Visible light brightness detection method and system coping with camera

Abstract

The invention relates to the technical field of photometry of an image processing camera, in particular to a visible light brightness detection method and system for a camera, and the method comprises the following steps: obtaining first brightness information of a scene in a view field of the camera when a linear polarizer is in a first vibration transmission direction in a target scene shot by the camera through a transparent medium, when the linear polarizer is in a direction different from the first vibration transmission direction, second brightness information of a scene in the field of view of the camera is obtained; based on the first brightness information and the second brightness information, the characteristic that the brightness changes along with the change of the vibration transmission direction of the linear polarizer is analyzed, so that a transparent medium surface reflected light part and a target scene transmission light part are distinguished, and the contribution amount of the transparent medium surface reflected light to the total brightness is determined; the brightness of the target scene is accurately determined by analyzing the characteristic that the brightness changes along with the vibration transmission direction of the polarizer.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a method and system for detecting the visible light brightness of a camera. Background Art

[0002] In many actual application scenarios, camera devices often need to collect images through a transparent medium. However, when a camera takes pictures through a transparent medium, the light received by the image sensor is not only from the target scene on the other side of the transparent medium (i.e., the effective transmitted light), but also includes interference light from the camera's own environment and reflected back through the surface of the transparent medium. When there are significant differences in the lighting conditions between the camera's environment and the target scene, such as bright indoors and dim outdoors, or sufficient lighting in the exhibition hall and local lighting inside the display cabinet, the reflection phenomenon on the surface of the transparent medium will seriously interfere with the brightness detection of the camera. The images formed by the reflection of light sources or high-brightness objects in the camera's environment on the surface of the transparent medium will be superimposed on the images of the target scene, enter the camera together and be captured.

[0003] In existing camera metering systems, regardless of the metering mode used (such as global average, center-weighted, zone metering, etc.), their basic principle is based on the analysis of the intensity of light entering the lens. Therefore, when calculating the brightness value of the metering area, these systems cannot intelligently distinguish the effective transmitted light from the target scene from the interference light reflected back through the surface of the transparent medium (including irregular points on it), and the obtained brightness value is the total intensity information after mixing these two kinds of light.

[0004] When the ambient brightness on one side of the camera is much higher than the brightness of the target scene, or the reflectivity of the transparent medium is relatively high, or there are strong reflection defects on the surface, the contribution of the reflected light to the metering reading will become significant and may even dominate the metering result. Especially when the metering area covers these strong reflection areas, the brightness value of this area will be greatly increased.

[0005] If the automatic exposure algorithm of the camera fails to effectively identify and compensate for this additional brightness introduced by the reflection on the surface of the transparent medium, it will make an exposure decision based on the brightness detection result that is interfered by the reflected light and is too high. If the main shooting intention is to clearly obtain the target scene on the other side of the transparent medium, the automatic exposure control unit will wrongly determine that the picture is too bright, and then reduce the exposure amount (such as shortening the shutter time, reducing the aperture, reducing the sensitivity). As a result, although the reflected bright spots may no longer be overexposed, the target scene that really needs to be recorded becomes dim due to insufficient exposure, and a large amount of detail, color, and morphological information is lost, resulting in a serious decline in the final image quality and unable to meet the image acquisition requirements of specific application scenarios. The root cause of this problem is that the existing visible light brightness detection method fails to effectively separate the transmitted light of the target scene from the reflected light at the interface of the transparent medium for independent evaluation.

[0006] In view of the above problems, the existing technology urgently needs to be improved. Summary of the Invention

[0007] The purpose of the present invention is to solve the deficiencies existing in the prior art, and a visible light brightness detection method and system for a camera are proposed.

[0008] In a first aspect, the present invention provides a visible light brightness detection method for a camera, and the method includes the following steps: S1: Obtain the first brightness information of the scene within the camera's field of view when the linear polarizer is in the first polarization direction when the camera shoots a target scene through a transparent medium, and the second brightness information of the scene within the camera's field of view when the linear polarizer is in a direction different from the first polarization direction; S2: Based on the first brightness information and the second brightness information, analyze the characteristics of the change in brightness with the change in the polarization direction of the linear polarizer to distinguish the part of the light reflected from the surface of the transparent medium and the part of the transmitted light of the target scene, so as to determine the contribution amount of the light reflected from the surface of the transparent medium to the total brightness; S3: Based on the contribution amount of the light reflected from the surface of the transparent medium to the total brightness, correct at least one or a combination of the first brightness information and the second brightness information to obtain the target brightness representing the target scene on the other side of the transparent medium.

[0009] In step S1 of this application, the brightness information of the scene captured by the camera in at least two different polarization directions of the linear polarizer is obtained. The reflected light on the surface of the transparent medium usually has polarization characteristics, while the light of the target scene may have different polarization states. Obtaining the brightness data in different polarization directions provides a basis for subsequent analysis of the polarization characteristics of the scene light. In step S2, based on the brightness information in different polarization directions obtained in S1, the change characteristics of the brightness with the polarization direction are analyzed. This change characteristic reflects the distribution of polarized light in the scene. By analyzing the polarization characteristics, the method can identify and distinguish the part of the light caused by the reflection from the surface of the transparent medium and the part of the transmitted light from the target scene. Furthermore, the method calculates the contribution amount of the reflected light to the total measured brightness. This processing process realizes the identification and quantification of the interfering light (reflected light). In step S3, the initial brightness information obtained in S1 is corrected by using the contribution amount of the reflected light determined in S2. By removing the contribution of the reflected light, the corrected brightness information represents the brightness of the target scene on the other side of the transparent medium. This processing process provides the brightness of the target scene after removing the reflection influence, and solves the problem that the original brightness measurement contains reflection interference.

[0010] In a second aspect, a visible light brightness detection system for a camera is provided, and the system includes: Polarization information acquisition module: Obtain the first brightness information of the scene within the camera's field of view when the linear polarizer is in the first polarization direction for the camera to capture the target scene through a transparent medium, and the second brightness information of the scene within the camera's field of view when the linear polarizer is in a direction different from the first polarization direction. Characteristic analysis and identification module: Based on the first brightness information and the second brightness information, analyze the characteristics of the brightness change as the polarization direction of the linear polarizer changes, so as to distinguish the part of the reflected light from the surface of the transparent medium and the part of the transmitted light of the target scene, and determine the contribution of the reflected light from the surface of the transparent medium to the total brightness. Brightness correction module: Based on the contribution of the reflected light from the surface of the transparent medium to the total brightness, correct at least one or a combination of the first brightness information and the second brightness information to obtain the target brightness characterizing the target scene on the other side of the transparent medium.

[0011] Compared with the prior art, the present invention has the following beneficial effects: By obtaining the brightness information in different polarization directions, analyzing the characteristics of brightness change, distinguishing the reflected light and the transmitted light, and performing correction, it is possible to distinguish the reflected light from the surface of the transparent medium and the transmitted light of the target scene, determine the contribution of the reflected light from the surface of the transparent medium to the total brightness, and obtain the brightness of the target scene. Description of the drawings

[0012] Figure 1 It is a flowchart of the present invention.

[0013] Figure 2 It is a structural diagram of the present invention.

[0014] In the figure: 201, polarization information acquisition module; 202, characteristic analysis and identification module; 203, brightness correction module. Detailed implementation manners

[0015] The following details the implementation manners of the present invention. The examples of the implementation manners are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The implementation manners described below by referring to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation to the present invention.

[0016] The terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.

[0017] Application scenario: How to solve the technical problem that the existing visible light brightness detection method cannot effectively distinguish and suppress the transmitted light of the target scene and the interfering light reflected from the surface of the transparent medium, including specular reflection and irregular strong reflection, resulting in the overall elevation of the photometric result by the interfering light, and then the automatic exposure system wrongly reduces the exposure amount, ultimately causing serious underexposure of the target subject on the other side of the transparent medium, a large amount of loss of image details and color information, and unable to meet the image acquisition quality requirements in specific scenarios.

[0018] Such as Figure 1 shown, a method for detecting the visible light brightness of a camera, the method includes the following steps: S1: Obtain the first brightness information of the scene in the camera's field of view when the linear polarizer in the camera is in the first polarization direction through the transparent medium to photograph the target scene, and the second brightness information of the scene in the camera's field of view when the linear polarizer is in a direction different from the first polarization direction; S2: Based on the first brightness information and the second brightness information, analyze the characteristics of the change in brightness as the polarization direction of the linear polarizer changes, so as to distinguish the part of the reflected light from the surface of the transparent medium and the part of the transmitted light of the target scene, and determine the contribution amount of the reflected light from the surface of the transparent medium to the total brightness; S3: Based on the contribution amount of the reflected light from the surface of the transparent medium to the total brightness, correct at least one or a combination of the first brightness information and the second brightness information to obtain the target brightness representing the target scene on the other side of the transparent medium.

[0019] Among them, in step S1, the brightness information of the scene captured by the camera in at least two different polarization directions of the linear polarizer is obtained. The reflected light on the surface of the transparent medium usually has polarization characteristics, while the light of the target scene may have different polarization states. Obtaining the brightness data in different polarization directions provides a basis for subsequent analysis of the polarization characteristics of the scene light.

[0020] In step S2, based on the brightness information in different polarization directions obtained in S1, the change characteristics of the brightness with the polarization direction are analyzed. This change characteristic reflects the distribution of polarized light in the scene. By analyzing the polarization characteristics, the method can identify and distinguish the part of the light caused by the reflection from the surface of the transparent medium and the part of the transmitted light from the target scene. Furthermore, the method calculates the contribution amount of the reflected light to the total measured brightness. This processing process realizes the identification and quantification of the interfering light (reflected light).

[0021] In step S3, the initial brightness information obtained in S1 is corrected by using the contribution amount of the reflected light determined in S2. By removing the contribution of the reflected light, the corrected brightness information represents the brightness of the target scene on the other side of the transparent medium. This processing process provides the brightness of the target scene after removing the reflection influence and solves the problem that the original brightness measurement contains reflection interference.

[0022] Specifically, when the camera captures the target scene through a transparent medium, reflected light is generated on the surface of the transparent medium. This part of the reflected light is superimposed on the transmitted light from the target scene, jointly constituting the total light signal received by the camera. Existing brightness detection methods cannot distinguish these two types of light, resulting in a relatively high measured total brightness and affecting subsequent exposure control. This method obtains the brightness information of the scene in different polarization states by introducing a linear polarizer and changing its polarization direction. Since the reflected light on the surface of the transparent medium usually has obvious polarization characteristics, while the polarization characteristics of the transmitted light of the target scene may be different, by comparing the brightness differences in at least two different polarization directions, the polarization characteristics of different regions in the scene can be analyzed. Based on this polarization characteristic analysis, the method can identify the regions mainly contributed by the reflected light on the surface of the transparent medium and the regions mainly contributed by the transmitted light of the target scene. Further, the method quantifies the specific contribution of the reflected light on the surface of the transparent medium to the total brightness. Finally, using the calculated contribution of the reflected light, the originally measured brightness information is corrected, for example, by subtracting the contribution of the reflected light, so as to obtain a brightness value that more accurately represents the target scene on the other side of the transparent medium. Thus, this method effectively separates and suppresses the interference of the reflected light on the surface of the transparent medium, obtains information closer to the true brightness of the target scene, and provides a basis for subsequent accurate automatic exposure control.

[0023] Further, step S2 includes: S21: Obtain each piece of brightness information collected in different polarization directions. Each piece of brightness information includes the first brightness information, the second brightness information, and optionally one or more pieces of optional brightness information collected in other different polarization directions; S22: Calculate the polarization characteristic parameters of the region where the brightness changes with the change of the polarization direction of the linear polarizer within the camera's field of view; S23: Compare the polarization characteristic parameters calculated for the brightness change region with the typical polarization characteristic parameter range in the preset reference model to generate a comparison result; S24: According to the comparison result, identify the source of the brightness change, and distinguish the part dominated by the reflected light on the surface of the transparent medium and the part dominated by the light source or object with polarization characteristics in the target scene itself; S25: Perform exclusion or compensation processing on the part of the brightness change dominated by the light source or object with polarization characteristics in the target scene itself to determine the contribution of the reflected light on the surface of the transparent medium to the total brightness.

[0024] Specifically, this method acquires the scene brightness information under at least two, preferably multiple different polarization directions, collecting more comprehensive polarization data. Further, using this multi-directional brightness information, the polarization characteristic parameters of the region where the brightness in the camera's field of view changes with the polarization direction of the polarizer are calculated, such as the degree of polarization and the direction of the principal polarization axis. These parameters quantitatively describe the polarization state of light. The reflected light from the surface of a transparent medium and the light emitted by a light source or object with polarization characteristics in the target scene usually have different ranges of polarization characteristic parameters. The calculated polarization characteristic parameters are compared with a preset reference model, which stores the ranges of polarization characteristic parameters of known sources (such as reflection from the glass surface, typical natural polarization sources). According to the comparison results, the source of the brightness change is identified, distinguishing which regions of the brightness change are mainly caused by the reflection from the surface of the transparent medium and which regions are mainly caused by the light source or object with polarization characteristics in the target scene itself. Finally, for the identified part of the brightness change caused by the polarization source in the target scene itself, exclusion or compensation processing is performed. This means that when calculating the contribution of the reflected light from the surface of the transparent medium to the total brightness, the influence of the polarized light from the target scene itself is subtracted or adjusted. Thus, the obtained contribution of the reflected light from the surface of the transparent medium to the total brightness more accurately reflects the influence of the reflected light, excluding the interference of the polarized light from the target scene itself, laying a foundation for accurately correcting the brightness and obtaining accurate brightness information characterizing the target scene in the subsequent process.

[0025] Further, step S22 includes: S221: The first brightness information and the second brightness information in step S21, and optionally one or more brightness information collected under other different polarization directions are collectively referred to as multi-frame polarization brightness information; S222: Determine the spatial offset of the corresponding scene regions in each frame of brightness information caused by the relative displacement or attitude change between the camera and the target scene during the acquisition of the multi-frame polarization brightness information; S223: According to the determined spatial offset, perform image registration processing on the multi-frame polarization brightness information to generate a set of spatially aligned multi-frame polarization brightness information, ensuring that the brightness values used to calculate the polarization characteristics of the same scene point in the spatially aligned multi-frame polarization brightness information correspond to the same physical region of the target scene; S224: Based on the spatially aligned multi-frame polarization brightness information, calculate the polarization characteristic parameters of the region where the brightness changes with the polarization direction of the linear polarizer in the camera's field of view.

[0026] Specifically, the present method aims to solve the problem of position offset of the same physical region in different frame images caused by the relative motion that may occur between the camera and the target scene during the acquisition of multiple frames of polarization brightness information. First, step S221 collects multiple frames of brightness information obtained under different transmission directions of the linear polarizer. Then, step S222 determines the spatial offset of each frame relative to a certain reference frame by analyzing the content of these brightness information frames. This offset reflects the relative motion of the camera or the scene during the acquisition. Next, step S223 performs image registration processing on all the brightness information frames according to the determined spatial offset. Image registration corrects the spatial misalignment between images by applying appropriate geometric transformations (such as translation, rotation, scaling), generating a set of images that are spatially aligned with each other. This processing ensures that in subsequent calculations, the brightness values extracted from different frames correspond to the same physical position of the target scene. Finally, step S224 calculates the polarization characteristic parameters of the region where the brightness changes with the transmission direction of the polarizer within the camera's field of view based on these spatially aligned multiple frames of brightness information. By eliminating the data misalignment caused by motion, the calculated polarization characteristic parameters are more accurate, providing a reliable data basis for subsequent differentiation between reflected light and transmitted light.

[0027] Furthermore, the polarization characteristic parameters include one or a combination of two items, namely the degree of polarization and the direction of the polarization principal axis, calculated based on the corresponding brightness information.

[0028] Furthermore, step S222 includes: S2221: For each frame image in the multiple frames of polarization brightness information, apply an image structure extraction algorithm to generate a structure feature map corresponding to this frame image; S2222: Select one frame from the multiple frames of polarization brightness information as the reference frame, and use the structure feature map corresponding to the reference frame image as the reference structure feature map; S2223: Between the reference structure feature map and the structure feature maps corresponding to the other frame images except the reference frame, determine the initial region correspondence by performing region-based cross-correlation operation or phase correlation operation; S2224: Perform consistency verification on the determined initial region correspondence, and determine the spatial offset of each frame relative to the reference frame based on the region correspondence that meets the consistency condition.

[0029] Among them, an image structure extraction algorithm is applied to each frame of the multi-frame polarization brightness information to generate a structure feature map. The structure feature map contains the edge information or gradient direction information of the image. Thus, the influence of the scene illumination change independent of the polarizer rotation on the pixel intensity in the multi-frame polarization brightness information is weakened. One frame in the multi-frame polarization brightness information is selected as a reference frame, and the structure feature map corresponding to the reference frame image is used as a reference structure feature map. Between the reference structure feature map and the structure feature maps corresponding to other frames except the reference frame, an initial region correspondence is determined by performing region-based cross-correlation operation or phase correlation operation. Consistency verification is performed on the determined initial region correspondence. Based on the region correspondences that meet the consistency conditions, the spatial offset of each frame relative to the reference frame is determined.

[0030] Specifically, when the camera captures the target scene through a transparent medium and acquires multi-frame polarization brightness information, due to the possible relative displacement or attitude change of the camera or the target scene, the position of the same physical region in different frame images is offset on the image sensor. Directly aligning images based on the original brightness information is easily affected by the brightness change caused by the rotation of the polarizer and the illumination change of the scene itself, resulting in inaccurate matching. To solve this problem, in this solution, an image structure extraction algorithm, such as edge detection or gradient calculation, is applied to each frame of the original brightness image to generate a structure feature map reflecting the geometric structure of the image. These structure feature maps are insensitive to the overall brightness change. Then, one of the frames is selected as the reference frame, and its structure feature map is used as the benchmark. Next, by performing region-based cross-correlation or phase correlation operation between the reference structure feature map and the structure feature maps of the remaining frames, the initial matching relationship of the corresponding regions in the images is found. These operations are performed on the structure feature maps, effectively reducing the interference of the brightness change on the matching process. The obtained initial region correspondences may contain incorrect matches, so consistency verification is required, such as by fitting a geometric transformation model and eliminating the correspondences that do not conform to the model. Finally, based on the reliable region correspondences that have been verified and meet the consistency conditions, the accurate spatial offset of each frame image relative to the reference frame is calculated. Thus, the accurate spatial offset information required for subsequent image registration is obtained, ensuring that when calculating the polarization characteristic parameters, the brightness values of different frames correspond to the same physical point in the scene, and improving the accuracy of polarization characteristic analysis.

[0031] Furthermore, the structure feature map contains the edge information or gradient direction information of the image to weaken the influence of the scene illumination change independent of the polarizer rotation on the pixel intensity in the multi-frame polarization brightness information.

[0032] Among them, the structural feature map contains the edge information or gradient direction information of the image. The edge information can be obtained by applying edge detection algorithms, such as the Canny edge detector or the Sobel operator, which can identify the boundaries of regions where the brightness in the image changes significantly. The gradient direction information can be obtained by calculating the gradients of the image in the horizontal and vertical directions and determining the direction in which the brightness changes fastest at each pixel point. Compared with the original pixel brightness values, this information is more robust to changes in the overall or local illumination intensity of the image. Thus, the generated structural feature map can more stably represent the geometric structure features of the image, which are the basis for determining the relative spatial positions between different frame images.

[0033] Further, step S2224 includes: S22241: Extract and match more than two sets of feature point coordinates from multiple frames of polarized brightness images, directly fit the affine transformation matrix using the least squares method, filter the matching points with excessive errors in the initial transformation according to the preset residual threshold, and re - perform the least squares fitting on the remaining matching points to obtain the final geometric transformation model; S22242: Calculate the degree of conformity between each corresponding relationship in the initial region correspondence and the geometric transformation model; S22243: Based on the calculated degree of conformity, determine which corresponding relationships in the initial region correspondence satisfy the region correspondence with the consistency condition.

[0034] When a camera captures a target scene through a transparent medium, due to the possible relative displacement or attitude change between the camera and the target scene, there is a spatial offset between multiple frames of polarization brightness information collected at different times. The initial region correspondence determined previously through region-based cross-correlation operation or phase correlation operation may contain inaccurate correspondences, affecting the accuracy of determining the spatial offset. To solve this problem, this solution introduces a global geometric transformation model based on feature point matching to verify the reliability of the initial region correspondence. First, feature points with unique textures or structures are extracted from multiple frames of polarization brightness images and matched. For example, feature extraction algorithms such as SIFT, SURF, or ORB can be used, and descriptor matching methods are used to find corresponding point pairs. Then, using the coordinates of these matched feature points, an affine transformation matrix describing the overall spatial transformation between frames is fitted by the least squares method. To improve the robustness of the model, an iterative optimization process is adopted. According to a preset residual threshold, the matching points with large deviations from the preliminary model are removed, and then the remaining matching points are used to refit to obtain a geometric transformation model that more accurately reflects the overall displacement between frames. This model provides a global and relatively reliable reference for the spatial correspondence between frames. Then, this geometric transformation model is used to evaluate each initial region correspondence. For each region correspondence obtained through region correlation operation, the difference or degree of conformity between this correspondence and the correspondence predicted by the global geometric transformation model is calculated.

[0035] For example, the distance between the position of the center point of the region predicted in the current frame after geometric transformation of the coordinates in the reference frame and the position in the current frame actually determined by this region correspondence can be calculated. Finally, according to the calculated degree of conformity, a discrimination criterion is set, such as the distance being less than a certain threshold. Only those region correspondences whose degree of conformity reaches the preset standard are considered region correspondences that meet the consistency condition. These region correspondences determined to be consistent are subsequently used to determine the precise spatial offset of each frame relative to the reference frame. By this method, using the global geometric model obtained by feature point matching as a reference, the incorrect correspondences that may occur in the region correlation operation are effectively filtered out, ensuring that the region correspondences used to determine the spatial offset are reliable, thereby improving the accuracy of spatial offset calculation and subsequent image registration, and ultimately contributing to more accurate calculation of polarization characteristic parameters.

[0036] Further, step S25 includes: S251: Monitor the temporal variation of the polarization characteristic parameters of the brightness change part during the process of collecting each brightness information at different polarization directions to obtain the temporal variation of the polarization characteristic parameters; S252: When the variation amplitude of the polarization parameter exceeds a preset threshold, adjust the strategy or parameters used for excluding or compensating the identified brightness change part dominated by the light source or object with polarization characteristics in the target scene according to the time variation of the obtained polarization characteristic parameters, so as to obtain the adjusted exclusion or compensation processing strategy or parameters; S253: Apply the adjusted exclusion or compensation processing strategy or parameters to exclude or compensate the identified brightness change part dominated by the light source or object with polarization characteristics in the target scene, and calculate the contribution amount of the reflected light on the transparent medium surface to the total brightness.

[0037] Specifically, first, during the process of collecting multiple frames of polarization brightness information, continuously calculate the polarization characteristic parameters of the brightness change region and record their time variation. Then, set a threshold for detecting significant changes in the polarization parameter. When the monitored variation amplitude of the polarization parameter exceeds this threshold, it indicates that the polarization source in the target scene has changed dynamically. At this time, the system dynamically adjusts the processing strategy or parameters used to exclude or compensate the influence of the polarization source of the target scene according to the monitored time variation of the polarization parameter, such as the variation amplitude or trend. For example, if the degree of polarization drops significantly, it may be necessary to reduce the compensation weight for the polarized light component in this region. Finally, apply the adjusted exclusion or compensation strategy or parameters to process the identified brightness change part dominated by the polarization source of the target scene. In this way, even if the polarization characteristics in the target scene change over time, this part of the influence can be more accurately separated from the total brightness change, so as to more precisely calculate the contribution amount of the reflected light on the transparent medium surface to the total brightness. Thereby, the ability to distinguish the reflected light of the transparent medium and the polarization light of the target scene itself in a complex dynamic scene is improved, and further the accuracy of determining the contribution amount of the reflected light on the transparent medium surface to the total brightness is enhanced.

[0038] In some specific embodiments, a time window can be set, such as the acquisition period of the most recent N frames of multi-frame polarization luminance information. After the end of each acquisition period, the polarization characteristic parameters (such as the degree of polarization) of the luminance change region are calculated. Step S251 can calculate the sequence of degrees of polarization of this region within the most recent N periods. Step S252 can calculate the standard deviation of this sequence of degrees of polarization or the difference between the maximum value and the minimum value. If the standard deviation or the difference exceeds a preset threshold (such as 0.05), it is determined that a significant change in the polarization characteristics has occurred. At this time, the compensation coefficient used to compensate for the influence of the polarization light of the target scene itself can be adjusted according to the magnitude of the standard deviation or the difference. For example, the compensation coefficient can be inversely proportional to this change amount, and the greater the change, the greater the adjustment amplitude of the compensation coefficient. Step S253 applies the adjusted compensation coefficient to perform a compensation process on the luminance change part caused by the polarization source of the target scene itself identified in the current acquisition period, so as to calculate the contribution amount of the reflected light on the surface of the transparent medium to the total luminance.

[0039] Further, step S3 includes: S31: Based on the contribution amount of the reflected light on the surface of the transparent medium to the total luminance, perform a first correction process on at least one of the first luminance information and the second luminance information or their combination to obtain intermediate luminance information with the influence of the reflected light on the surface of the transparent medium weakened; S32: Obtain the spectral transmittance characteristics or color characteristic parameters of the transparent medium; S33: Based on the obtained spectral transmittance characteristics or color characteristic parameters, perform a second correction process on the intermediate luminance information. The second correction process is used to compensate for the spectral change or intensity attenuation caused by the transparent medium to the transmitted light of the target scene, so as to obtain the target luminance characterizing the target scene on the other side of the transparent medium.

[0040] Among them, the first correction process removes the reflected light component from the original luminance information based on the contribution amount of the reflected light on the surface of the transparent medium to the total luminance. This can be achieved by subtracting the calculated reflected light contribution amount from the total luminance, or by establishing a model to separate the reflected light and transmitted light components. Obtaining the spectral transmittance characteristics or color characteristic parameters of the transparent medium can be achieved in various ways. For example, when the system is installed, the spectral transmittance of the transparent medium is measured, or according to the type of the transparent medium (such as ordinary glass, low-iron glass, coated glass, etc.), the preset optical parameter database is consulted. These parameters can be represented as transmittance curves at different wavelengths, or simplified as a color correction matrix or look-up table describing color shift and overall attenuation. The second correction process uses these obtained parameters to compensate the intermediate luminance information after the first correction. If the obtained is a spectral transmittance curve, the spectral distribution of the intermediate luminance information can be adjusted according to this curve; if the obtained are color characteristic parameters, the corresponding color correction algorithm or look-up table can be applied to compensate for the color deviation and luminance attenuation caused by the transparent medium.

[0041] Specifically, the method first uses polarization information analysis technology to separate the contribution of the reflected light from the transparent medium surface to the total luminance from the total luminance information including reflected light and transmitted light collected by the camera. This separation process identifies the characteristics of the luminance changing with the polarization direction to distinguish the reflected light and the transmitted light of the target scene. Based on this contribution, a first correction process is performed to subtract or compensate for the influence of the reflected light from the original luminance information, obtaining an intermediate luminance information that primarily reflects the transmitted light. This step solves the problem of the overestimated total luminance caused by the interference of the reflected light. However, the transparent medium itself is not perfectly transparent. It will absorb or scatter part of the light and may have different attenuations for lights of different wavelengths, thus changing the intensity and color of the transmitted light. Therefore, the spectral transmission characteristics or color characteristic parameters of the transparent medium are further obtained. Finally, a second correction process is performed on the intermediate luminance information based on these parameters. This process compensates for the attenuation and spectral change of the transmitted light by the transparent medium, enabling the finally obtained luminance information to more accurately reflect the true luminance of the target scene on the other side of the transparent medium. By combining the removal of the reflected light and the compensation of the medium transmission, this method improves the accuracy of determining the luminance of the target scene in the presence of the interference of the transparent medium, providing accurate input for subsequent automatic exposure control.

[0042] Reference Figure 2 As shown in the figure, a visible light luminance detection system for a camera includes: Polarization information acquisition module: acquires the first luminance information of the scene within the camera's field of view when the linear polarizer is in the first polarization direction when the camera shoots the target scene through the transparent medium, and the second luminance information of the scene within the camera's field of view when the linear polarizer is in a direction different from the first polarization direction; Characteristic analysis and identification module: based on the first luminance information and the second luminance information, analyzes the characteristics of the luminance changing with the polarization direction of the linear polarizer to distinguish the part of the reflected light from the transparent medium surface and the part of the transmitted light of the target scene, and determines the contribution of the reflected light from the transparent medium surface to the total luminance; Luminance correction module: based on the contribution of the reflected light from the transparent medium surface to the total luminance, corrects at least one or a combination of the first luminance information and the second luminance information to obtain the target luminance characterizing the target scene on the other side of the transparent medium.

[0043] Specifically, the system solves the problem that when the camera shoots through a transparent medium, the reflected light on the surface of the transparent medium interferes with the brightness measurement, resulting in insufficient exposure of the target scene. When the camera shoots through a transparent medium such as a glass window, the light received by the camera is a mixture of the transmitted light of the target scene and the reflected light on the surface of the transparent medium. When the reflected light is strong, the existing brightness measurement methods cannot distinguish these two kinds of light, resulting in a higher measured total brightness. If automatic exposure is performed based on this higher brightness, the camera will wrongly reduce the exposure amount, making the target scene that really needs to be photographed under-exposed. The system obtains the scene brightness information under at least two different polarization directions by introducing a polarization information acquisition module. This information is sent to the characteristic analysis and recognition module, which analyzes the variation law of brightness with the polarization direction by using the difference in the polarization characteristics of the reflected light and the transmitted light. Through this analysis, the system can identify and quantify the specific contribution of the reflected light on the surface of the transparent medium to the total brightness. Subsequently, the brightness correction module uses this contribution amount to correct the original brightness information. For example, the contribution of the reflected light can be subtracted from the total brightness to obtain information closer to the true brightness of the target scene. The finally output corrected brightness information accurately characterizes the brightness of the target scene on the other side of the transparent medium, excluding the interference of the reflected light. Based on this accurate target brightness information for automatic exposure control, it can ensure that the target scene obtains appropriate exposure, avoid under-exposure caused by reflection interference, and thus improve the image acquisition quality.

[0044] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed.

Claims

1. A method for detecting the visible light brightness of a camera, characterized in that, The method includes the following steps: S1: Obtain the first brightness information of the scene within the camera's field of view when the linear polarizer in the target scene is in the first polarization direction through the transparent medium by the camera, and the second brightness information of the scene within the camera's field of view when the linear polarizer is in a direction different from the first polarization direction; S2: Based on the first brightness information and the second brightness information, analyze the characteristics of the brightness change as the polarization direction of the linear polarizer changes, so as to distinguish the part of the reflected light from the surface of the transparent medium and the part of the transmitted light of the target scene, and determine the contribution amount of the reflected light from the surface of the transparent medium to the total brightness; S3: Based on the contribution amount of the reflected light from the surface of the transparent medium to the total brightness, correct at least one or a combination of the first brightness information and the second brightness information to obtain the target brightness characterizing the target scene on the other side of the transparent medium.

2. The visible light brightness detection method for a camera according to claim 1, wherein Step S2 includes: S21: Obtain the brightness information collected at different polarization directions, and each brightness information includes the first brightness information, the second brightness information, and optionally one or more optional brightness information collected at other different polarization directions; S22: Calculate the polarization characteristic parameters of the region where the brightness changes as the polarization direction of the linear polarizer changes within the camera's field of view; S23: Compare the polarization characteristic parameters calculated for the brightness change region with the range of typical polarization characteristic parameters in a preset reference model to generate a comparison result; S24: According to the comparison result, identify the source of the brightness change, and distinguish the part dominated by the reflected light from the surface of the transparent medium and the part dominated by the light source or object with polarization characteristics in the target scene itself; S25: Exclude or compensate the part of the brightness change dominated by the light source or object with polarization characteristics in the target scene itself to determine the contribution amount of the reflected light from the surface of the transparent medium to the total brightness.

3. The visible light brightness detection method for a camera according to claim 2, wherein, Step S22 includes: S221: Collectively refer to the first brightness information and the second brightness information in step S21, and optionally one or more brightness information collected at other different polarization directions as multi-frame polarization brightness information; S222: Determine the spatial offset amount of the corresponding scene regions in each frame of brightness information caused by the relative displacement or attitude change between the camera and the target scene during the acquisition of the multi-frame polarization brightness information; S223: According to the determined spatial offset amount, perform image registration processing on the multi-frame polarization brightness information to generate a set of spatially aligned multi-frame polarization brightness information, ensuring that the brightness values used to calculate the polarization characteristics of the same point in the spatially aligned multi-frame polarization brightness information correspond to the same physical region of the target scene; S224: Based on the spatially aligned multi-frame polarization brightness information, calculate the polarization characteristic parameters of the region where the brightness changes as the polarization direction of the linear polarizer occurs within the camera's field of view.

4. The visible light brightness detection method for a camera according to claim 3, wherein, The polarization characteristic parameters include one or a combination of two items, namely the degree of polarization and the polarization main axis direction calculated according to the corresponding brightness information.

5. A method for detecting the visible light brightness of a camera according to claim 3, characterized in that, Step S222 includes: S2221: Apply an image structure extraction algorithm to each frame image in the multi-frame polarization luminance information to generate a structure feature map corresponding to the frame image; S2222: Select one frame from the multi-frame polarization luminance information as a reference frame, and use the structure feature map corresponding to the reference frame image as a reference structure feature map; S2223: Determine an initial region correspondence between the reference structure feature map and the structure feature maps corresponding to other frame images except the reference frame by performing region-based cross-correlation operation or phase correlation operation; S2224: Perform consistency verification on the determined initial region correspondence, and determine the spatial offset of each frame relative to the reference frame based on the region correspondence that meets the consistency condition.

6. The visible light brightness detection method for a camera according to claim 5, characterized in that, The structure feature map contains edge information or gradient direction information of the image to weaken the influence of scene illumination changes independent of the rotation of the polarizer on the pixel intensity in the multi-frame polarization luminance information.

7. A method for detecting the visible light brightness of a camera according to claim 5, characterized in that, Step S2224 includes: S22241: Extract and match more than two sets of feature point coordinates from the multi-frame polarization luminance images, directly fit an affine transformation matrix using the least squares method, filter the matching points with excessive errors in the initial transformation according to a preset residual threshold, and re-perform least squares fitting on the remaining matching points to obtain a final geometric transformation model; S22242: Calculate the degree of conformity between each correspondence in the initial region correspondence and the geometric transformation model; S22243: Determine which correspondences in the initial region correspondence meet the region correspondence of the consistency condition based on the calculated degree of conformity.

8. The visible light brightness detection method for a camera according to claim 2, wherein Step S25 includes: S251: Monitor the time variation of the polarization characteristic parameters of the luminance change part during the acquisition of each luminance information at different transmission directions to obtain the time variation of the polarization characteristic parameters; S252: When the change amplitude of the polarization parameter exceeds a preset threshold, adjust the strategy or parameters used for excluding or compensating the identified luminance change part dominated by the light source or object with polarization characteristics in the target scene based on the obtained time variation of the polarization characteristic parameters to obtain an adjusted exclusion or compensation processing strategy or parameters; S253: Apply the adjusted exclusion or compensation processing strategy or parameters to exclude or compensate the identified luminance change part dominated by the light source or object with polarization characteristics in the target scene, and calculate the contribution amount of the reflected light on the surface of the transparent medium to the total luminance.

9. A method for detecting the visible light brightness of a camera according to claim 1, characterized in that, Step S3 includes: S31: Perform a first correction process on at least one of or a combination of the first luminance information and the second luminance information based on the contribution amount of the reflected light on the surface of the transparent medium to the total luminance to obtain intermediate luminance information with the influence of the reflected light on the surface of the transparent medium weakened; S32: Obtain the spectral transmission characteristic or color characteristic parameters of the transparent medium; S33: Based on the obtained spectral transmission characteristics or color characteristic parameters, perform a second correction process on the intermediate brightness information. The second correction process is used to compensate for the spectral change or intensity attenuation caused by the transparent medium to the transmitted light of the target scene, so as to obtain the target brightness characterizing the target scene on the other side of the transparent medium.

10. A visible light brightness detection system for a camera, characterized in that, The system includes: A polarization information acquisition module: acquires the first brightness information of the scene within the camera's field of view when the linear polarizer is in the first polarization direction for the camera to photograph the target scene through the transparent medium, and the second brightness information of the scene within the camera's field of view when the linear polarizer is in a direction different from the first polarization direction; A characteristic analysis and identification module: based on the first brightness information and the second brightness information, analyze the characteristics of the change in brightness with the change in the polarization direction of the linear polarizer, so as to distinguish the part of the reflected light from the surface of the transparent medium and the part of the transmitted light of the target scene, and determine the contribution amount of the reflected light from the surface of the transparent medium to the total brightness; A brightness correction module: based on the contribution amount of the reflected light from the surface of the transparent medium to the total brightness, correct at least one or a combination of the first brightness information and the second brightness information, so as to obtain the target brightness characterizing the target scene on the other side of the transparent medium.

Images