image to be processed in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; wherein the target texture features include the first texture features and / or the second texture features.

[0015] In one of the embodiments, the method further includes: downsampling the image to be processed to obtain a derivative image of the image to be processed; extracting a third texture feature of the derivative image based on a generalized Gaussian distribution; and extracting a fourth texture feature of the derivative image in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; wherein the target texture feature also includes the third texture feature and / or the fourth texture feature.

[0016] In one of the embodiments, extracting the target texture features of the image to be processed includes: dividing the image to be processed into at least one candidate image block; for each candidate image block, determining the sharpness data of the candidate image block according to the distribution of different pixel values ​​in the candidate image block; selecting a target image block from the at least one candidate image block according to the sharpness data of each candidate image block; and extracting the target texture features of the target image block as the target texture features of the image to be processed.

[0017] In one of the embodiments, determining the second exposure data of the current image according to the image quality data includes: determining the second exposure data of the current image according to the image quality data based on a target mapping function; wherein the target mapping function is generated in the following manner: determining reference exposure data of a reference image captured under preset quality acquisition requirements, and a reference image quality of the reference image; performing polynomial fitting based on the reference exposure data and corresponding reference image qualities of at least two reference images to obtain the target mapping function.

[0018] In one of the embodiments, the method also includes: obtaining the acquisition environment type of the current image; selecting a candidate mapping function corresponding to the acquisition environment type from at least one candidate mapping function as the target mapping function; wherein different candidate mapping functions generate corresponding reference images with different acquisition environment types.

[0019] In a second aspect, the present application further provides an exposure data adjustment device, comprising:

[0020] A first determination module, used to determine histogram data and image quality data of a current image;

[0021] a second determining module, configured to determine first exposure data of the current image according to the histogram data; and

[0022] A third determining module, configured to determine second exposure data of the current image according to the image quality data;

[0023] a data fusion module, configured to fuse the first exposure data and the second exposure data to obtain expected exposure data;

[0024] A data adjustment module is used to adjust the first exposure data according to the difference between the first exposure data and the expected exposure data.

[0025] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0026] Determining histogram data and image quality data of the current image;

[0027] Determining first exposure data of the current image according to the histogram data; and

[0028] determining second exposure data of the current image according to the image quality data;

[0029] fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0030] The first exposure data is adjusted according to a difference between the first exposure data and the expected exposure data.

[0031] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0032] Determining histogram data and image quality data of the current image;

[0033] Determining first exposure data of the current image according to the histogram data; and

[0034] determining second exposure data of the current image according to the image quality data;

[0035] fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0036] The first exposure data is adjusted according to a difference between the first exposure data and the expected exposure data.

[0037] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0038] Determining histogram data and image quality data of the current image;

[0039] Determining first exposure data of the current image according to the histogram data; and

[0040] determining second exposure data of the current image according to the image quality data;

[0041] fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0042] The first exposure data is adjusted according to a difference between the first exposure data and the expected exposure data.

[0043] The above exposure data adjustment method, device and computer equipment introduce the image quality data and histogram data of the current image to replace the traditional technology of determining the expected exposure data based only on the histogram data, so that the expected exposure data can fully consider the impact of the image quality data, thereby improving the accuracy of the determined expected exposure data. Correspondingly, based on the difference between the first exposure data and the expected exposure data, the first exposure data is adjusted, so that the adjustment process of the exposure parameters is more reasonable, thereby improving the accuracy of the first exposure parameter adjustment result, and further helping to improve the image quality of the captured image when the image is captured based on the adjusted result. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 is a flow chart of an exposure data adjustment method in one embodiment;

[0046] Figure 2A is a flow chart of a step of determining image quality data in one embodiment;

[0047] Figure 2B is a schematic diagram of a pixel extension direction in one embodiment;

[0048] Figure 2C is a schematic diagram of a screening result of a target image block in an embodiment;

[0049] Figure 2D is a schematic diagram of a screening result of another target image block in one embodiment;

[0050] Figure 3 is a schematic flow chart of a step of determining second exposure data in one embodiment;

[0051] Figure 4 is a flow chart of an exposure data adjustment method according to another embodiment;

[0052] Figure 5 is a structural block diagram of an exposure data adjustment device in one embodiment;

[0053] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0055] In an exemplary embodiment, Figure 1 As shown, an exposure data adjustment method is provided. This embodiment takes the method applied to a terminal as an example. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0056] S110, determining histogram data and image quality data of the current image.

[0057] The current image can be understood as the image currently captured by the image acquisition device that needs to adjust the exposure parameters. The histogram data is used to characterize the distribution of pixel brightness in the current image, which may include at least one of the image brightness distribution, image pixel distribution and histogram type data. This application does not make any new restrictions on this. The image quality data is used to evaluate the image quality from at least one dimension. This application does not make any restrictions on the evaluation dimension, for example, it may include the image clarity dimension.

[0058] It is worth noting that the histogram data and image quality data of the current image can be determined by at least one method in the prior art, and the present application does not impose any limitation on the specific technical means used to determine the two.

[0059] In an optional embodiment, the current image can be directly acquired by a computing device that executes the exposure data adjustment method, and the histogram data and image quality data of the current image can be determined. Alternatively, in order to reduce the amount of data calculation of the computing device that executes the exposure data adjustment method, thereby improving the adjustment efficiency of the exposure data adjustment by the computing device, the histogram data and image quality data of the current image can be determined by an image acquisition device that acquires the current image, or other devices that are communicatively connected to the image acquisition device and the computing device that executes the exposure data adjustment method, and when the computing device that executes the exposure data adjustment method needs to adjust the exposure data, the histogram data and image quality data can be directly acquired from the corresponding device.

[0060] S120, determining first exposure data of the current image according to the histogram data.

[0061] It is worth noting that at least one of the conventional technologies may be used to determine the first exposure data of the current image according to the histogram data, and the present application does not impose any limitation on the method for determining the first exposure data.

[0062] For example, the exposure parameter information carried in the histogram data may be directly read as the first exposure data of the current image, or the first exposure data of the current image may be determined based on the histogram data by calling the exposure parameter interface.

[0063] S130, determining second exposure data of the current image according to the image quality data.

[0064] Exemplarily, a mapping relationship between different image quality data and exposure data may be pre-constructed, and when exposure data adjustment is required, the image quality data of the current image is used to search for exposure data corresponding to the mapping relationship as the second exposure data.

[0065] The above mapping relationship may be constructed artificially or obtained through a large number of experiments, and this application does not impose any limitation on this.

[0066] It is worth noting that the numerical value of the aforementioned first exposure data reflects the distribution of image brightness, while the numerical value of the second exposure data reflects the quality of the image.

[0067] It should be noted that S120 may be executed before or after S130, or both may be executed simultaneously or crosswise, and the present application does not impose any limitation on the specific execution order of the two.

[0068] S140, fusing the first exposure data and the second exposure data to obtain expected exposure data.

[0069] The expected exposure data is used to characterize the idealized optimal exposure data corresponding to the current image, that is, a clear image of the current image that meets the preset quality acquisition requirements and is acquired by the image acquisition device under the same image acquisition conditions.

[0070] In an optional embodiment, the first exposure data and the second exposure data may be directly summed, and the summed result is used as the expected exposure data.

[0071] In another optional embodiment, the weights of the first exposure data and the second exposure data may be preset, and the weighted sum of the first exposure data and the second exposure data under the corresponding weights may be used as the expected exposure data. The weights of the first exposure data and the second exposure data may be set or adjusted by a technician according to needs or experience, or may be determined repeatedly through a large number of experiments, and it is only necessary to ensure that the sum of the two is 1, and this application does not impose any restrictions on this. In a specific implementation, the weight of the first exposure data may be 0.7, and the weight of the second exposure data may be 0.3.

[0072] S150: Adjust the first exposure data according to the difference between the first exposure data and the expected exposure data.

[0073] The first exposure data is exposure data determined only based on the histogram data, which has a certain one-sidedness and can reflect the actual exposure parameter setting of the image acquisition device corresponding to the current image when performing image acquisition. The expected exposure number is the exposure data after the first exposure data and the second exposure data are fused, which can more accurately reflect the exposure parameter setting requirements of the current image when the preset quality acquisition requirements are met. Therefore, the difference between the first exposure data and the expected exposure data can be used to determine whether the first exposure data needs to be adjusted. Furthermore, in the case where the first exposure data needs to be adjusted, the adjustment direction and adjustment range can also be numerically quantified according to the specific difference value between the first exposure data and the expected exposure data.

[0074] In an optional embodiment, the absolute value of the difference between the first exposure data and the expected exposure data can be used as the difference data to quantify the difference between the two; the exposure difference interval to which the difference data belongs is determined, and the first exposure data is adjusted according to the adjustment direction and adjustment range preset in the exposure difference interval. The adjustment direction and adjustment range corresponding to different exposure difference intervals can be set by technicians according to needs or experience, or determined repeatedly through a large number of experiments, and this application does not impose any limitation on this.

[0075] In a specific implementation, if the difference data is less than the first difference threshold, there is no need to adjust the first exposure data; if the difference data is not less than the first difference threshold and less than the second difference threshold, the first exposure data is adjusted according to the first adjustment range; if the difference data is not less than the second difference threshold and less than the third difference threshold, the first exposure data is adjusted according to the second adjustment range; if the difference data is not less than the third difference threshold, the result obtained by amplifying the current exposure data by the first preset multiple can be used as the third adjustment range to adjust the first exposure data. The first difference threshold is less than the second difference threshold, and the second difference threshold is less than the third difference threshold; the first adjustment range is less than the second adjustment range; and the second adjustment range is less than the third adjustment range. At least one of the first preset multiple, the first difference threshold, the second difference threshold, the third difference threshold, the first adjustment range, and the second adjustment range can be set or adjusted by a technician according to needs or experience, or repeatedly determined through a large number of experiments, and this application does not impose any limitation on this.

[0076] In another specific implementation, when it is determined that the first exposure data needs to be adjusted, the adjustment direction of the first exposure data may be determined according to the relative size relationship between the first exposure data and the expected exposure data.

[0077] The embodiment of the present application introduces the image quality data and histogram data of the current image, replacing the conventional technique of determining the expected exposure data based only on the histogram data, so that the expected exposure data can fully consider the impact of the image quality data, thereby improving the accuracy of the determined expected exposure data. Accordingly, based on the difference between the first exposure data and the expected exposure data, the first exposure data is adjusted, so that the adjustment process of the exposure parameters is more reasonable, thereby improving the accuracy of the adjustment result of the first exposure parameters, and further helping to improve the image quality of the captured image when the image is captured based on the adjusted result.

[0078] Based on the technical solutions of the above embodiments, the present application also provides an optional embodiment. In this optional embodiment, the step of determining the image quality data in S110 is refined to improve the convenience of determining the image quality data result.

[0079] See also Figure 2A The steps for determining the image quality data shown include:

[0080] S210, obtaining texture distribution features obtained by extracting features from an image to be processed; wherein the image to be processed includes a current image and a reference image.

[0081] Among them, the reference image can be understood as an image whose image quality meets the preset quality requirements, such as a clear image whose clarity meets the clarity requirements. Among them, the texture distribution characteristics of the image to be processed are used to characterize the distribution of texture features in the image to be processed, and may include the mean vector and covariance matrix of the texture features, etc., and this application does not impose any restrictions on this. Among them, the mean vector reflects the distribution center position of the texture features in the image to be processed, corresponding to the center position of the color, brightness and edge texture of the entire image; the covariance matrix reflects the correlation between different features in the texture features, as well as the shape, direction and spatial relationship of objects in the image.

[0082] Exemplarily, the computing device executing the exposure data adjustment method can directly extract features from the image to be processed to obtain corresponding texture distribution features. Alternatively, in order to reduce the amount of data calculation of the computing device and improve the efficiency of exposure data adjustment, other devices that are in communication with the computing device can also extract texture distribution features from the image to be processed, and when exposure data adjustment is required, the corresponding texture distribution features can be searched and obtained from the device.

[0083] In an optional embodiment, the target texture features of the image to be processed may be directly extracted, and distribution features of the target texture features of the image to be processed may be extracted based on a multivariate Gaussian model to obtain texture distribution features of the image to be processed.

[0084] Among them, the target texture features of the image to be processed carry the key texture information of the image to be processed, and represent the richness and diversity of the information carried in the image to be processed.

[0085] Optionally, a trained texture feature extraction network may be directly used to extract target texture features of the image to be processed.

[0086] In an optional embodiment, the target texture feature may include a first texture feature; accordingly, the first texture feature of the image to be processed may be extracted based on a generalized Gaussian distribution.

[0087] The first texture feature may carry at least one of the information of the shape, kurtosis and distribution width of the texture distribution curve in the image to be processed.

[0088] Among them, Generalized Gaussian Distribution (GGD) is a continuous probability distribution, which is a generalized form of normal distribution and Laplace distribution. It has a flexible form and the advantage of simulating different data distribution characteristics by adjusting its shape parameters.

[0089] The following probability density function of the generalized Gaussian distribution can be used to determine the distribution characteristics of the image to be processed:

[0090] ;

[0091] ;

[0092] ;

[0093] in, and are the shape parameters of the distribution; is the width parameter of the distribution, corresponding to the standard deviation or local standard deviation of the image to be processed; is a real or complex number; is a gamma function that is used to ensure the normalization of the distribution.

[0094] For example, the shape parameters can be pre-set Set the search interval and search step size to give the shape parameters For example, Set it to [0.2:0.001:10], that is, in the search interval [0.2, 10], with a step size of 0.001, The setting of candidate values.

[0095] All of the above The candidate values ​​are substituted into the following ratio function :

[0096] ;

[0097] Use the following formula to compare the functions Make an estimate and get an estimated value :

[0098] ;

[0099] ;

[0100] ;

[0101] Among them, the random variable X(x1,x2...,x m ) corresponds to the normalized pixel values ​​of the image to be processed.

[0102] Sure and The distance and the shape parameter of the case with the minimum distance are calculated Corresponding ; and determine the shape parameters based on the following formula Corresponding :

[0103] .

[0104] Accordingly, we can generate , shape parameters and width parameters The first texture feature.

[0105] It can be understood that the texture distribution curve of the image to be processed is presented by introducing the generalized Gaussian distribution. Due to the different causes of image distortion in the image quality assessment process, the shape of the generalized Gaussian distribution curve will be quite different. Therefore, in view of the above characteristics of the Gaussian generalized distribution, the abnormal distribution in the image to be processed can be presented and used as a constituent element of the target texture feature of the image to be processed to participate in subsequent operations.

[0106] In another optional embodiment, the target texture feature may include a second texture feature; accordingly, the second texture feature of the image to be processed in at least one pixel extension direction may be extracted based on an asymmetric generalized Gaussian distribution.

[0107] The second texture feature may carry at least one of the information such as the shape, center position and diffusion degree of the texture distribution curve in the image to be processed. The second texture features in different pixel extension directions can reflect the texture correlation in the corresponding pixel extension directions. The number of pixel extension directions is at least one, and can usually include at least two, so as to improve the richness and comprehensiveness of the second texture feature.

[0108] In a specific implementation, see Figure 2B As shown in the schematic diagram of pixel extension direction, for a certain pixel point I(i, j) of the image to be processed, the following four directions can be used as the pixel extension directions of the second texture feature: I(i, j)I(i+1, j), I(i, j)I(i-1, j+1), I(i, j)I(i+1, j+1) and I(i, j)I(i, j+1).

[0109] Among them, Asymmetric Generalized Gaussian Distribution (AGGD) is an important probability distribution function, which can describe the distribution of data with asymmetric characteristics.

[0110] For any pixel extension direction, the distribution characteristics of the image to be processed can be determined based on the following probability density function:

[0111] ;

[0112] Among them, x is a random variable, corresponding to each pixel of the normalized image to be processed; is the shape parameter of the distribution, which controls the peak value of the distribution, that is, the sharpness of the distribution; is the scale parameter of the negative semi-axis, controlling the width of the left side of the distribution; is the scale parameter of the positive semi-axis, controlling the width of the right side of the distribution; and reflects the degree of diffusion of the distribution. The two can be different, thus reflecting the asymmetry of the distribution. Among them, when x≤0, the density of the distribution increases with the increase of -x according to The ratio of decays rapidly; when x ≥ 0, the density of the distribution increases as x increases according to The ratio decays rapidly.

[0113] For central location This distribution feature can be presented using the following formula:

[0114] ;

[0115] in, and The distribution is the gamma function in and The gamma function can be found in the above content and will not be described here.

[0116] For example, the shape parameters can be pre-set Set the search interval and search step size to give the shape parameters For example, Set it to [0.2:0.001:10], that is, in the search interval [0.2, 10], with a step size of 0.001, The setting of candidate values.

[0117] All of the above The candidate values ​​are substituted into the following ratio function :

[0118] ;

[0119] Use the following formula to determine the scale parameters of the negative semi-axis of the asymmetric generalized Gaussian grouping: and the scale parameter of the positive semi-axis :

[0120] ;

[0121] ;

[0122] in, is the number of samples in the random variable whose pixel values ​​are located on the negative semi-axis; is the number of samples in the random variable whose pixel values ​​are on the positive semi-axis.

[0123] The asymmetry of the asymmetric generalized Gaussian distribution can be estimated using the following formula:

[0124] .

[0125] The above ratio function is given by the following formula: Make an estimate and get the estimated :

[0126] ;

[0127] ;

[0128] ;

[0129] in, is the mean of the absolute values ​​of the random variables; is the mean of the squares of the random variables.

[0130] Sure and The distance and the shape parameter of the case with the minimum distance are calculated Corresponding ; and determine the scale parameter based on the following formula Corresponding , and the scale parameter Corresponding :

[0131] ;

[0132] ;

[0133] Accordingly, for any pixel extension direction, a shape parameter can be generated. , scale parameters and , and central location These four second texture features.

[0134] It should be noted that for each image to be processed, a first texture feature including a shape parameter and a width parameter can be generated, and a second texture feature including a shape parameter, a center position parameter, a negative semi-axis width parameter, and a positive semi-axis width parameter (i.e., a diffusion degree parameter) can be generated in each pixel extension direction. If there are 4 pixel extension directions, a total of 18 target texture features can be generated, including 2 first texture features and 4×4 second texture features.

[0135] In order to improve the richness of the extracted target texture features, further, for each image to be processed, the image to be processed can be downsampled to obtain a derivative image of the image to be processed, and texture features can be extracted based on the derivative image, thereby achieving the purpose of enriching the target texture features.

[0136] Among them, the number of downsampling times and the downsampling sampling interval can be set or adjusted by technical personnel according to needs or experience values, and this application does not impose any limitations on this.

[0137] Exemplarily, the third texture feature of the derived image can be extracted based on the generalized Gaussian distribution; and the fourth texture feature of the derived image in at least one pixel extension direction can be extracted based on the asymmetric generalized Gaussian distribution; wherein the target texture feature also includes the third texture feature and / or the fourth texture feature.

[0138] Taking downsampling by 1 / 2 once to generate a derivative image as an example, the derivative image can be used as a new random variable to replace the aforementioned image to be processed to re-extract texture features and obtain new target texture features.

[0139] In a specific implementation, for the derived image, if there are 4 pixel extension directions, a total of 18 target texture features can be generated, including 2 third texture features and 4×4 fourth texture features.

[0140] It is worth noting that when extracting the target texture features, the feature extraction can be performed on the entire image to be processed, or on a local area in the image to be processed. In order to reduce the amount of data calculation in the feature extraction process, in an optional embodiment, the image to be processed can be divided into at least one candidate image block; a target image block is selected from the at least one candidate image block; and only the target texture features of the target image block are extracted as the target texture features of the image to be processed.

[0141] Exemplarily, for each candidate image block, the sharpness data of the candidate image block may be determined according to the distribution of different pixel values ​​in the candidate image block; and the target image block may be selected from at least one candidate image block according to the sharpness data of each candidate image block.

[0142] The sharpness data reflects the regional changes of the information carried in the candidate image block.

[0143] In a specific implementation, the image to be processed can be divided into at least one candidate image block according to a preset size, and an index value (b=1, 2, ..., B) of each candidate image block is established; for each candidate image block, the sharpness data of the candidate image block is determined according to the average value of the standard deviation (or local standard deviation) corresponding to each pixel point in the candidate image block; the sharpness data of each candidate image block is sorted, and the sharpness screening threshold is determined according to the sorting result; and the candidate image block whose sharpness data is greater than the sharpness screening threshold is used as the target image block.

[0144] Optionally, the sharpness data at a preset percentile (e.g., 80%) in the ascending sorting result may be used as the sharpness screening threshold. Alternatively, the product of the maximum value in the sharpness data and a preset percentage (e.g., 80%) may be used as the sharpness screening threshold.

[0145] See also Figure 2C and Figure 2D The schematic diagram of the screening result of the target image block is shown, wherein the image block retained in the figure is the target image block, and compared with the area of ​​the non-target image block, the target image block carries relatively rich texture information.

[0146] It is worth noting that the above-mentioned normalization process of the image to be processed can be implemented by using at least one traditional normalization method.

[0147] In an optional embodiment, the original acquired image to be processed may be normalized in the following manner:

[0148] ;

[0149] in, is the pixel mean of the image to be processed; is the pixel standard deviation of the image to be processed; is a preset value to avoid the denominator being 0, for example, it can be 1; The pixel point of the image to be processed The pixel value on .

[0150] It can be understood that the image normalization result obtained by normalizing the image to be processed in the above manner has a low dependence on the strength of the image texture, and the image features subsequently extracted have higher practicality.

[0151] The pixel mean and pixel standard deviation may be implemented by using a traditional mean and standard deviation determination method, or by using a local image processing method, and this application does not impose any limitation on this.

[0152] In a specific implementation, the following formula can be used to construct a Gaussian model for the image, and determine the local mean and local standard deviation. Accordingly, the local mean can be used as the aforementioned pixel mean, and the local standard deviation can be used as the aforementioned pixel standard deviation.

[0153] ;

[0154] ;

[0155] Wherein, K is the extended length of the pixel point in the length direction of the image; L is the extended length of the pixel point in the height direction of the image; The number of the corresponding pixel.

[0156] The above is a detailed description of the process of extracting target texture features of the image to be processed. The following is an explanation of the multivariate Gaussian model and its use process.

[0157] In an optional embodiment, the multivariate Gaussian model is obtained by fitting target texture features extracted from at least one sample image.

[0158] The sample image may be an image that is acquired under preset quality acquisition conditions and meets quality requirements.

[0159] Exemplarily, the target texture features of the image to be processed can be input into the expression of the multivariate Gaussian model; the expression of the multivariate Gaussian model is estimated, for example, by maximum likelihood estimation, to obtain the mean vector and covariance matrix of the model as the texture distribution features of the image to be processed.

[0160] In a specific implementation, the multivariate Gaussian model can be expressed using the following formula:

[0161] ;

[0162] in, is the multivariate Gaussian model function; is the target texture feature containing k elements; is the mean vector of the multivariate Gaussian model; is the covariance matrix of the multivariate Gaussian model.

[0163] S220, determining image quality data of the current image according to the difference between the texture distribution characteristics of the current image and the texture distribution characteristics of the reference image.

[0164] Exemplarily, the distance between the texture distribution feature of the current image and the texture distribution feature of the reference image can be determined, so as to numerically quantify the difference between the texture distribution features of the two through the distance value, and determine the image quality data of the current image based on the distance value. The image quality data changes monotonically with the change of the distance value, for example, the image quality data can increase monotonically or decrease monotonically with the distance value.

[0165] In a specific implementation, the following formula may be used to determine the image quality data of the current image:

[0166] ;

[0167] in, is the image quality data of the current image; and are the mean vector and covariance matrix of the reference image respectively; and are the mean vector and covariance matrix of the current image respectively.

[0168] The embodiment of the present application introduces the texture distribution characteristics of the current image and the texture distribution characteristics of the reference image, and determines the image quality data of the current image based on the difference between the two. The determination process is convenient and fast, and the amount of data calculation is small, which helps to improve the execution efficiency of the exposure data adjustment method.

[0169] Based on the technical solutions of the above embodiments, the present application also provides an optional embodiment. In this optional embodiment, the step of determining the second exposure data in S130 is refined to improve the accuracy of the second exposure data determination result.

[0170] See also Figure 3 The step of determining the second exposure data shown includes:

[0171] S310, determining second exposure data of the current image based on the target mapping function and the image quality data.

[0172] The target mapping function is used to characterize the mapping transformation relationship between the image quality data and the exposure data, and through the mapping transformation relationship, the second exposure data corresponding to different image quality data can be determined. The input data of the target mapping function corresponds to the image quality data; the output data of the target mapping function corresponds to the second exposure data of the current image.

[0173] Exemplarily, the image quality data is used as input data of the target mapping function to obtain second exposure data of the current image.

[0174] In one embodiment, the target mapping function can be generated in the following manner: determining the reference exposure data of a reference image captured under preset quality acquisition requirements, and the reference image quality of the reference image; performing polynomial fitting based on the reference exposure data and corresponding reference image qualities of at least two reference images to obtain the target mapping function.

[0175] The reference image acquired under the preset quality acquisition requirement can be understood as a standard clear image, that is, an image whose corresponding image quality data meets the requirements. The reference exposure data of the reference image can be understood as the exposure data used when acquiring the reference image under the preset quality acquisition requirement, or it can also be the exposure data determined based on the histogram data of the reference image. The reference image quality can be understood as the image quality data corresponding to the reference image determined by the image quality data determination step in the aforementioned embodiment.

[0176] Exemplarily, the exposure data of multiple reference images can be used as independent variables, the quality of the corresponding reference images can be used as dependent variables, polynomial fitting can be performed based on a preset polynomial model, and the fitting result can be used as the target mapping function.

[0177] Since there are usually certain differences in the quality of images collected in different image acquisition scenarios, for example, in driving scenarios, the road environments are different, such as highways, urban roads, rural roads, garages, tunnels, etc., and in high-light environments such as day and night. For the same area, the differences in image quality and exposure parameters of reference images that meet the preset quality acquisition requirements are relatively significant. Therefore, when constructing the mapping function between exposure data and image quality data, it is also possible to introduce differentiated treatment of scene-related factors.

[0178] In an optional embodiment, the reference exposure data of the reference images captured under preset quality acquisition requirements and the reference image quality of the reference images can be determined in advance for different acquisition environment types (corresponding to different scenes) in the corresponding scenes respectively; for the same acquisition environment type, polynomial fitting is performed according to the reference exposure data and the corresponding reference image quality of each reference image under the acquisition environment type to obtain a candidate mapping function under the acquisition environment type.

[0179] Accordingly, in the process of determining the second exposure data of the current image, the acquisition environment type of the current image can also be obtained in advance; from at least one candidate mapping function, a candidate mapping function corresponding to the acquisition environment type is selected as the target mapping function. Among them, different candidate mapping function generation formulas have different acquisition environment types for reference images. The advantage of doing so is that the candidate mapping functions can be constructed in a differentiated manner for different acquisition environment types. Accordingly, in the process of determining the second exposure data, the acquisition environment type of the current image is introduced, which can avoid the influence of different acquisition environment types on the determined second exposure data, improve the accuracy of the second exposure data determination result, thereby improving the accuracy of the expected exposure data determination result, and further helping to improve the accuracy of the first exposure data adjustment result.

[0180] The embodiment of the present application obtains a target mapping function by introducing reference exposure data of at least two reference images and corresponding reference image quality for polynomial fitting, thereby directly determining the second exposure data based on the image quality data based on the target mapping function. The above scheme is used to replace the traditional table lookup method to determine the second exposure data, which has higher computational efficiency and does not require a large amount of manpower and material resources to construct a data table in advance, thereby reducing the initial cost investment for determining the second exposure data.

[0181] Based on the technical solutions of the above embodiments, the present application also provides an optional embodiment, in which the exposure data adjustment method is described in detail.

[0182] See also Figure 4 An exposure data adjustment method shown includes:

[0183] S401, determining histogram data of the current image.

[0184] S402, dividing the image blocks: dividing the image to be processed into at least one candidate image block. The image to be processed includes a current image and a reference image.

[0185] S403, screening image blocks: for each candidate image block, determining sharpness data of the candidate image block according to the distribution of different pixel values ​​in the candidate image block; and selecting a target image block from at least one candidate image block according to the sharpness data of each candidate image block.

[0186] S404, extracting texture features: based on a generalized Gaussian distribution, extracting a first texture feature of a target image block in an image to be processed; wherein the image to be processed includes a current image and a reference image; based on an asymmetric generalized Gaussian distribution, extracting a second texture feature of a target image block in an image to be processed in at least one pixel extension direction; downsampling the image to be processed to obtain a derivative image of the target image block in the image to be processed; based on a generalized Gaussian distribution, extracting a third texture feature of the target image block in the derivative image; based on an asymmetric generalized Gaussian distribution, extracting a fourth texture feature of the target image block in the derivative image in at least one pixel extension direction; wherein the target texture features of the image to be processed include the first texture feature, the second texture feature, the third texture feature and the fourth texture feature.

[0187] S405, determining image quality data: extracting distribution features of target texture features of the image to be processed based on a multivariate Gaussian model to obtain texture distribution features of the image to be processed; and determining image quality data of the current image based on the difference between the texture distribution features of the current image and the texture distribution features of the reference image. The multivariate Gaussian model is obtained by fitting the target texture features extracted from at least one sample image.

[0188] S406, determining expected exposure data: determining first exposure data of the current image according to the histogram data; obtaining the acquisition environment type of the current image; selecting a candidate mapping function corresponding to the acquisition environment type from at least one candidate mapping function as a target mapping function; determining second exposure data of the current image according to the image quality data based on the target mapping function; fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0189] The types of acquisition environments of reference images corresponding to different candidate mapping functions are different. The target mapping function is generated in the following manner: determining reference exposure data of a reference image acquired under preset quality acquisition requirements and the reference image quality of the reference image; performing polynomial fitting based on the reference exposure data and the corresponding reference image quality of at least two reference images to obtain the target mapping function.

[0190] S407, adjusting exposure parameters: adjusting the first exposure data according to the difference between the first exposure data and the expected exposure data.

[0191] When the first exposure data needs to be adjusted, the target gain coefficient can be first determined based on the difference between the first exposure data and the expected exposure data; the target gain coefficient can be divided according to the preset damping coefficient to obtain the current gain coefficient for each gain adjustment; based on the current gain data, gain is allocated to different exposure parameters in the first exposure data; and the allocated exposure parameters are made effective in a preset time window, thereby realizing gradual adjustment of the first exposure data.

[0192] It should be noted that each step-by-step adjustment may correspond to a preset number of frames of images, such as a single frame of image. The preset damping coefficient may be determined based on the frame rate of the image acquisition device and the acquisition scene in which the image acquisition device is located.

[0193] Exemplarily, the target gain coefficient may be determined by the following formula:

[0194] ;

[0195] in, is the expected exposure data; is the first exposure data; is the target gain factor.

[0196] Exemplarily, the following formula may be used to determine the current gain coefficient for adjusting exposure data one by one:

[0197] ;

[0198] in, is the preset damping coefficient; is the target gain coefficient; is the previous gain coefficient; is the current gain coefficient of the current time.

[0199] For example, for a vehicle-mounted camera with a frame rate of 30FPS, the preset damping coefficient can be first set to the first coefficient (such as 0.2, that is, 80% of the old parameters and 20% of the new parameters), and after the cumulative adjusted image frames reach the first preset number (such as 10) frames, the preset damping coefficient is adjusted to the second coefficient (such as 0.9, that is, the proportion of the old parameters drops to 10%). Furthermore, after the cumulative adjusted image frames reach the second preset number (such as 30) frames, the preset damping coefficient is set to the third coefficient (such as 1, that is, ignoring the proportion of the old parameters), so as to gradually converge the new parameters to the target gain coefficient. For the application scenarios of vehicle-mounted cameras, especially in high-speed scenarios, the dynamic changes are fast, and it is generally necessary to transition to the ideal brightness value within 5 frames.

[0200] In an optional embodiment, when performing gain allocation for different exposure parameters in the first exposure data according to the current gain data, the current gain data can be mapped to control parameters such as sensor (image sensor in the image acquisition device) exposure time, sensor gain, ISP gain, etc. Among them, the exposure time can improve the signal-to-noise ratio of the image, and a common strategy is to use up the exposure time as much as possible within the constraint boundary, and then allocate it to the sensor gain, and finally allocate the remaining gain to the ISP digital gain.

[0201] It is worth noting that the above-mentioned decomposed control parameters need to take effect synchronously in order for the picture to obtain the expected exposure brightness, and all parameters need to take effect within a specific time window, that is, the time when the previous frame image has ended and the next frame image has not yet begun, that is, the vertical blanking window of the sensor. This time window is relatively short, generally around 3 to 5 milliseconds.

[0202] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0203] Based on the same inventive concept, the embodiment of the present application also provides an exposure data adjustment device for implementing the exposure data adjustment method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the exposure data adjustment device provided below can refer to the limitations of the exposure data adjustment method above, and will not be repeated here.

[0204] In an exemplary embodiment, Figure 5 As shown, an exposure data adjustment device is provided, including: a first determination module 510, a second determination module 520, a third determination module 530, a data fusion module 540 and a data adjustment module 550. Among them,

[0205] A first determination module 510, configured to determine histogram data and image quality data of a current image;

[0206] A second determination module 520 is used to determine first exposure data of the current image according to the histogram data; and

[0207] A third determination module 530 is used to determine second exposure data of the current image according to the image quality data;

[0208] A data fusion module 540, configured to fuse the first exposure data and the second exposure data to obtain expected exposure data;

[0209] The data adjustment module 550 is used to adjust the first exposure data according to the difference between the first exposure data and the expected exposure data.

[0210] In one embodiment, the first determination module 510 includes: a first acquisition unit, used to obtain texture distribution features obtained by feature extraction of a to-be-processed image; wherein the to-be-processed image includes a current image and a reference image; and the first determination unit, used to determine image quality data of the current image based on the difference between the texture distribution features of the current image and the texture distribution features of the reference image.

[0211] In one embodiment, the first determination module 510 also includes: a first extraction unit, used to extract target texture features of the image to be processed; a second extraction unit, used to extract distribution features of the target texture features of the image to be processed based on a multivariate Gaussian model, to obtain texture distribution features of the image to be processed; wherein the multivariate Gaussian model is obtained by fitting the target texture features extracted from at least one sample image.

[0212] In one embodiment, the first extraction unit includes: a first extraction subunit, used to extract a first texture feature of the image to be processed based on a generalized Gaussian distribution; and a second extraction subunit, used to extract a second texture feature of the image to be processed in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; wherein the target texture feature includes the first texture feature and / or the second texture feature.

[0213] In one embodiment, the first extraction unit also includes: a downsampling subunit, used to downsample the image to be processed to obtain a derivative image of the image to be processed; a third extraction subunit, used to extract a third texture feature of the derivative image based on a generalized Gaussian distribution; and a fourth extraction subunit, used to extract a fourth texture feature of the derivative image in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; wherein the target texture feature also includes the third texture feature and / or the fourth texture feature.

[0214] In one of the embodiments, the first extraction unit includes: a first division subunit, used to divide the image to be processed into at least one candidate image block; a first determination subunit, used to determine the sharpness data of each candidate image block according to the distribution of different pixel values ​​in the candidate image block; a first selection subunit, used to select a target image block from at least one candidate image block according to the sharpness data of each candidate image block; and a fifth extraction subunit, used to extract target texture features of the target image block as target texture features of the image to be processed.

[0215] In one embodiment, the third determination module 530 includes: a second determination unit, used to determine the second exposure data of the current image based on the target mapping function and the image quality data; wherein the target mapping function is generated in the following manner: determining the reference exposure data of the reference image captured under the preset quality acquisition requirements, and the reference image quality of the reference image; performing polynomial fitting based on the reference exposure data and the corresponding reference image qualities of at least two reference images to obtain the target mapping function.

[0216] In one embodiment, the third determination module 530 also includes: a second acquisition unit, used to obtain the acquisition environment type of the current image; a second selection unit, used to select a candidate mapping function corresponding to the acquisition environment type from at least one candidate mapping function as the target mapping function; wherein different candidate mapping functions generate different acquisition environment types for corresponding reference images.

[0217] Each module in the above-mentioned exposure data adjustment device can be implemented in whole or in part by software, hardware or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0218] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 6As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, an exposure data adjustment method is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0219] In an alternative embodiment, the computer device may be a vehicle-mounted rearview mirror.

[0220] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0221] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0222] Determining histogram data and image quality data of the current image;

[0223] Determining first exposure data of the current image according to the histogram data; and

[0224] determining second exposure data of the current image according to the image quality data;

[0225] Fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0226] The first exposure data is adjusted according to the difference between the first exposure data and the expected exposure data.

[0227] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining texture distribution features obtained by extracting features from an image to be processed; wherein the image to be processed includes a current image and a reference image; and determining image quality data of the current image based on the difference between the texture distribution features of the current image and the texture distribution features of the reference image.

[0228] In one embodiment, when the processor executes the computer program, the following steps are also implemented: extracting target texture features of the image to be processed; based on a multivariate Gaussian model, performing distribution feature extraction on the target texture features of the image to be processed to obtain texture distribution features of the image to be processed; wherein the multivariate Gaussian model is obtained by fitting the target texture features extracted from at least one sample image.

[0229] In one embodiment, when the processor executes the computer program, the following steps are also implemented: based on the generalized Gaussian distribution, extracting the first texture feature of the image to be processed; and based on the asymmetric generalized Gaussian distribution, extracting the second texture feature of the image to be processed in at least one pixel extension direction; wherein the target texture feature includes the first texture feature and / or the second texture feature.

[0230] In one embodiment, when the processor executes the computer program, the following steps are also implemented: downsampling the image to be processed to obtain a derivative image of the image to be processed; extracting a third texture feature of the derivative image based on a generalized Gaussian distribution; and extracting a fourth texture feature of the derivative image in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; wherein the target texture feature also includes the third texture feature and / or the fourth texture feature.

[0231] In one embodiment, when the processor executes the computer program, the following steps are also implemented: dividing the image to be processed into at least one candidate image block; for each candidate image block, determining the sharpness data of the candidate image block according to the distribution of different pixel values ​​in the candidate image block; selecting a target image block from at least one candidate image block according to the sharpness data of each candidate image block; and extracting the target texture features of the target image block as the target texture features of the image to be processed.

[0232] In one embodiment, when the processor executes the computer program, the following steps are also implemented: based on the target mapping function, according to the image quality data, second exposure data of the current image is determined; wherein the target mapping function is generated in the following manner: reference exposure data of a reference image captured under preset quality acquisition requirements and a reference image quality of the reference image are determined; polynomial fitting is performed based on the reference exposure data of at least two reference images and the corresponding reference image qualities to obtain the target mapping function.

[0233] In one embodiment, when the processor executes the computer program, it also implements the following steps: obtaining the acquisition environment type of the current image; selecting a candidate mapping function corresponding to the acquisition environment type from at least one candidate mapping function as the target mapping function; wherein different candidate mapping functions generate different acquisition environment types for the corresponding reference images.

[0234] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0235] Determining histogram data and image quality data of the current image;

[0236] Determining first exposure data of the current image according to the histogram data; and

[0237] determining second exposure data of the current image according to the image quality data;

[0238] Fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0239] The first exposure data is adjusted according to the difference between the first exposure data and the expected exposure data.

[0240] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: obtaining texture distribution features obtained by feature extraction of an image to be processed; wherein the image to be processed includes a current image and a reference image; and determining image quality data of the current image based on the difference between the texture distribution features of the current image and the texture distribution features of the reference image.

[0241] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: extracting target texture features of the image to be processed; based on a multivariate Gaussian model, performing distribution feature extraction on the target texture features of the image to be processed to obtain texture distribution features of the image to be processed; wherein the multivariate Gaussian model is obtained by fitting the target texture features extracted from at least one sample image.

[0242] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: based on a generalized Gaussian distribution, extracting a first texture feature of the image to be processed; and based on an asymmetric generalized Gaussian distribution, extracting a second texture feature of the image to be processed in at least one pixel extension direction; wherein the target texture feature includes the first texture feature and / or the second texture feature.

[0243] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: downsampling the image to be processed to obtain a derivative image of the image to be processed; extracting a third texture feature of the derivative image based on a generalized Gaussian distribution; and extracting a fourth texture feature of the derivative image in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; wherein the target texture feature also includes the third texture feature and / or the fourth texture feature.

[0244] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: dividing the image to be processed into at least one candidate image block; for each candidate image block, determining the sharpness data of the candidate image block according to the distribution of different pixel values ​​in the candidate image block; selecting a target image block from at least one candidate image block according to the sharpness data of each candidate image block; and extracting the target texture features of the target image block as the target texture features of the image to be processed.

[0245] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: based on the target mapping function, according to the image quality data, second exposure data of the current image is determined; wherein the target mapping function is generated in the following manner: reference exposure data of a reference image captured under preset quality acquisition requirements and a reference image quality of the reference image are determined; polynomial fitting is performed based on the reference exposure data of at least two reference images and the corresponding reference image qualities to obtain the target mapping function.

[0246] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining the acquisition environment type of the current image; selecting a candidate mapping function corresponding to the acquisition environment type from at least one candidate mapping function as the target mapping function; wherein different candidate mapping functions generate corresponding reference images with different acquisition environment types.

[0247] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0248] Determining histogram data and image quality data of the current image;

[0249] Determining first exposure data of the current image according to the histogram data; and

[0250] determining second exposure data of the current image according to the image quality data;

[0251] Fusing the first exposure data and the second exposure data to obtain expected exposure data;

[0252] The first exposure data is adjusted according to the difference between the first exposure data and the expected exposure data.

[0253] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: obtaining texture distribution features obtained by feature extraction of an image to be processed; wherein the image to be processed includes a current image and a reference image; and determining image quality data of the current image based on the difference between the texture distribution features of the current image and the texture distribution features of the reference image.

[0254] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: extracting target texture features of the image to be processed; based on a multivariate Gaussian model, performing distribution feature extraction on the target texture features of the image to be processed to obtain texture distribution features of the image to be processed; wherein the multivariate Gaussian model is obtained by fitting the target texture features extracted from at least one sample image.

[0255] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: based on a generalized Gaussian distribution, extracting a first texture feature of the image to be processed; and based on an asymmetric generalized Gaussian distribution, extracting a second texture feature of the image to be processed in at least one pixel extension direction; wherein the target texture feature includes the first texture feature and / or the second texture feature.

[0256] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: downsampling the image to be processed to obtain a derivative image of the image to be processed; extracting a third texture feature of the derivative image based on a generalized Gaussian distribution; and extracting a fourth texture feature of the derivative image in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; wherein the target texture feature also includes the third texture feature and / or the fourth texture feature.

[0257] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: dividing the image to be processed into at least one candidate image block; for each candidate image block, determining the sharpness data of the candidate image block according to the distribution of different pixel values ​​in the candidate image block; selecting a target image block from at least one candidate image block according to the sharpness data of each candidate image block; and extracting the target texture features of the target image block as the target texture features of the image to be processed.

[0258] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: based on the target mapping function, according to the image quality data, second exposure data of the current image is determined; wherein the target mapping function is generated in the following manner: reference exposure data of a reference image captured under preset quality acquisition requirements and a reference image quality of the reference image are determined; polynomial fitting is performed based on the reference exposure data of at least two reference images and the corresponding reference image qualities to obtain the target mapping function.

[0259] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining the acquisition environment type of the current image; selecting a candidate mapping function corresponding to the acquisition environment type from at least one candidate mapping function as the target mapping function; wherein different candidate mapping functions generate different acquisition environment types for the corresponding reference images.

[0260] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0261] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0262] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0263] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for adjusting exposure data, characterized in that: include: Determining histogram data and image quality data of the current image; Determining first exposure data of the current image according to the histogram data; as well as, determining second exposure data of the current image according to the image quality data; fusing the first exposure data and the second exposure data to obtain expected exposure data; The first exposure data is adjusted according to a difference between the first exposure data and the expected exposure data.

2. The method according to claim 1, characterized in that Determine image quality data for the current image, including: Acquire texture distribution features obtained by extracting features from an image to be processed; wherein the image to be processed includes the current image and a reference image; The image quality data of the current image is determined according to the difference between the texture distribution characteristics of the current image and the texture distribution characteristics of the reference image.

3. The method according to claim 2, characterized in that The texture distribution characteristics of the image to be processed are determined in the following way: Extracting target texture features of the image to be processed; Based on a multivariate Gaussian model, distribution features of target texture features of the image to be processed are extracted to obtain texture distribution features of the image to be processed; The multivariate Gaussian model is obtained by fitting target texture features extracted from at least one sample image.

4. The method according to claim 3, characterized in that The step of extracting target texture features of the image to be processed comprises: Extracting a first texture feature of the image to be processed based on a generalized Gaussian distribution; and Extracting a second texture feature of the image to be processed in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; Wherein, the target texture feature includes the first texture feature and / or the second texture feature.

5. The method according to claim 4, characterized in that The method further comprises: Downsampling the image to be processed to obtain a derivative image of the image to be processed; Extracting a third texture feature of the derivative image based on a generalized Gaussian distribution; and Extracting a fourth texture feature of the derivative image in at least one pixel extension direction based on an asymmetric generalized Gaussian distribution; Wherein, the target texture feature also includes the third texture feature and / or the fourth texture feature.

6. The method according to claim 3, characterized in that The step of extracting target texture features of the image to be processed comprises: Dividing the image to be processed into at least one candidate image block; For each candidate image block, determining sharpness data of the candidate image block according to the distribution of different pixel values ​​in the candidate image block; Selecting a target image block from the at least one candidate image block according to the sharpness data of each candidate image block; The target texture features of the target image block are extracted as the target texture features of the image to be processed.

7. The method according to any one of claims 1 to 6, characterized in that: The determining, according to the image quality data, second exposure data of the current image comprises: Determining second exposure data of the current image based on the target mapping function and according to the image quality data; The target mapping function is generated in the following manner: Determining reference exposure data of a reference image captured under a preset quality capture requirement, and a reference image quality of the reference image; The target mapping function is obtained by performing polynomial fitting according to reference exposure data of at least two reference images and corresponding reference image qualities.

8. The method according to claim 7, characterized in that The method further comprises: Acquire the acquisition environment type of the current image; Selecting, from at least one candidate mapping function, a candidate mapping function corresponding to the acquisition environment type as the target mapping function; Among them, different candidate mapping functions are generated according to different acquisition environment types of corresponding reference images.

9. An exposure data adjustment device, characterized in that: include: A first determination module, used to determine histogram data and image quality data of a current image; A second determination module, configured to determine first exposure data of the current image according to the histogram data; as well as, A third determining module, configured to determine second exposure data of the current image according to the image quality data; a data fusion module, configured to fuse the first exposure data and the second exposure data to obtain expected exposure data; A data adjustment module is used to adjust the first exposure data according to the difference between the first exposure data and the expected exposure data.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images