WO2021223113A1

WO2021223113A1 - Metering method, camera, electronic device, and computer-readable storage medium

Info

Publication number: WO2021223113A1
Application number: PCT/CN2020/088814
Authority: WO
Inventors: 王浩伟; 韩守谦; 郑子翔
Original assignee: 深圳市大疆创新科技有限公司
Priority date: 2020-05-06
Filing date: 2020-05-06
Publication date: 2021-11-11

Abstract

The embodiments of the present application disclose a metering method, comprising: acquiring brightness information respectively corresponding to a plurality of pixels in an image to be metered; determining weight information respectively corresponding to the plurality of pixels, the weight information being determined according to the category of image content corresponding to the pixels; and calculating the overall brightness of said image according to the brightness information and the weight information respectively corresponding to the plurality of pixels. In the metering method, when calculating the overall brightness of said image, the weight information corresponding to the pixels used is determined according to the category of image content corresponding to the pixels; in this way, a camera can focus on the image content itself, and focus on different objects in a scene. Compared with the existing metering method, the present invention has broader scene applications.

Description

Photometry method, camera, electronic device, computer readable storage medium

Technical field

This application relates to the field of image processing technology, and in particular to a photometry method, a camera, an electronic device, and a computer-readable storage medium.

Background technique

Metering is the method used to measure the brightness of the object being photographed. Through metering, the camera can use its own automatic exposure algorithm to adjust the shooting parameters accordingly, so that the captured images can have proper exposure. There are multiple metering modes, such as evaluative metering, center-weighted average metering, spot metering, etc. Although these metering modes have their own applicable scenes, each metering mode still has its own limitations and cannot be applied in some scenes.

Summary of the invention

In view of this, the embodiments of the present application provide a photometry method, a camera, an electronic device, and a computer-readable storage medium, which have wider applicability and break the limitations of the traditional photometry mode.

The first aspect of the embodiments of the present application provides a photometry method, including:

Obtain the brightness information corresponding to each of the multiple pixels in the light image to be measured;

Determining weight information corresponding to each of the multiple pixels; wherein the weight information is determined according to the category of the image content corresponding to the pixel;

Calculate the overall brightness of the light image to be measured according to the brightness information and the weight information corresponding to each of the plurality of pixels.

A second aspect of the embodiments of the present application provides a camera, including: a body, a lens connected to the body, an image sensor, a processor, and a memory storing a computer program provided in the body;

The image sensor is used to collect an image to be measured through the lens;

The processor implements the following steps when executing the computer program:

Acquiring, from the image sensor, brightness information corresponding to each of the plurality of pixels in the light image to be measured;

A third aspect of the embodiments of the present application provides an electronic device, including: a camera mounted on the electronic device, a processor, and a memory storing a computer program;

Acquiring, by the camera, brightness information corresponding to each of multiple pixels in the light image to be measured;

The fourth aspect of the embodiments of the present application provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements any of the photometry provided in the above-mentioned first aspect. method.

In the light metering method provided by the embodiments of this application, the weight information corresponding to the pixel used when calculating the overall brightness of the light image to be metered is determined according to the category of the image content corresponding to the pixel. Such a light metering method can focus on the image content itself. , Pay attention to different objects in the scene, and have wider scene applicability compared with the existing metering methods. Taking portrait photography as an example, if the evaluative metering mode is used, the brightness of the characters in the images taken under and without backlight is completely different. The reason for this is that evaluative metering cannot identify which part of the image is a person. It integrates the brightness of each area in the screen, and these areas contain objects such as the sky or the road that the user does not care about. However, this application The light metering method provided by the embodiment focuses on the image content, so it can focus on people in various environments for light metering, so that the brightness of the people in the picture is relatively suitable regardless of whether it is shot under backlight or without backlight. .

Description of the drawings

In order to more clearly describe the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings that need to be used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained from these drawings without creative labor.

Fig. 1 is a flowchart of a photometry method provided by an embodiment of the present application.

Fig. 2 is a schematic diagram of semantic segmentation provided by an embodiment of the present application.

Fig. 3 is a schematic structural diagram of a camera provided by an embodiment of the present application.

Fig. 4 is a schematic structural diagram of an unmanned aerial vehicle provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of this application.

For electronic devices with shooting functions, automatic exposure (AE, Auto Exposure) is a very common function. For example, electronic devices such as smart phones, tablet computers, drones, cameras (including sports cameras, infrared cameras, etc.) are all equipped with automatic exposure algorithms in their software. The automatic exposure algorithm can automatically adjust the camera's shooting parameters (mainly adjusting the three elements of exposure: aperture value, shutter time, and ISO) according to the metering results, so that the captured images have better effects in brightness and quality.

Metering is the method used to measure the brightness of the object being photographed. Specific to the image data level, photometry can be a process of calculating image brightness. When the image sensor collects an image, it can collect the brightness information corresponding to each pixel in the image. For the entire image, a brightness value is needed as the overall brightness of the image. It is easy to understand that when using multiple pixels When the corresponding brightness information calculates the overall brightness of an image, there are multiple calculation methods, and different calculation methods correspond to different metering modes, such as evaluative metering, center-weighted average metering, and spot metering.

Different metering modes can meet different shooting scenes. For example, evaluative metering is suitable for sceneries such as scenery and capturing. In terms of metering method, it divides the viewfinder into several metering areas, and each metering area calculates the brightness separately and then integrates the weighting. Therefore, the overall brightness of the image calculated by the evaluative metering takes into account the conditions of each area of the image. After the shooting parameters are adjusted according to the overall brightness through the automatic exposure algorithm, the brightness of each area in the captured image is then taken. All are relatively moderate, and there will be no over-exposure or under-exposure in a certain area. Another example is spot metering. Spot metering uses a small area (1% to 3%) in the center of the viewfinder screen as the reference point, and calculates the brightness of this small area as the overall brightness of the image. Therefore, it is based on the overall brightness. After the shooting parameters are adjusted, the brightness of the small area where the reference point is located in the captured image is moderate, while the part outside the small area may not be overexposed or underexposed.

Although the various metering modes mentioned above have their own applicable scenes, each metering mode still has its own limitations, and there are many incompetent scenes.

To this end, an embodiment of the present application provides a photometry method, which can be applied to a camera or an electronic device equipped with a camera. Refer to Fig. 1, which is a flow chart of a photometry method provided by an embodiment of the present application. The method includes the following steps:

S101: Acquire brightness information corresponding to each of multiple pixels in the light image to be measured.

S102. Determine weight information corresponding to each of the multiple pixels according to the category of the image content corresponding to the multiple pixels.

S103: Calculate the overall brightness of the light image to be measured according to the brightness information and weight information corresponding to each of the multiple pixels.

When the image sensor collects an image, it can collect information corresponding to each pixel of the image, which includes the brightness information corresponding to each pixel. For example, in an implementation, if the YUV color coding method is adopted, the information corresponding to the pixel may include the luminance information of the Y channel and the chromaticity information of the UV channel. In short, the brightness information corresponding to the pixel can be easily obtained by those skilled in the art.

When performing photometry, it is not necessary to use the brightness information of all pixels of the photometric image. For example, some existing metering modes, such as spot metering, partial metering, etc., only use the brightness information of pixels in a part of the image for metering. Therefore, in step S101, the brightness information corresponding to the required pixels can be obtained according to the photometry methods in different embodiments.

The category corresponding to the pixel is the category on the image content. Specifically, the image content is the content that can be visually observed and expressed on the image. In one implementation, the pixel corresponding to the category can be the sky, ground, water, people, cars, plants, animals and other environmental and/or object categories, in another implementation, it can also be high temperature, low temperature, high radioactivity , Low-level radioactivity and other attributes.

There are many ways to determine the category of the image content corresponding to the pixel. For example, in one implementation, after a certain target is identified in the image, the center point of the target can be located, and the center point near the center point is determined. The category corresponding to the pixels in the range is determined to be the category corresponding to the target. When recognizing a target in an image, you can also flexibly select a certain area in the image for recognition. For example, considering people’s shooting habits, you can only recognize the central part of the image, and assign the recognition result to the central part of the pixel. Corresponding categories, and the corresponding types of pixels in other parts can be directly the default types.

In another implementation, this can be achieved by semantic segmentation of the photometric image. Semantic segmentation is to classify images at the pixel level. It can be considered that after semantic segmentation, each pixel in the image will be assigned a category. For example, the pixels corresponding to the sky in the image content can be assigned or labeled with the category information of the sky. (Category tag), the pixel of the building corresponding to the image content can be assigned or marked with the category information of the building. The category information can be specifically represented by various symbols or combinations of symbols. For example, the category information of the sky can be 0, and the category information of the road can be 1 and so on.

There are many ways to implement semantic segmentation. In one implementation, you can use traditional methods such as TextonForest, random forest classifier, gray-scale segmentation, etc., while in another implementation, you can use semantic segmentation models to perform semantics. segmentation. Specifically, the light image to be measured may be input to the semantic segmentation model to obtain the first semantic map output by the semantic segmentation model.

The semantic segmentation model may be a model based on deep learning. There are many kinds of semantic segmentation models based on deep learning. In an optional implementation, the semantic segmentation model can choose the full convolutional neural network FCN model, which is based on the CNN model. Of course, U-Net, MultiScale and other models can also be used.

The first semantic map, in a specific example, may be a matrix corresponding to the size of the photometric image. Refer to FIG. 2, which is a schematic diagram of semantic segmentation provided by an embodiment of the present application. In Figure 2, the size of the photometric image can be 12*9, that is, it includes 12*9 pixels, including people, sky, and road surfaces. After semantic segmentation is performed on the light image to be measured, the output first semantic map can be obtained. The first semantic map may be a matrix with a size of 12*9, each element of which corresponds to an area in the light image to be measured, and the value of each element corresponds to the category of the area. It can be seen that in the example of FIG. 2, the category information corresponding to the sky is 0, the category information corresponding to the road surface is 1, and the category information corresponding to the person is 2. It should be noted that the example shown in Figure 2 is only an easy to understand example. The size of the image in the example is different from the size of the actual image. For example, the size of the actual photometric image can be 4080*2720. , Which includes 4080*2720 pixels.

It is easy to understand that, in order to reduce the amount of calculation of semantic segmentation, the image input to the semantic segmentation model may not be the original image of the photometric image, but an image after the photometric image is down-sampled.

Since the first semantic map contains category information corresponding to each pixel of the photometric image, the weight information corresponding to each pixel of the photometric image can be determined through the first semantic map. But considering that when calculating the overall brightness of the image, it is not necessary to use the brightness information of all pixels of the light image to be measured (this approach will take up larger computing resources), therefore, in an optional implementation, Before determining the weight information corresponding to the pixel through the first semantic map, the first semantic map may be down-sampled to obtain a second semantic map of a preset size, and then the weight information corresponding to the pixel is determined on the second semantic map.

It should be noted that the preset size corresponds to the size of the predetermined metering weight matrix. The metering weight matrix is a matrix composed of weight information, and the value of each element in the matrix can correspond to a weight value. Through the dot product calculation of the matrix formed by the metering weight matrix and the corresponding brightness information, the overall brightness of the image can be calculated. The metering weight matrix is usually much smaller than the size of the image. For example, the metering weight matrix can be 16×16 or 32×32. Therefore, to determine the metering weight matrix, you can first downsample the first semantic map to metering The size corresponding to the weight matrix is then used to determine the photometric weight matrix formed by the weight information by using the down-sampled semantic map (ie, the second semantic map).

When down-sampling the first semantic map, the usual approach is to down-sample the matrix every other row and every column. However, the embodiment of the present application provides an optional implementation manner to obtain a more accurate reflection of the first semantic map. The overall second semantic map. Specifically, the first semantic map may be divided into multiple sub-areas according to the size to be reduced (ie, the preset size). For example, the preset size is 32×32, and the first semantic map may be divided into 32×32 sub-regions. area. For each sub-region, the category corresponding to each pixel can be counted, and the category with the highest proportion among the categories corresponding to each pixel is determined as the category corresponding to the sub-region. In this way, the matrix formed by the categories corresponding to each sub-region forms the second Semantic map.

When determining the weight information corresponding to the pixel through the second semantic map, the category corresponding to each pixel in the second semantic map can be matched with the target correspondence relationship, so as to determine the weight information corresponding to each pixel in the second semantic map. The target correspondence relationship may be a correspondence relationship including weight information corresponding to each category, for example, category 0 corresponds to a weight value of 10, category 1 corresponds to a weight value of 4, and so on. The target correspondence is also diverse in specific manifestations, such as tables, images, curves, and so on.

In an implementation, the target correspondence can be preset by the manufacturer and other internal personnel and configured in the product's memory, or the product's processor can also be automatically initialized to generate the target correspondence, and the target correspondence can be connected to the network when the target correspondence is generated. Obtained from the server, or generated according to the default settings. In another implementation, the target correspondence can also be customized by the user. For example, in a scene, users often use cameras or electronic devices equipped with the camera to take pictures of power equipment, then the user can increase the weight information corresponding to the power equipment in the target correspondence relationship, so that the power in the captured image The brightness of the device is appropriate; for another example, in another scene, the user often uses the camera or the electronic device equipped with the camera to shoot people, then the user can increase the weight information corresponding to the person in the target correspondence relationship, so that the shooting The brightness of the characters in the resulting image is moderate.

In an implementation, the category-weight information correspondence in the target correspondence may also be adjusted according to the captured image. For example, the weight information corresponding to the category can be determined according to the proportion of the image area corresponding to the category in the photometric image. Specifically, for example, the person in the photometric image occupies 50% of the screen, and the sky occupies 25%. If the road occupies 25%, it can be considered that the user's main shooting object is a person, and the weight information corresponding to the person can be adjusted or determined to be larger, while the weight information corresponding to the sky and the road can be smaller. In this implementation, the weight information corresponding to the category with the largest proportion in the photometric image will also be the largest.

In another implementation, in the target correspondence relationship, the weight information corresponding to the category may also be determined according to the position of the image area corresponding to the category in the photometric image. For example, for example, in the photometric image, the person is located in the central area, but only 15% of the photometric image, but the car next to the person is on the side of the image, but it is on the side of the image. The proportion of the image in the image is 50%. In this implementation, although the character is relatively small, the character is located in the center of the image. Therefore, the weight information corresponding to the character can be larger. Although the car is relatively large, it is located in the image. Side, so the weight information corresponding to the car can be smaller. In this implementation, the category of the pixel corresponding to the central area of the image to be measured is the category with the largest weight information.

After determining the weight information corresponding to each pixel in the second semantic map, that is, after the photometric weight matrix is determined, the photometric weight matrix can be multiplied by the brightness information matrix of the photometric image of the corresponding size to calculate the photometric image The overall brightness. Wherein, the photometric image of the corresponding size can also be obtained by down-sampling. For example, the photometric weight matrix is 32×32, and the photometric image can be down-sampled to a size of 32×32.

The calculated overall brightness can have multiple applications. For example, in one application, the overall brightness can be used to guide the adjustment of the camera's shooting parameters, or in other words, the calculated overall brightness can be provided to the automatic exposure algorithm to pass the automatic The exposure algorithm adjusts the shooting parameters of the camera. It is easy to understand that the shooting parameters of the camera can be one or a combination of three elements of exposure (aperture value, shutter time, ISO).

When the shooting parameters are specifically adjusted according to the overall brightness, the current ambient light brightness can be calculated or estimated according to the overall brightness and the current shooting parameters through the exposure equation of the APEX system. According to the determined ambient light brightness, the corresponding target shooting parameters can be determined by querying the exposure meter. Furthermore, the current shooting parameters can be adjusted according to the target shooting parameters.

For the overall brightness, in another application, the calculated overall brightness can also be compared with the preset reference brightness or standard brightness, so as to output corresponding exposure prompts to the user, such as overexposure prompt and/or underexposure prompt hint. When the camera is in manual gear, the user can also give corresponding prompts when the user selects the shooting parameters. For example, when the user reduces the shutter time to 1/2000, it can prompt under-exposure for the shutter time of 1/2000, increasing to 1. At /125, overexposure etc. can be prompted for the shutter time of 1/125.

The above is a detailed description of a photometry method provided by an embodiment of the present application. Please refer to FIG. 3 below, which is a schematic structural diagram of a camera provided by an embodiment of the present application. The camera includes: a body 310, a lens 320 connected to the body, an image sensor 311, a processor 312, and a memory 313 storing computer programs in the body;

The image sensor 311 is used to collect an image to be measured through the lens 320;

The processor 312 implements the following steps when executing the computer program:

Acquiring, from the image sensor 311, brightness information corresponding to each of a plurality of pixels in the photometric image;

Optionally, the category of the image content corresponding to the pixel is obtained by semantically segmenting the light image to be measured.

Optionally, when the processor executes the step of semantically segmenting the light image to be measured, it is used to input the light image to be measured into a semantic segmentation model to obtain a first semantic map output by the semantic segmentation model ; Wherein, the first semantic map includes the category of the image content corresponding to the pixel.

Optionally, the semantic segmentation model is a pre-trained convolutional neural network CNN model.

Optionally, the image input to the semantic segmentation model is the down-sampled photometric image.

Optionally, the processor is further configured to down-sample the first semantic map to obtain a second semantic map of a preset size before determining the weight information corresponding to the pixel through the first semantic map .

Optionally, the preset size corresponds to a predetermined size of the photometric weight matrix.

Optionally, when the processor executes the step of down-sampling the first semantic map, the processor is configured to divide the first semantic map into a plurality of sub-areas according to the preset size; and determine each of the The category corresponding to the sub-region; the category corresponding to the sub-region is the category with the highest proportion among the categories corresponding to the pixels in the sub-region; the second semantic map is generated according to the category corresponding to each of the sub-regions.

Optionally, the weight information is determined according to the corresponding relationship between the category of the image content corresponding to the pixel and the target, and the target corresponding relationship includes weight information corresponding to each category.

Optionally, the target correspondence relationship is preset by the user or the system.

Optionally, in the target correspondence relationship, the weight information corresponding to each category is determined according to the proportion of the image area corresponding to the category in the photometric image.

Optionally, the image area corresponding to the category with the largest weight information occupies the largest proportion in the photometric image.

Optionally, in the target correspondence relationship, the weight information corresponding to each category is determined according to the position of the image area corresponding to the category in the photometric image.

Optionally, the image area corresponding to the category with the largest weight information is located in the center of the photometric image.

Optionally, the overall brightness is used to guide the adjustment of shooting parameters.

Optionally, when the processor executes the step of adjusting the shooting parameters according to the overall brightness, it is used to determine the brightness of the ambient light according to the overall brightness and the current shooting parameters; and to determine the target shooting parameters according to the brightness of the ambient light ; Adjust the current shooting parameters according to the target shooting parameters.

Optionally, the shooting parameters include any one of the following: aperture value, shutter time, and sensitivity ISO.

Optionally, the overall brightness is used to compare with a preset reference brightness to output a corresponding exposure prompt.

Optionally, the exposure prompt includes: an overexposure prompt and/or an underexposure prompt.

In the camera provided by the embodiment of the present application, the weight information corresponding to the pixel used when calculating the overall brightness of the light image to be measured is determined according to the category of the image content corresponding to the pixel. In this way, the image content itself can be focused on the scene. Compared with the existing camera's metering method, it has wider scene applicability. Taking portrait photography as an example, if the evaluative metering mode is used, the brightness of the characters in the images taken under and without backlight is completely different. The reason for this is that evaluative metering cannot identify which part of the image is a person. It integrates the brightness of each area in the screen, and these areas contain objects such as the sky or the road that the user does not care about. However, this application The camera provided by the embodiment focuses on the image content, so it can perform light metering focusing on the person in various environments, so that the brightness of the person in the picture is relatively suitable whether it is shot under backlight or without backlight.

Optionally, the foregoing process may also be executed by a device including a processor and a memory, and the image sensor of another device may send the collected image to the device to execute the process involved in the foregoing photometry method.

The foregoing is a detailed description of a camera provided by an embodiment of the present application. The embodiment of the present application also provides an electronic device. The electronic device is an electronic device equipped with a camera. Specifically, it may be a mobile phone, a smart speaker, a drone, and so on. The following takes a drone as an example for description. You can refer to FIG. 4. FIG. 4 is a schematic structural diagram of a drone provided by an embodiment of the present application. The drone includes a camera 401 and a processor 402 mounted on the drone. And a memory 403 storing computer programs.

The processor 402 implements the following steps when executing the computer program:

Acquiring, through the camera 401, brightness information corresponding to each of multiple pixels in the light image to be measured;

Optionally, the category of the image content corresponding to the pixel is obtained by performing semantic segmentation on the light image to be measured.

Optionally, when the processor executes the step of down-sampling the first semantic map, the processor is configured to divide the first semantic map into a plurality of sub-regions according to the preset size; and determine each of the The category corresponding to the sub-region; the category corresponding to the sub-region is the category with the highest proportion among the categories corresponding to the pixels in the sub-region; and the second semantic map is generated according to the category corresponding to each of the sub-regions.

Optionally, the electronic equipment includes a drone.

In the drone provided by the embodiments of the present application, the weight information corresponding to the pixels used by the camera to calculate the overall brightness of the light image to be measured is determined according to the category of the image content corresponding to the pixel, so that the camera can focus on the image The content itself pays attention to different objects in the scene, and has wider scene applicability than the light metering method of the camera mounted in the existing drone. Taking portrait photography as an example, if the evaluative metering mode is used, the brightness of the characters in the images taken under and without backlight is completely different. The reason for this is that evaluative metering cannot identify which part of the image is a person. It integrates the brightness of each area in the screen, and these areas contain objects such as the sky or the road that the user does not care about. However, this application The drone provided by the embodiment, due to the focus on the content of the image, can measure light focusing on the characters in various environments. In this way, the brightness of the characters in the picture is relatively suitable regardless of whether it is shot under backlight or without backlight. .

As long as there is no conflict or contradiction between the technical features provided in the above embodiments, those skilled in the art can combine the various technical features according to actual conditions to form various embodiments. However, the document of this application is limited in length and does not describe various embodiments. However, it is understandable that various embodiments also belong to the scope of the disclosure of the embodiments of this application.

The embodiments of the present application may adopt the form of a computer program product implemented on one or more storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing program codes. Computer usable storage media include permanent and non-permanent, removable and non-removable media, and information storage can be realized by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to: phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, Magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media can be used to store information that can be accessed by computing devices.

It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply one of these entities or operations. There is any such actual relationship or order between. The terms "including", "including" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements that are not explicitly listed. Elements, or also include elements inherent to such processes, methods, articles, or equipment. If there are no more restrictions, the element defined by the sentence "including a..." does not exclude the existence of other identical elements in the process, method, article, or equipment that includes the element.

The method and equipment provided by the embodiments of the application are described in detail above. Specific examples are used in this article to describe the principles and implementations of the embodiments of the application. The description of the above embodiments is only used to help understand the embodiments of the application. At the same time, for those of ordinary skill in the art, according to the ideas of the embodiments of this application, there will be changes in the specific implementation and the scope of application. In summary, the content of this specification should not be It is understood as a limitation to the embodiments of the present application.

Claims

A photometric method, characterized in that it comprises:

Obtain the brightness information corresponding to each of the multiple pixels in the light image to be measured;

Determining weight information corresponding to each of the multiple pixels; wherein the weight information is determined according to the category of the image content corresponding to the pixel;

Calculate the overall brightness of the light image to be measured according to the brightness information and the weight information corresponding to each of the plurality of pixels.
The photometry method according to claim 1, wherein the category of the image content corresponding to the pixel is obtained by semantically segmenting the photometry image.
The photometry method of claim 2, wherein the semantic segmentation of the photometric image comprises:

The light image to be measured is input into a semantic segmentation model to obtain a first semantic map output by the semantic segmentation model; wherein, the first semantic map includes the category of the image content corresponding to the pixel.
The photometry method according to claim 3, wherein the semantic segmentation model is a pre-trained convolutional neural network CNN model.
The photometry method according to claim 3, wherein the image input to the semantic segmentation model is the down-sampled photometry image.
The photometry method according to claim 3, wherein before determining the weight information corresponding to the pixel through the first semantic map, the method further comprises:

Down-sampling the first semantic map to obtain a second semantic map of a preset size.
7. The photometry method of claim 6, wherein the preset size corresponds to a predetermined size of the photometry weight matrix.
The photometry method of claim 6, wherein the down-sampling of the first semantic map comprises:

Dividing the first semantic map into a plurality of sub-areas according to the preset size;

Determine the category corresponding to each of the sub-regions; the category corresponding to the sub-region is the category with the highest proportion among the categories corresponding to the pixels in the sub-region;

The second semantic map is generated according to the category corresponding to each of the sub-regions.
The photometry method according to claim 1, wherein the weight information is determined according to the corresponding relationship between the category of the image content corresponding to the pixel and the target, and the target corresponding relationship includes weight information corresponding to each category.
The photometry method according to claim 9, wherein the target correspondence relationship is preset by the user or the system.
The photometry method according to claim 9, wherein in the target correspondence relationship, the weight information corresponding to the category is determined according to the proportion of the image area corresponding to the category in the photometric image.
The photometry method according to claim 11, wherein the image area corresponding to the category with the largest weight information occupies the largest proportion in the photometry image.
The photometry method according to claim 9, wherein in the target correspondence relationship, the weight information corresponding to the category is determined according to the position of the image area corresponding to the category in the photometric image.
The photometry method according to claim 13, wherein the image area corresponding to the category with the largest weight information is located in the center of the photometry image.
The light metering method according to claim 1, wherein the overall brightness is used to guide the adjustment of shooting parameters.
The light metering method according to claim 15, wherein adjusting the shooting parameters according to the overall brightness comprises:

Determine the ambient light brightness according to the overall brightness and the current shooting parameters;

Determining target shooting parameters according to the ambient light brightness;

The current shooting parameter is adjusted according to the target shooting parameter.
The light metering method according to claim 15, wherein the shooting parameters include any one of the following: aperture value, shutter time, and sensitivity ISO.
The light metering method according to claim 1, wherein the overall brightness is used to compare with a preset reference brightness to output a corresponding exposure prompt.
The light metering method according to claim 18, wherein the exposure prompt comprises: an overexposure prompt and/or an underexposure prompt.
The light metering method according to claim 1, wherein the image to be metered is collected by an image sensor.
The photometry method according to claim 1, wherein the method is applied to a camera.
A camera, characterized by comprising: a body, a lens connected to the body, an image sensor, a processor, and a memory storing a computer program arranged in the body;

The image sensor is used to collect an image to be measured through the lens;

The processor implements the following steps when executing the computer program:

Acquiring, from the image sensor, brightness information corresponding to each of the plurality of pixels in the light image to be measured;

Determining weight information corresponding to each of the multiple pixels; wherein the weight information is determined according to the category of the image content corresponding to the pixel;

Calculate the overall brightness of the light image to be measured according to the brightness information and the weight information corresponding to each of the plurality of pixels.
The camera according to claim 22, wherein the category of the image content corresponding to the pixel is obtained by semantically segmenting the photometric image.
22. The camera of claim 23, wherein the processor is configured to input the photometric image into a semantic segmentation model when executing the step of semantically segmenting the photometric image to obtain the The first semantic map output by the semantic segmentation model; wherein, the first semantic map includes the category of the image content corresponding to the pixel.
The camera of claim 24, wherein the semantic segmentation model is a pre-trained convolutional neural network CNN model.
The camera of claim 24, wherein the image input to the semantic segmentation model is the down-sampled photometric image.
The camera according to claim 24, wherein the processor is further configured to down-sample the first semantic map before determining the weight information corresponding to the pixel through the first semantic map, Obtain the second semantic map of the preset size.
The camera of claim 27, wherein the preset size corresponds to the size of a predetermined metering weight matrix.
The camera according to claim 27, wherein when the processor performs the step of down-sampling the first semantic map, the processor is configured to divide the first semantic map according to the preset size Are multiple sub-regions; determine the category corresponding to each of the sub-regions; the category corresponding to the sub-region is the category with the highest proportion among the categories corresponding to each pixel in the sub-region; according to the category corresponding to each of the sub-regions, Generate the second semantic map.
The camera according to claim 22, wherein the weight information is determined according to a corresponding relationship between a category of image content corresponding to the pixel and a target, and the target corresponding relationship includes weight information corresponding to each category.
The camera of claim 30, wherein the target correspondence relationship is preset by the user or the system.
The camera according to claim 30, wherein, in the target correspondence relationship, the weight information corresponding to each category is determined according to the proportion of the image area corresponding to the category in the photometric image.
The camera of claim 32, wherein the image area corresponding to the category with the largest weight information occupies the largest proportion in the photometric image.
The camera according to claim 30, wherein, in the target correspondence relationship, the weight information corresponding to each category is determined according to the position of the image area corresponding to the category in the photometric image.
The camera of claim 34, wherein the image area corresponding to the category with the largest weight information is located in the center of the photometric image.
The camera of claim 22, wherein the overall brightness is used to guide the adjustment of shooting parameters.
The camera according to claim 36, wherein when the processor executes the step of adjusting the shooting parameters according to the overall brightness, the processor is configured to determine the brightness of the ambient light according to the overall brightness and the current shooting parameters; The ambient light brightness determines a target shooting parameter; and the current shooting parameter is adjusted according to the target shooting parameter.
The camera of claim 36, wherein the shooting parameters comprise any one of the following: aperture value, shutter time, and ISO sensitivity.
The camera of claim 22, wherein the overall brightness is used to compare with a preset reference brightness to output a corresponding exposure prompt.
The camera of claim 39, wherein the exposure prompt comprises: an overexposure prompt and/or an underexposure prompt.
An electronic device, characterized by comprising: a camera mounted on the electronic device, a processor, and a memory storing a computer program;

The processor implements the following steps when executing the computer program:

Acquiring, by the camera, brightness information corresponding to each of multiple pixels in the light image to be measured;

Determining weight information corresponding to each of the multiple pixels; wherein the weight information is determined according to the category of the image content corresponding to the pixel;

Calculate the overall brightness of the light image to be measured according to the brightness information and the weight information corresponding to each of the plurality of pixels.
The electronic device according to claim 41, wherein the category of the image content corresponding to the pixel is obtained by semantically segmenting the light image to be measured.
The electronic device according to claim 42, wherein when the processor executes the step of semantically segmenting the light image to be measured, it is used to input the light image to be measured into a semantic segmentation model to obtain the The first semantic map output by the semantic segmentation model; wherein, the first semantic map includes the category of the image content corresponding to the pixel.
The electronic device according to claim 43, wherein the semantic segmentation model is a pre-trained convolutional neural network CNN model.
The electronic device according to claim 43, wherein the image input to the semantic segmentation model is the down-sampled photometric image.
The electronic device according to claim 43, wherein the processor is further configured to down-sample the first semantic map before determining the weight information corresponding to the pixel through the first semantic map , To obtain the second semantic map of the preset size.
The electronic device according to claim 46, wherein the preset size corresponds to a predetermined size of the photometric weight matrix.
The electronic device according to claim 46, wherein, when the processor executes the step of down-sampling the first semantic map, it is configured to convert the first semantic map according to the preset size Divide into multiple sub-regions; determine the category corresponding to each of the sub-regions; the category corresponding to the sub-region is the category with the highest proportion among the categories corresponding to each pixel in the sub-region; according to the category corresponding to each of the sub-regions , Generating the second semantic map.
The electronic device according to claim 41, wherein the weight information is determined according to a corresponding relationship between a category of image content corresponding to the pixel and a target, and the target corresponding relationship includes weight information corresponding to each category.
The electronic device according to claim 49, wherein the target correspondence relationship is preset by a user or a system.
The electronic device according to claim 49, wherein, in the target correspondence relationship, the weight information corresponding to each category is determined according to the proportion of the image area corresponding to the category in the photometric image .
The electronic device according to claim 51, wherein the image area corresponding to the category with the largest weight information occupies the largest proportion in the photometric image.
The electronic device according to claim 49, wherein in the target correspondence relationship, the weight information corresponding to each category is determined according to the position of the image area corresponding to the category in the photometric image.
The electronic device according to claim 53, wherein the image area corresponding to the category with the largest weight information is located in the center of the photometric image.
The electronic device according to claim 41, wherein the overall brightness is used to guide the adjustment of shooting parameters.
The electronic device according to claim 55, wherein when the processor executes the step of adjusting the shooting parameters according to the overall brightness, the processor is configured to determine the brightness of the ambient light according to the overall brightness and the current shooting parameters; The target shooting parameter is determined according to the ambient light brightness; the current shooting parameter is adjusted according to the target shooting parameter.
The electronic device according to claim 55, wherein the shooting parameters comprise any one of the following: aperture value, shutter time, and sensitivity ISO.
The electronic device according to claim 41, wherein the overall brightness is used to compare with a preset reference brightness to output a corresponding exposure prompt.
The electronic device according to claim 58, wherein the exposure prompt comprises: an overexposure prompt and/or an underexposure prompt.
The electronic device according to claim 41, wherein the electronic device comprises a drone.
A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the photometry method according to any one of claims 1 to 21 is realized.