Camera automatic exposure method and device based on fuzzy control and image recognition equipment

Abstract

The invention provides an automatic camera exposure method and device based on fuzzy control and image recognition equipment. The method comprises the following steps: firstly, determining a basic increment value of an exposure parameter according to basic image brightness; secondly, judging an ambient light state through an ambient light fuzzy rule and generating an ambient light estimation parameter, namely mathematical representation of ambient light; and converting the influence of the ambient light into a correction value of the exposure parameter by using the ambient light estimation parameter and a fuzzy control method, and finally obtaining the corrected exposure parameter. According to the method provided by the invention, a fuzzy control technology is adopted, calculation is carried out by setting a fuzzy rule conforming to human intuition and experience, exposure parameters are controlled when an environment state changes, and meanwhile, rapid convergence can be achieved without massive calculation; by using the method provided by the invention, automatic exposure can be realized in a scene with a complex and changeable environment, and a clear image can be acquired.

Description

Technical Field

[0001] The present application relates to the field of machine vision, and in particular to a camera automatic exposure method, device and image recognition device based on fuzzy control. Background Art

[0002] With the development of camera technology and artificial intelligence, vision cameras have been widely used. A key factor affecting the imaging quality of vision cameras is the exposure strategy of the cameras. The existing automatic exposure strategies mainly include the following several types:

[0003] (1) An automatic exposure scheme with exposure priority. The principle of this scheme is to collect an image and calculate the image brightness of the current frame, continuously adjust the exposure time according to the image brightness and collect the next frame of image until the image brightness meets the requirements. If the exposure time is adjusted to the limit value, the gain value is adjusted.

[0004] (2) An automatic exposure scheme with gain priority. The principle of this scheme is to collect an image and calculate the image brightness of the current frame, continuously adjust the gain value according to the image brightness and collect the next frame of image until the image brightness meets the requirements. If the gain value is adjusted to the limit value, the exposure time is adjusted.

[0005] (3) Establish a mapping relationship between the current brightness, exposure time, gain value, exposure time increment and gain value increment by means of a look-up table or a function expression.

[0006] The inventors of the present application found that the existing automatic exposure strategies have the following problems: in the above-mentioned schemes (1) and (2), it is difficult to set the increments of the exposure time and the gain value each time. If the parameter increment is too small, multiple adjustments are required to converge; if the parameter increment is too large, oscillations are likely to occur, resulting in an increase in the number of adjustments. Especially in some complex scenarios such as those with large changes in the intensity of ambient light, it is more difficult to determine the parameter increments of automatic exposure; although the above-mentioned scheme (3) can adjust the parameters once to reach the target brightness, when using the look-up table method, the established three-dimensional look-up table has high complexity, large volume and is difficult to apply; when using the function expression method, there are problems such as abstract function expressions and unclear parameter tuning directions in establishing traditional function models, and multiple trials and errors are required, consuming a large amount of time and effort. Summary of the Invention

[0007] According to a first aspect of the present application, a method for automatic exposure control of a camera based on fuzzy control is proposed. The method may include: setting at least one set of exposure parameters as basic exposure parameters; acquiring a first image with the basic exposure parameters and determining the basic image brightness according to the first image; determining a basic exposure increment parameter according to the basic image brightness; generating an ambient light estimation parameter according to the basic image brightness, the basic exposure parameters, the ambient light fuzzy rule, and the fuzzy set for ambient light estimation; generating an exposure increment correction parameter according to the ambient light estimation parameter, the exposure correction fuzzy rule, and the fuzzy set for exposure correction; generating a correction increment parameter according to the basic exposure increment parameter and the exposure increment correction parameter; determining a transition exposure parameter according to the correction increment parameter and the basic exposure parameters; and determining a target exposure parameter according to the transition exposure parameter.

[0008] According to a second aspect of the present application, a device for automatic exposure control of a camera based on fuzzy control is proposed. The device may include: a preprocessing module, which can be used to set at least one set of exposure parameters as basic exposure parameters; acquire a first image with the basic exposure parameters and determine the basic image brightness according to the first image; a basic automatic exposure module, which can be used to determine a basic exposure increment parameter according to the basic image brightness; an ambient light estimation module, which can be used to generate an ambient light estimation parameter according to the basic image brightness, the basic exposure parameters, the ambient light fuzzy rule, and the fuzzy set for ambient light estimation; an exposure correction module, which can be used to generate an exposure increment correction parameter according to the ambient light estimation parameter, the exposure correction fuzzy rule, and the fuzzy set for exposure correction; generating a correction increment parameter according to the basic exposure increment parameter and the exposure increment correction parameter; determining a transition exposure parameter according to the correction increment parameter and the basic exposure parameters; a parameter determination module, which can be used to determine a target exposure parameter according to the transition exposure parameter.

[0009] According to a third aspect of the present application, an image recognition device is proposed. The device may include the device for automatic exposure control of a camera based on fuzzy control as described in the second aspect of the present application.

[0010] According to a fourth aspect of the present application, an electronic device is proposed. The electronic device may include: a processor; a memory storing a computer program, and when the computer program is executed by the processor, the processor is caused to execute the method as described in the first aspect of the present application.

[0011] According to a fifth aspect of the present application, a non-transitory computer-readable storage medium is proposed. A computer-readable instruction is stored thereon, and when the instruction is executed by a processor, the processor is caused to execute the method as described in the first aspect of the present application.

[0012] The method proposed in this application first determines the basic increment value of the exposure parameter according to the basic image brightness; then judges the ambient light state through the ambient light blur rule and generates the ambient light estimation parameter, that is, the mathematical representation of the ambient light; uses the ambient light estimation parameter and the fuzzy control method to convert the influence of the ambient light into the correction value of the exposure parameter, and finally obtains the corrected exposure parameter. The method proposed in this application adopts the fuzzy control technology, calculates by setting fuzzy rules that conform to human intuition and experience, controls the exposure parameter when the environmental state changes, and can achieve fast convergence without a large amount of calculation. Using the method proposed in this application, automatic exposure can be realized in scenes with complex and changeable environments, and clear images can be captured. Brief Description of the Drawings

[0013] To more clearly illustrate the technical solutions in the embodiments of this application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following described drawings are only some embodiments of this application. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings without exceeding the scope of protection required by this application.

[0014] Figure 1 It is a schematic diagram of the steps of the camera automatic exposure method 1000 based on fuzzy control of this application;

[0015] Figure 2 For Figure 1 it is a schematic diagram of the steps of step S104 in the method 1000;

[0016] Figure 3 For Figure 2 it is a schematic diagram of the ambient light blur rule R100 in step S104 in

[0017] Figure 4 For Figure 2 it is a schematic diagram of the fuzzy set F100 for ambient light estimation in step S104 in

[0018] Figure 5 For Figure 1 it is a schematic diagram of the exposure correction fuzzy rule R200 in step S105 in

[0019] Figure 6 For Figure 1 it is a schematic diagram of the fuzzy set F200 for exposure correction in step S105 in

[0020] Figure 7 For Figure 1 it is a schematic diagram of the steps of step S108 in the method 1000 of

[0021] Figure 8Schematic diagram of the automatic exposure device 2000 for cameras based on fuzzy control according to the present application;

[0022] Figure 9 Schematic diagram of the image recognition device 3000 according to the present application;

[0023] Figure 10 Structural diagram of an electronic device according to the present application.

[0024] Description of reference numerals:

[0025] 2000: Automatic exposure device for cameras based on fuzzy control; 201: Preprocessing module; 202: Basic automatic exposure module; 203: Ambient light estimation module; 204: Exposure correction module; 205: Parameter determination module;

[0026] 3000: Image recognition device; 301: Automatic exposure device for cameras. Detailed implementation manners

[0027] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0028] Figure 1 Schematic diagram of the steps of the automatic exposure method 1000 for cameras based on fuzzy control according to the present application. As Figure 1 shown, the method 1000 includes steps S101 - step S108.

[0029] In step S101, the automatic exposure device for cameras based on fuzzy control (such as a processor with a camera, hereinafter referred to as the processor) sets at least one set of exposure parameters as the basic exposure parameters. In some specific embodiments, the basic exposure parameters are preset default values. In some specific embodiments, the exposure parameters include camera parameters that affect the brightness of the captured image, such as exposure time and gain value.

[0030] In step S102, the processor acquires the first image with the basic exposure parameters and determines the basic image brightness according to the first image. In some specific embodiments, the processor determines the basic image brightness according to the first image by selecting the image area of interest in the first image and calculating the average gray level of the image area. In some specific embodiments, the processor selects the central area in the first image as the image area of interest. In some specific embodiments, the processor selects the target detection object area in the first image as the image area of interest, such as the face frame area, palm area, etc.

[0031] In step S103, the processor determines a base exposure increment parameter according to the base image brightness. In some specific embodiments, in step S103, the processor uses an existing exposure optimization method to determine the base exposure increment according to the base image brightness. For example, the processor adjusts the exposure parameters using the exposure priority or gain priority method to determine the base exposure increment parameter.

[0032] In some specific embodiments, in step S103, the processor determines the base exposure increment parameter based on a fuzzy control method, specifically including generating the base exposure increment parameter according to the base image brightness, the exposure increment fuzzy rule, and the exposure increment fuzzy set. In some specific embodiments, the exposure increment fuzzy set includes a brightness fuzzy set and an exposure control fuzzy set.

[0033] In some specific embodiments, in step S103, the brightness fuzzy set represents the functional relationship of the membership degree of the image brightness and the image brightness state. The membership degree of the image brightness to different image brightness states represents the degree to which the image brightness belongs to the image brightness state. The exposure control fuzzy set represents the functional relationship of the increment value of the exposure parameter and the membership degree of the exposure parameter change direction. The exposure parameter change direction represents the change of the exposure parameter. For example, the exposure parameter change direction includes negative growth, maintaining, and positive growth.

[0034] In some specific embodiments, in step S103, the membership degree of the increment value of the exposure parameter to different exposure parameter change directions represents the degree to which the increment value of the exposure parameter belongs to the exposure parameter change direction. The exposure increment fuzzy rule represents the corresponding relationship between the exposure parameter change direction and the image brightness state. In some specific embodiments, in step S103, the processor can obtain the exposure parameter response set of the exposure control fuzzy set to the first membership degree according to the exposure increment fuzzy rule. In some specific embodiments, defuzzification is performed on the exposure parameter response set to obtain the base exposure increment parameter.

[0035] In step S104, the processor generates an ambient light estimation parameter according to the base image brightness, the base exposure parameter, the ambient light fuzzy rule, and the fuzzy set for ambient light estimation. In some specific embodiments, in step S104, the fuzzy set for ambient light estimation includes a target brightness fuzzy set and an exposure state fuzzy set.

[0036] In some specific embodiments, the processor determines a first membership degree data set according to the basic exposure parameters and the exposure status fuzzy set. In step S104, the exposure status fuzzy set represents the functional relationship between the value of the exposure parameter and the membership degree of the exposure parameter status. The exposure parameter status represents the numerical size of the exposure parameter, and the number of exposure parameter statuses can be set according to the requirements of the actual application. For example, in some specific embodiments, the number of exposure parameter statuses is 3, including small exposure parameters, general exposure parameters, and large exposure parameters. For example, in some specific embodiments, the number of exposure parameter statuses is 2, including large exposure parameters and small exposure parameters.

[0037] In some specific embodiments, the exposure parameter includes the exposure time. For example, the basic exposure time is 4 ms, the membership degree to the small exposure time is 0.56, and the membership degree to the large exposure time is 0.11. In some specific embodiments, the exposure parameter includes the gain value. For example, the basic gain value is 200, the membership degree to the small gain value is 0.5, and the membership degree to the large gain value is 0.5. At this time, {{0.56, 0.11}, {0.5, 0.5}} is the first membership degree data set representing the exposure parameter status.

[0038] In some specific embodiments, in step S104, the processor determines a second membership degree data set according to the basic image brightness and the target brightness fuzzy set. In step S104, the target brightness fuzzy set represents the functional relationship between the image brightness and the membership degree of the brightness status. The brightness status represents the relative brightness of the ambient light corresponding to the image brightness, and the number of brightness statuses can be set according to the requirements of the actual application. For example, in some specific embodiments, the number of brightness statuses is 3, including dark brightness, general brightness, and bright brightness. For example, in some specific embodiments, the number of brightness statuses is 2, including dark brightness and bright brightness.

[0039] The membership degree of the image brightness to different brightness statuses represents the degree to which the image brightness belongs to this brightness status. For example, in some specific embodiments, the brightness status includes dark brightness and bright brightness. When the basic image brightness is 60, the membership degree to the dark brightness is 0.2, and the membership degree to the bright brightness is 0. At this time, {0.2, 0} is the second membership degree data set.

[0040] In some specific embodiments, in step S104, the ambient light fuzzy rule represents the corresponding relationship between the ambient light estimation status and the exposure parameter status. For example, the ambient light fuzzy rule includes: when the image brightness is in the bright state and the exposure parameter is in the small state, the ambient light estimation status is in the bright state.

[0041] In some specific embodiments, in step S104, the processor evaluates the membership degrees corresponding to the image brightness state and the exposure parameter state respectively according to the ambient light blurring rule, and obtains the response value of the ambient light state as the ambient light estimation parameter. For example, the method by which the processor evaluates the membership degrees corresponding to the image brightness state and the exposure parameter state respectively to obtain the response value of the ambient light state can be adjusted according to the actual application, and different methods can be used for evaluation.

[0042] Optionally, the processor can obtain the response value of the ambient light state by taking the minimum value of the membership degrees corresponding to the image brightness state and the exposure parameter state respectively. For example, the membership degree of the image brightness in the bright state is 0.11, the membership degree of the exposure parameter state in the small state is 0.5, and the response value of the ambient light state in the bright state is min{0.11, 0.5}, that is, 0.11. Similarly, the ambient light estimation parameter including the ambient light state and the corresponding membership degree can be obtained. For example, the ambient light estimation parameter can be {(dark ambient light: 0.11), (bright ambient light: 0)}.

[0043] Optionally, the processor can obtain the response value of the ambient light state by deblurring and reblurring the membership degrees corresponding to the image brightness state and the exposure parameter state respectively.

[0044] In some specific embodiments, in step S105, the processor generates an exposure increment correction parameter according to the ambient light estimation parameter, the exposure correction blurring rule, and the fuzzy set for exposure correction. In step S105, the processor generates an exposure increment response correction parameter according to the ambient light estimation parameter, the exposure correction blurring rule, and the fuzzy set for exposure correction. In step S105, the processor deblurs the exposure increment response correction parameter to generate the exposure increment correction parameter. In some specific embodiments, in step S105, the processor uses the centroid method to deblur the gain value correction response parameter and the exposure time correction response parameter to obtain the gain value correction value and the exposure time correction value, which are the exposure increment correction parameters.

[0045] In step S106, the processor generates a correction increment parameter according to the basic exposure increment parameter and the exposure increment correction parameter. In some specific embodiments, in step S106, the processor calculates the exposure parameter increment value in the basic exposure increment parameter and the correction amplitude in the exposure increment correction parameter. For example, for the exposure time, the increment value is 4 and the correction amplitude is 3, then the corrected exposure time is 4 + 3 = 7, with the unit of ms. For example, for the gain value, the increment value is 25 and the correction amplitude is 0.96, then the corrected gain value is 25 * 0.96 = 24.

[0046] In step S107, the processor determines the transitional exposure parameter according to the correction increment parameter and the basic exposure parameter. In some specific embodiments, in step S107, the processor sums the basic exposure parameter and the correction increment parameter to obtain the transitional exposure parameter. For example, if the exposure parameter includes the exposure time, the basic exposure time is 4 ms, and the correction exposure time is 7 ms, then the transitional exposure time is 4 + 7 = 11 ms. For example, if the exposure parameter includes the gain value, the basic gain value is 200, and the correction gain value is 24, then the transitional gain value is 200 + 24 = 224.

[0047] In some specific embodiments, in step S108, the processor determines the target exposure parameter according to the transitional exposure parameter. In some specific embodiments, when the exposure control termination condition is satisfied, the processor determines the transitional exposure parameter as the target exposure parameter. In some specific embodiments, when the exposure control termination condition is not satisfied or not set, the processor uses the exposure parameter in the transitional exposure parameter as the basic exposure parameter and returns to step S102.

[0048] According to the Figure 1 embodiment shown, the method proposed in this application first determines the basic increment value of the exposure parameter according to the basic image brightness; then judges the ambient light state through the ambient light blur rule and generates the ambient light estimation parameter, that is, the mathematical representation of the ambient light; uses the ambient light estimation parameter and the fuzzy control method to convert the influence of the ambient light into the correction value of the exposure parameter, and finally obtains the corrected exposure parameter. The method proposed in this application adopts the fuzzy control technology, calculates by setting fuzzy rules that conform to human intuition and experience, controls the exposure parameter when the environmental state changes, and can achieve fast convergence without a large amount of calculation. Using the method proposed in this application, automatic exposure can be realized in scenes with complex and changeable environments, and clear images can be captured.

[0049] Figure 2 For Figure 1 the method 1000, it is the step schematic diagram of step S104. As Figure 2 shown, step S104 includes steps S1041 - S1044. Figure 3 For Figure 2 the ambient light blur rule R100 in step S104 in Figure 3 it, as shown, the ambient light blur rule R100 includes rules R11 - R13. Figure 4 For Figure 2 the fuzzy set F100 for ambient light estimation in step S104 in

[0050] As Figure 4As shown, the fuzzy set F100 for ambient light estimation includes a target brightness fuzzy set F11, an exposure time state fuzzy set F12, and a gain value state fuzzy set F13. For example, in Figure 4 the base brightness is i, and the target brightness fuzzy set F11 can be expressed as:

[0051]

[0052]

[0053] In step S1041, the processor determines a first membership data set according to the base exposure parameters and the exposure state fuzzy set. In some specific embodiments, in step S1041, the exposure state fuzzy set represents the functional relationship between the value of the exposure parameter and the membership degree of the exposure parameter state. The exposure parameter state represents the numerical magnitude of the exposure parameter, and the number of exposure parameter states can be set according to the requirements of the actual application. For example, in some specific embodiments, the number of exposure parameter states is 2, for example, the exposure parameter states include large and small.

[0054] In some specific embodiments, the exposure parameters include exposure time and gain value, and the base exposure parameters include base exposure time and base gain value. In some specific embodiments, in step S1041, the gain value state fuzzy set represents the degree of association between the magnitude of the gain value and the gain value state, and the number of functions included in the gain value state fuzzy set and their expressions can be determined according to the number of gain value states in the actual application. For example, in some specific embodiments, the gain value state fuzzy set includes a small gain value membership function and a large gain value membership function.

[0055] In some specific embodiments, in step S1041, the exposure time state fuzzy set represents the degree of association between the magnitude of the exposure time and the exposure time state, and the number of functions included in the exposure time state fuzzy set and their expressions can be determined according to the number of exposure time states in the actual application. For example, in some specific embodiments, the exposure time state fuzzy set includes a small exposure time membership function and a large exposure time membership function.

[0056] In some specific embodiments, similar to Figure 4 the brightness fuzzy set F11 in, as Figure 4 shown, the exposure time state fuzzy set F12 and the gain value state fuzzy set F13 include their corresponding function expressions, which will not be elaborated here.

[0057] For example, in step S1041, as Figure 4As shown, the value of the base exposure time is between TS2 and TS3, the membership degree to the small exposure time is N1, the membership degree to the large exposure time is N2, the base gain value is between IS2 and IS3, the membership degree to the small gain value is N3, and the membership degree to the large gain value is N4. At this time, {{(small gain: N3), (large gain: N4)}, {(small exposure time: N1, (large exposure time: N2)}} is the second membership degree set representing the exposure parameter state.

[0058] In step S1042, the processor determines the second membership degree data set according to the base image brightness and the target brightness fuzzy set. In step S1042, the target brightness fuzzy set represents the functional relationship between the image brightness and the membership degree of the brightness state. The brightness state represents the relative brightness of the ambient light corresponding to the image brightness, and the number of functions included in the target brightness fuzzy set and their expressions can be determined according to the number of brightness states in the actual application. For example, in some specific embodiments, the brightness states include dark and bright. The membership degree of the image brightness to different brightness states represents the degree to which the image brightness belongs to that brightness state, and the number of functions included in the target brightness fuzzy set and their expressions can be determined according to the number of brightness states in the actual application. For example, in some specific embodiments, the base image brightness is 60, the membership degree to the dark brightness is 0.2, and the membership degree to the bright brightness is 0. At this time, {0.2, 0} is the second membership degree data set.

[0059] In step S1043, the processor determines the ambient light estimation parameter according to the ambient light fuzzy rule, the first membership degree data set, and the second membership degree data set.

[0060] In some specific embodiments, in step S1043, the ambient light fuzzy rule represents the corresponding relationship between the image brightness and the exposure parameter state. In some specific embodiments, as Figure 3 shown, in step S1043, the ambient light fuzzy rule R100 includes:

[0061] Rule R11: When the base image brightness belongs to the bright fuzzy set, the base gain value belongs to the small gain value fuzzy set, and the base exposure time belongs to the small exposure time fuzzy set, the ambient light response is bright ambient light;

[0062] Rule R12: When the base image brightness belongs to the dark fuzzy set, the base gain value belongs to the large gain value fuzzy set, and the base exposure time belongs to the large exposure time fuzzy set, the ambient light response is dark ambient light;

[0063] Rule R13: In other cases, the ambient light response is general ambient light. Optionally, the content and number of the ambient light states can be set according to the requirements of the actual application for the ambient light response state.

[0064] For example, in step S1043, the first membership degree data set is: {{(small gain: 0.68), (large gain: 0.1)}, {(small exposure time: 0.72), (large exposure time: 0.11)}}, and the second membership degree data set is: {(dark environment: 0), (general environment: 0.58), (bright environment: 0.21)}. According to the ambient light fuzzy rule R100, the condition for the ambient light response to be bright is a bright environment, small gain, and small exposure time, i.e., {0.21, 0.68, 0.72}.

[0065] In some specific embodiments, in step S1043, the processor evaluates the membership degrees corresponding to the image brightness state and the exposure parameter state according to the ambient light fuzzy rule to obtain the response value of the ambient light state, which is used as the ambient light estimation parameter. Optionally, since the relationship between the above three conditions is an AND relationship, the response value of the ambient light state can be obtained by taking the minimum value of the membership degrees corresponding to the image brightness state and the exposure parameter state respectively. For example, the estimated value for the ambient light response to be bright is min{0.21, 0.68, 0.72} = 0.21. Similarly, the estimated value for the ambient light response to be dark and the estimated value for the ambient light response to be general can be obtained, and the above estimated values are used as the ambient light estimation set. For example, the ambient light estimation set can be {(bright: 0.21), (general: 0.5), (dark: 0)}.

[0066] In some specific embodiments, in step S1043, the processor evaluates the membership degrees corresponding to the image brightness state and the exposure parameter state according to the ambient light fuzzy rule to obtain the response value of the ambient light state, which is used as the ambient light estimation parameter. Optionally, the response value of the ambient light state can be obtained by defuzzifying and then refuzzifying the membership degrees corresponding to the image brightness state and the exposure parameter state respectively.

[0067] For example, as Figure 4 shown, the processor defuzzifies the membership degree data set for the ambient light response to be bright, i.e., {0.21, 0.68, 0.72}, using the centroid method according to the ambient light state fuzzy set F14 to obtain the ambient light brightness value, and then refuzzifies the ambient light brightness value using the ambient light state fuzzy set F14 to obtain the response value of the ambient light state.

[0068] It can be understood that the ambient light fuzzy rule R100 shown as Figure 3 including rules R11 - R13 is only one example of the ambient light fuzzy rule of the present application, and the content and quantity of the ambient light fuzzy rule can be set according to actual application requirements. Similarly, as Figure 4The fuzzy set F100 for ambient light estimation shown is only one example of the fuzzy sets for ambient light estimation in this application. The content and quantity of the fuzzy sets for ambient light estimation can be set according to actual application requirements.

[0069] According to the above embodiments, the method proposed in this application determines the response results of ambient light to image brightness and exposure parameters according to ambient light fuzzy rules or directly obtains the state of ambient light from the fuzzy set of ambient light states as the ambient light estimation parameters. The method according to the above embodiments can quickly and accurately obtain the mathematical representation of ambient light in various environments.

[0070] Figure 5 is Figure 1 a schematic diagram of the exposure correction fuzzy rule R200 in step S105 in Figure 5 As shown, the exposure correction fuzzy rule R200 includes rules R21 - R24. Figure 6 is Figure 1 a schematic diagram of the fuzzy set F200 for exposure correction in step S105 in Figure 6 As shown, the fuzzy set F200 for exposure correction includes an exposure time correction fuzzy set F21 and a gain value correction fuzzy set F22.

[0071] In some specific embodiments, the fuzzy set for exposure correction represents the functional relationship between the correction amplitude of exposure parameters and the membership degree of the correction state of exposure parameters, and the correction state of exposure parameters represents the numerical size of the correction of exposure parameters. In some specific embodiments, the exposure parameters include exposure time and gain value. As Figure 10 shown, the fuzzy set for exposure correction includes an exposure time correction fuzzy set F21 and a gain value correction fuzzy set F22.

[0072] In some specific embodiments, the exposure time correction fuzzy set F21 includes a membership function for decreasing exposure amplitude and a membership function for increasing exposure amplitude, and the gain value correction fuzzy set F22 includes a membership function for decreasing gain amplitude and a membership function for increasing gain amplitude. Similar to Figure 4 the target brightness fuzzy set F11 in Figure 10 as shown, the exposure time correction fuzzy set F21 and the gain value correction fuzzy set F22 include their corresponding functional expressions, which will not be elaborated here.

[0073] The exposure correction fuzzy rule R200 represents the corresponding relationship between the ambient light estimation result and the correction state of exposure parameters. As Figure 5 shown, the exposure correction fuzzy rule R200 includes:

[0074] R21. When the ambient light is bright ambient light, the adjustment range response of the exposure time is that the exposure adjustment range decreases, and the adjustment range response of the gain value is that the gain adjustment range increases;

[0075] R22. When the ambient light is dim ambient light, the adjustment range response of the exposure time is that the exposure adjustment range increases, and the adjustment range response of the gain value is that the gain adjustment range decreases;

[0076] R23. When the adjustment range response of the exposure time exceeds the adjustable range of the exposure, the adjustment range response of the gain value is that the gain adjustment range increases;

[0077] R34. In other cases, the adjustment range responses of the exposure time and the gain value remain unchanged. Optionally, the adjustment range responses of the exposure time and the gain value can be set according to the requirements of the actual application for the content and quantity of the adjustment range responses of the exposure time and the gain value.

[0078] For example, in step S105, the ambient light estimation set is {(bright: 0.21), (dim: 0)}. According to the exposure correction fuzzy rule R200, the processor can obtain that the response of the decrease in the adjustment range of the exposure time and the increase in the adjustment range of the gain value is 0.21, and the response of the increase in the adjustment range of the exposure time and the decrease in the adjustment range of the gain value is 0. At this time, the response values of the exposure time correction and the gain value correction are {{(increase in gain amplitude: 0.21), (decrease in gain amplitude: 0)}, {(increase in exposure time amplitude: 0), (decrease in exposure time amplitude: 0.21)}}. Using the response values of the gain value and the exposure time in each state to Figure 6 truncate the exposure time correction fuzzy set F21 and the gain value correction fuzzy set F22 in, that is, limit the maximum value of the function value range of each in the exposure time correction fuzzy set F21 and the gain value correction fuzzy set F22 to the response value, and the gain value correction response set and the exposure time correction response set can be obtained. In step S105, the gain value correction response set and the exposure time correction response set are used as the exposure increment response correction parameters.

[0079] In step S105, the processor defuzzifies the exposure increment response correction parameters to generate the exposure increment correction parameters. In some specific embodiments, in step S105, the processor uses the centroid method to defuzzify the gain value correction response set and the exposure time correction response set to obtain the gain value correction value and the exposure time correction value, which are the exposure increment correction parameters.

[0080] According to the above embodiments, the method proposed in this application determines the response result of the exposure correction amplitude value to the ambient light state according to the exposure correction fuzzy rule, and uses it as the correction value of the exposure increment value. According to the method of the above embodiments, it is possible to appropriately correct the exposure increment value according to the ambient light state, so as to quickly and accurately obtain the optimized exposure parameter result.

[0081] It can be understood that the exposure correction fuzzy rule R200 shown in Figure 5 only includes Rule R21 - Rule R24, which is only one example of the exposure correction fuzzy rule of this application. The content and quantity of the exposure correction fuzzy rule can be set according to actual application requirements. Similarly, the fuzzy set F200 for exposure correction shown in Figure 6 is only one example of the fuzzy set for exposure correction of this application. The content and quantity of the fuzzy set for exposure correction can be set according to actual application requirements.

[0082] Figure 7 For Figure 1 is a schematic diagram of the steps of step S108 in method 1000. As shown in Figure 7 , step S108 includes steps S1081 - step S1082.

[0083] In step S1081, when the exposure control termination condition is met, the processor uses the transitional exposure parameter as the target exposure parameter. In step S1082, when the exposure control termination condition is not met or not set, the processor uses the transitional exposure parameter as the basic exposure parameter, returns to collect the first image with the basic exposure parameter, and determines the basic image brightness according to the first image. In some specific embodiments, in step S108, the exposure control termination condition includes that the image brightness of the second image collected with the transitional exposure parameter is within the target brightness range.

[0084] In some specific embodiments, the exposure control termination condition includes that the difference between the current one or more transitional exposure parameters and the previous one or more corresponding transitional exposure parameters is less than one or more thresholds in the preset first threshold set. In some specific embodiments, the exposure parameters include exposure time and gain value, and the first threshold set includes an exposure time threshold and a gain value threshold.

[0085] In some specific embodiments, when the difference between the current exposure time and the previously generated exposure time is less than the exposure time threshold, and the difference between the current gain value and the previously generated gain value is less than the gain value threshold, the processor uses the current exposure time and gain value as the target exposure parameters.

[0086] In some specific embodiments, in step S1082, when no exposure control termination condition is set, the processor uses the overexposure parameter as the basic exposure parameter and returns to step S102 in Figure 1 above.

[0087] According to the embodiment as Figure 7 shown, the method proposed in this application can achieve automatic camera exposure under various requirements such as continuous image acquisition and single-frame image acquisition by flexibly adjusting the exposure control termination condition. The method proposed in this application can be applied in various situations and can also be adjusted according to requirements, with adaptability and practicality.

[0088] Figure 8 FIG. is a schematic structural diagram of an automatic camera exposure device 2000 based on fuzzy control according to this application. As Figure 8 shown, the device 2000 includes a preprocessing module 201, a basic automatic exposure module 202, an ambient light estimation module 203, an exposure correction module 204, and a parameter determination module 205.

[0089] In some specific embodiments, the preprocessing module 201 sets at least one set of exposure parameters as the basic exposure parameters. In some specific embodiments, the basic exposure parameters are preset default values. In some specific embodiments, the exposure parameters include camera parameters that affect the brightness of the captured image, such as exposure time and gain value.

[0090] In some specific embodiments, the preprocessing module 201 captures a first image with the basic exposure parameters and determines the basic image brightness according to the first image. In some specific embodiments, the preprocessing module 201 determining the basic image brightness according to the first image includes selecting an image region of interest in the first image and calculating the average gray level of the image region. In some specific embodiments, the preprocessing module 201 selects the central region in the first image as the image region of interest. In some specific embodiments, the preprocessing module 201 selects the target detection object region in the first image as the image region of interest, such as a face frame region, a palm region, etc.

[0091] In some specific embodiments, the basic automatic exposure module 202 determines the basic exposure increment parameter according to the basic image brightness. In some specific embodiments, the basic automatic exposure module 202 uses an existing automatic exposure method to determine the basic exposure increment according to the basic image brightness. For example, the basic automatic exposure module 202 adjusts the exposure parameters using the exposure priority or gain priority method to determine the basic exposure increment parameter.

[0092] In some specific embodiments, the ambient light estimation module 203 determines a basic exposure increment parameter based on a fuzzy control method, specifically including generating a basic exposure increment parameter according to a basic image brightness, an exposure increment fuzzy rule, and an exposure increment fuzzy set. In some specific embodiments, the exposure increment fuzzy set includes a brightness fuzzy set and an exposure control fuzzy set.

[0093] In some specific embodiments, the brightness fuzzy set represents a functional relationship of the membership degrees of an image brightness and an image brightness state. The membership degree of the image brightness to different image brightness states represents the degree to which the image brightness belongs to the image brightness state. The exposure control fuzzy set represents a functional relationship of the increment value of an exposure parameter and the membership degree of the exposure parameter change direction. The exposure parameter change direction represents the change of the exposure parameter. For example, the exposure parameter change direction includes negative growth, maintaining, and positive growth.

[0094] In some specific embodiments, the membership degree of the increment value of the exposure parameter to different exposure parameter change directions represents the degree to which the increment value of the exposure parameter belongs to the exposure parameter change direction. The exposure increment fuzzy rule represents the corresponding relationship between the exposure parameter change direction and the image brightness state. In some specific embodiments, the ambient light estimation module 203 can obtain an exposure parameter response set of the exposure control fuzzy set to the first membership degree according to the exposure increment fuzzy rule. In some specific embodiments, defuzzification is performed on the exposure parameter response set to obtain the basic exposure increment parameter.

[0095] The ambient light estimation module 203 generates an ambient light estimation parameter according to the basic image brightness, the basic exposure parameter, an ambient light fuzzy rule, and a fuzzy set for ambient light estimation. In some specific embodiments, the fuzzy set for ambient light estimation includes a target brightness fuzzy set and an exposure state fuzzy set.

[0096] In some specific embodiments, the ambient light estimation module 203 determines a first membership degree data set according to the basic exposure parameter and the exposure state fuzzy set. The exposure state fuzzy set represents a functional relationship between the value of the exposure parameter and the membership degree of the exposure parameter state. The exposure parameter state represents the numerical size of the exposure parameter, and the number of exposure parameter states can be set according to the requirements of actual applications. For example, in some specific embodiments, the number of exposure parameter states is 3, including a small exposure parameter, a general exposure parameter, and a large exposure parameter. For example, in some specific embodiments, the number of exposure parameter states is 2, including a large exposure parameter and a small exposure parameter.

[0097] In some specific embodiments, the exposure parameters include the exposure time. For example, the base exposure time is 4 ms, the membership degree to a small exposure time is 0.56, and the membership degree to a large exposure time is 0.11. In some specific embodiments, the exposure parameters include the gain value. For example, the base gain value is 200, the membership degree to a small gain value is 0.5, and the membership degree to a large gain value is 0.5. At this time, {{0.56, 0.11}, {0.5, 0.5}} is the first membership degree data set representing the exposure parameter state.

[0098] In some specific embodiments, the ambient light estimation module 203 determines a second membership degree data set according to the base image brightness and the target brightness fuzzy set. The target brightness fuzzy set represents the functional relationship between the image brightness and the membership degree of the brightness state. The brightness state represents the relative brightness degree of the ambient light corresponding to the image brightness, and the number of brightness states can be set according to the requirements of the actual application. For example, in some specific embodiments, the number of brightness states is 3, including a dark brightness, a general brightness, and a bright brightness. For example, in some specific embodiments, the number of brightness states is 2, including a dark brightness and a bright brightness.

[0099] The membership degree of the image brightness to different brightness states represents the degree to which the image brightness belongs to this brightness state. For example, in some specific embodiments, the brightness states include a dark brightness and a bright brightness. When the base image brightness is 60, the membership degree to the dark brightness is 0.2, and the membership degree to the bright brightness is 0. At this time, {0.2, 0} is the second membership degree data set.

[0100] In some specific embodiments, the ambient light fuzzy rule represents the corresponding relationship between the ambient light estimation state and the exposure parameter state. For example, the ambient light fuzzy rule includes: when the image brightness is in the bright state and the exposure parameter is in the small state, the ambient light estimation state is in the bright state.

[0101] In some specific embodiments, the ambient light estimation module 203 evaluates the membership degrees corresponding to the image brightness state and the exposure parameter state according to the ambient light fuzzy rule, and obtains the response value of the ambient light state as the ambient light estimation parameter. For example, the method by which the ambient light estimation module 203 evaluates the membership degrees corresponding to the image brightness state and the exposure parameter state to obtain the response value of the ambient light state can be adjusted according to the actual application, and different methods can be used for evaluation.

[0102] Optionally, the ambient light estimation module 203 may obtain the response value of the ambient light state by taking the minimum value of the membership degrees corresponding to the image brightness state and the exposure parameter state respectively. For example, the membership degree of the image brightness being in the bright state is 0.11, the membership degree of the exposure parameter state being in the small state is 0.5, and the response value of the ambient light state being bright is min{0.11, 0.5}, that is, 0.11. Similarly, the ambient light estimation parameters including the ambient light state and the corresponding membership degrees can be obtained. For example, the ambient light estimation parameters may be {(dark ambient light: 0.11), (bright ambient light: 0)}.

[0103] Optionally, the ambient light estimation module 203 may obtain the response value of the ambient light state by deblurring and reblurring the membership degrees corresponding to the image brightness state and the exposure parameter state respectively.

[0104] In some specific embodiments, the exposure correction module 204 generates an exposure increment correction parameter according to the ambient light estimation parameters, the exposure correction fuzzy rules, and the fuzzy set for exposure correction. The exposure correction module 204 generates an exposure increment response correction parameter according to the ambient light estimation parameters, the exposure correction fuzzy rules, and the fuzzy set for exposure correction. The exposure correction module 204 deblurs the exposure increment response correction parameter to generate an exposure increment correction parameter. In some specific embodiments, the exposure correction module 204 uses the centroid method to deblur the gain value correction response parameter and the exposure time correction response parameter to obtain the gain value correction value and the exposure time correction value, which are the exposure increment correction parameters.

[0105] The exposure correction module 204 generates a correction increment parameter according to the basic exposure increment parameter and the exposure increment correction parameter. In some specific embodiments, the exposure correction module 204 calculates the exposure parameter increment value in the basic exposure increment parameter and the correction amplitude in the exposure increment correction parameter. For example, for the exposure time, the increment value is 4 and the correction amplitude is 3, then the corrected exposure time is 4 + 3 = 7, with the unit of ms. For example, for the gain value, the increment value is 25 and the correction amplitude is 0.96, then the corrected gain value is 25 * 0.96 = 24.

[0106] The exposure correction module 204 determines the transitional exposure parameter according to the correction increment parameter and the basic exposure parameter. In some specific embodiments, the exposure correction module 204 sums the basic exposure parameter and the correction increment parameter to obtain the transitional exposure parameter. For example, the exposure parameter includes the exposure time, the basic exposure time is 4 ms, and the corrected exposure time is 7 ms, then the transitional exposure time is 4 + 7 = 11 ms. For example, the exposure parameter includes the gain value, the basic gain value is 200, and the corrected gain value is 24, then the transitional gain value is 200 + 24 = 224.

[0107] In some specific embodiments, the parameter determination module 205 determines the target exposure parameter according to the transitional exposure parameter. In some specific embodiments, when the exposure control termination condition is met, the parameter determination module 205 determines the transitional exposure parameter as the target exposure parameter. In some specific embodiments, when the exposure control termination condition is not met or not set, the parameter determination module 205 uses the exposure parameter in the transitional exposure parameter as the base exposure parameter, returns to collect the first image with the base exposure parameter, and determines the base image brightness according to the first image.

[0108] Figure 9 FIG. is a schematic structural diagram of the image recognition device 3000 of the present application. As Figure 9 shown, the device 3000 includes an automatic exposure device 301.

[0109] In some specific embodiments, the device 3000 is a door lock, and its working process is as follows:

[0110] Set the default exposure time and default gain value as the base exposure parameter, use the automatic exposure device 301 to collect an image, select the face area as the image area of interest, and count the average gray value of the face area as the base image brightness; the automatic exposure device 301 determines the base exposure increment parameter according to the base image brightness; the automatic exposure device 301 generates the ambient light estimation parameter according to the base image brightness, the base exposure parameter, the ambient light blur rule, and the fuzzy set for ambient light estimation; the automatic exposure device 301 generates the exposure increment correction parameter according to the ambient light estimation parameter, the exposure correction blur rule, and the fuzzy set for exposure correction; the automatic exposure device 301 generates the correction increment parameter according to the base exposure increment parameter and the exposure increment correction parameter; the automatic exposure device 301 determines the transitional exposure parameter according to the correction increment parameter and the base exposure parameter; the automatic exposure device 301 determines the target exposure parameter according to the transitional exposure parameter set. After the door lock collects the target image with the target exposure parameter, it starts the face recognition process according to the target image. When the face recognition result is passed, it controls to unlock, otherwise it does not act.

[0111] Figure 10 FIG. is a structural diagram of an electronic device provided by the present application, including a processor and a memory. The memory stores computer instructions, and when the computer instructions are executed by the processor, the processor executes the computer instructions to implement the method and refinement scheme as Figure 1 shown.

[0112] It should be understood that the above device embodiments are merely illustrative, and the devices disclosed in the present invention can also be implemented in other ways. For example, the division of the units / modules described in the above embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0113] In addition, without special instructions, in each embodiment of the present invention, each functional unit / module can be integrated in one unit / module, or each unit / module can exist physically alone, or two or more units / modules can be integrated together. The above integrated unit / module can be implemented in the form of hardware or in the form of a software program module.

[0114] When the integrated unit / module is implemented in the form of hardware, the hardware can be a digital circuit, an analog circuit, etc. The physical implementation of the hardware structure includes but is not limited to transistors, memristors, etc. Without special instructions, the processor or chip can be any suitable hardware processor, such as CPU, GPU, FPGA, DSP, and ASIC, etc. Without special instructions, the on-chip cache, off-chip memory, and memory can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc.

[0115] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this disclosure. The aforementioned memory includes: various media such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), external hard drives, magnetic disks, or optical discs that can store program codes.

[0116] Embodiments of this application also provide a non-transitory computer storage medium storing a computer program, which, when executed by multiple processors, causes the processors to execute the method and refinement solutions as Figure 1 shown.

[0117] The above has introduced the embodiments of this application in detail. Specific examples are used in this article to elaborate on the principle and implementation manner of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application. At the same time, any changes or deformations made by those skilled in the art based on the idea of this application, within the specific implementation manner and application scope of this application, fall within the protection scope of this application. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. An automatic exposure method for a camera based on fuzzy control, characterized in that, Including: Setting at least one set of exposure parameters as the basic exposure parameters; Collecting a first image with the basic exposure parameters and determining the basic image brightness according to the first image; Determining the basic exposure increment parameter according to the basic image brightness; Generating an ambient light estimation parameter according to the basic image brightness, the basic exposure parameters, the ambient light blur rule, and the fuzzy set for ambient light estimation; Generating an exposure increment correction parameter according to the ambient light estimation parameter, the exposure correction blur rule, and the fuzzy set for exposure correction; Generating a correction increment parameter according to the basic exposure increment parameter and the exposure increment correction parameter; Determining a transition exposure parameter according to the correction increment parameter and the basic exposure parameters; Determining a target exposure parameter according to the transition exposure parameter.

2. The method according to claim 1, wherein Determining a target exposure parameter according to the transition exposure parameter includes: When the exposure control termination condition is satisfied, using the transition exposure parameter as the target exposure parameter; or When the exposure control termination condition is not set or not satisfied, using the transition exposure parameter as the basic exposure parameter, returning to collect a first image with the basic exposure parameter, and determining the basic image brightness according to the first image.

3. The method according to claim 2, wherein The number of the transition exposure parameters is at least one set, and the exposure control termination condition includes: The difference between the current at least one set of the transition exposure parameters and the corresponding at least one set of the transition exposure parameters generated last time is less than at least one preset threshold; or The image brightness of the second image collected with the transition exposure parameter is within the target brightness range.

4. The method according to any one of claims 1 to 3, characterized in that, The fuzzy set for ambient light estimation includes a target brightness fuzzy set and an exposure state fuzzy set. Generating an ambient light estimation parameter according to the basic image brightness, the basic exposure parameters, the ambient light blur rule, and the fuzzy set for ambient light estimation includes: Determining a first membership degree data set according to the basic exposure parameters and the exposure state fuzzy set; Determining a second membership degree data set according to the basic image brightness and the target brightness fuzzy set; Determining an ambient light estimation parameter according to the ambient light blur rule, the first membership degree data set, and the second membership degree data set.

5. The method according to claim 4, wherein The exposure parameters include an exposure time and a gain value, and the basic exposure parameters include a basic exposure time and a basic gain value. ， The exposure state fuzzy set at least includes a gain value state fuzzy set and an exposure time state fuzzy set. The gain value state fuzzy set at least includes a small gain value fuzzy set and a large gain value fuzzy set. The exposure time state fuzzy set at least includes a small exposure time fuzzy set and a large exposure time fuzzy set. The target brightness fuzzy set at least includes a dark fuzzy set and a bright fuzzy set.

6. The method according to claim 5, characterized in that, The ambient light blur rule includes: When the basic image brightness belongs to the bright fuzzy set, the basic gain value belongs to the small gain value fuzzy set, and the basic exposure time belongs to the small exposure time fuzzy set, the ambient light response is bright ambient light; When the basic image brightness belongs to the dark fuzzy set, the basic gain value belongs to the large gain value fuzzy set, and the basic exposure time belongs to the large exposure time fuzzy set, the ambient light response is dark ambient light; otherwise The ambient light response is general ambient light.

7. The method according to any one of claims 1 to 3, characterized in that The exposure parameters include an exposure time and a gain value, the basic exposure parameters include a basic exposure time and a basic gain value, the fuzzy sets for exposure correction at least include an exposure time correction fuzzy set and a gain value correction fuzzy set, the exposure time correction fuzzy set at least includes an exposure amplitude reduction fuzzy set and an exposure amplitude increase fuzzy set, and the gain value correction fuzzy set at least includes a gain amplitude reduction fuzzy set and a gain amplitude increase fuzzy set.

8. The method according to claim 7, wherein The exposure correction fuzzy rules include: In the case where the ambient light is bright ambient light, the exposure time adjustment amplitude response is an exposure adjustment amplitude reduction, and the gain value adjustment amplitude response is a gain adjustment amplitude increase; In the case where the ambient light is dark ambient light, the exposure time adjustment amplitude response is an exposure adjustment amplitude increase, and the gain value adjustment amplitude response is a gain adjustment amplitude reduction; In the case where the exposure time adjustment amplitude response exceeds the exposure adjustable range, the gain value adjustment amplitude response is a gain adjustment amplitude increase; otherwise The exposure time adjustment amplitude and the gain value adjustment amplitude response remain unchanged.

9. An automatic exposure device for a camera based on fuzzy control, characterized in that, Including: A preprocessing module for setting at least one set of exposure parameters as basic exposure parameters; Collecting a first image with the basic exposure parameters and determining the basic image brightness according to the first image; A basic automatic exposure module for determining basic exposure increment parameters according to the basic image brightness; An ambient light estimation module for generating ambient light estimation parameters according to the basic image brightness, the basic exposure parameters, the ambient light fuzzy rules, and the fuzzy sets for ambient light estimation; An exposure correction module for generating exposure increment correction parameters according to the ambient light estimation parameters, the exposure correction fuzzy rules, and the fuzzy sets for exposure correction; Generating correction increment parameters according to the basic exposure increment parameters and the exposure increment correction parameters; Determining transition exposure parameters according to the correction increment parameters and the basic exposure parameters; A parameter determination module for determining target exposure parameters according to the transition exposure parameters.

10. An image recognition device, characterized in that, Including the camera automatic exposure device based on fuzzy control according to claim 9.

11. An electronic device, including: A processor; A memory storing a computer program, which when executed by the processor causes the processor to execute the method according to any one of claims 1-8.

12. A non-transitory computer-readable storage medium, having stored thereon computer-readable instructions, which when executed by a processor cause the processor to execute the method according to any one of claims 1-8.

Images