WO2016206344A1

WO2016206344A1 - White balance correction method, device and computer storage medium

Info

Publication number: WO2016206344A1
Application number: PCT/CN2015/100016
Authority: WO
Inventors: 王妮绒
Original assignee: 中兴通讯股份有限公司
Priority date: 2015-06-25
Filing date: 2015-12-30
Publication date: 2016-12-29
Also published as: CN106331661A

Abstract

A white balance correction method and device, the method comprising: acquiring reference color information of a reference image (S110); acquiring collected color information of a collected image (S120); determining a white balance correction parameter of the collected image according to the reference color information and the collected color information (S130); and performing a white balance correction on the collected image according to the white balance correction parameter (S140). Also provided is a computer storage medium.

Description

White balance processing method, device and computer storage medium

Technical field

The present invention relates to the field of information processing, and in particular, to a white balance processing method, apparatus, and computer storage medium.

Background technique

When using a mobile phone or a camera to take pictures, there are often reasons such as bad weather or poor light, which may cause poor white balance processing when taking pictures on the mobile phone or camera, and the picture is not ideal. Especially in different weather and light, the same image has a large color difference after the automatic white balance processing.

White balance (WB for short) is a process of removing abnormal colors. The human eye can naturally adjust the color of the object to be seen according to the current color temperature of the light source, while the camera equipment such as a digital camera is often difficult to achieve perfect automatic white balance. Just because the sensor does not have the color constancy at different illumination color temperatures of the human eye, the white balance module needs to restore the color of the object that appears to the human eye to make it appear white in the photo. To put it simply, it is the process of correcting the color cast of the image.

Therefore, in the prior art, a method for automatically adjusting white balance of an electronic device is proposed, which is an urgent problem to be solved.

Summary of the invention

In view of this, embodiments of the present invention are expected to provide a white balance processing method, apparatus, and computer storage medium for satisfying user image collection requirements.

The technical solution of the present invention is implemented as follows: The embodiment of the present invention provides a white balance processing method, where the method includes:

Obtaining reference color information of the reference image;

Obtaining the collected color information of the captured image;

Determining a white balance calibration parameter of the acquired image according to the reference color information and the collected color information;

Performing white balance calibration on the acquired image according to the white balance calibration parameter.

Based on the foregoing solution, the reference color information of the reference image is obtained by:

Matching a graphic object in the reference image with an acquisition object in the captured image;

And if the collection object matches the graphic object in the reference image, extracting color information of the graphic object as the reference color information.

Based on the above solution, the graphic object in the reference image includes a character object;

The matching the graphic object in the reference image with the collected object in the collected image comprises:

Determining whether the acquisition object is a graphic object in the reference image by using face recognition and/or graphic matching;

If the collection object is the graphic object, the graphic object is considered to match the collection object.

Based on the foregoing solution, the obtaining reference color information of the reference image includes:

Obtaining, by the reference image, a first mean value of three primary color components of the graphic object;

The acquiring color information of the collected image includes:

Obtaining a second mean value of three primary color components of the specified collection object in the acquired image;

Determining, according to the reference color information and the collected color information, a white balance calibration parameter of the collected image, including:

Obtaining gain coefficients of the three primary color components respectively according to the first mean value and the second mean value;

Obtaining the white balance calibration parameter according to the gain coefficient;

Wherein, the gain coefficient is the white balance calibration parameter.

Based on the foregoing solution, the acquiring the white balance calibration parameter according to the gain coefficient includes:

When the reference image is at least two, the white balance calibration parameter is obtained according to a preset function relationship and at least two gain coefficient calculation function values respectively formed according to the acquired image and each of the reference images.

Based on the foregoing solution, the method further includes:

The reference image is set according to a user instruction or according to a built-in indication before acquiring the reference color information of the reference image.

A second aspect of the embodiments of the present invention provides a white balance processing apparatus, where the apparatus includes:

a first acquiring unit configured to acquire reference color information of the reference image;

a second acquiring unit configured to acquire collected color information of the collected image;

a determining unit, configured to determine a white balance calibration parameter of the collected image according to the reference color information and the collected color information;

And a calibration unit configured to perform white balance calibration on the acquired image according to the white balance calibration parameter.

The first acquiring unit is configured to match the graphic object in the reference image with the collected object in the collected image; if the collected object matches the graphic object in the reference image, Then, color information of the graphic object is extracted as the reference color information.

The first acquiring unit is configured to determine, by using face recognition and/or graphic matching, whether the collected object is a graphic object in the reference image; if the collected object is the graphic object, the graphic is considered The object matches the collection object.

The first obtaining unit is configured to acquire a first mean value of three primary color components of the reference image specifying graphic object, according to the foregoing solution;

The second acquiring unit is configured to acquire a second average value of three primary color components of the specified collection object in the collected image;

The determining unit is configured to obtain gain coefficients of the three primary color components respectively according to the first mean value and the second mean value; and obtain the white balance calibration parameter according to the gain coefficient;

Wherein, the gain coefficient is the white balance calibration parameter.

The determining unit is configured to: when the reference image is at least two, according to a preset function relationship and at least two gain coefficient calculation functions respectively formed according to the collected image and each of the reference images Value, the white balance calibration parameter is obtained.

Based on the above solution, the device further includes:

And a setting unit configured to set the reference image according to a user indication or according to a built-in indication before acquiring reference color information of the reference image.

A third aspect of the embodiments of the present invention provides a computer storage medium, where the computer storage medium stores computer executable instructions, and the computer executable instructions are used to execute at least one of the foregoing white balance processing methods.

The white balance processing method, device and computer storage medium according to the embodiment of the present invention, when performing white balance processing, enter a reference image to perform white balance processing on the collected object, and after the collected image is subjected to white balance processing, It can be similar to the white balance processing of the reference image, thus providing a new method of white balance correction. If the white balance of the reference image is usually an image formed by white balance that satisfies the user's needs, then a simple implementation of white balance processing that satisfies the user's needs is achieved.

DRAWINGS

1 is a schematic flowchart of a white balance processing method according to an embodiment of the present invention;

2 is a second schematic flowchart of a white balance processing method according to an embodiment of the present invention;

3 is a schematic structural diagram of a white balance processing apparatus according to an embodiment of the present invention;

4 is a third schematic flowchart of a white balance processing method according to an embodiment of the present invention;

FIG. 5 is a second schematic structural diagram of a white balance processing apparatus according to an embodiment of the present invention; FIG.

FIG. 6 is a schematic diagram of a reference image according to an embodiment of the present invention; FIG.

FIG. 7 is a schematic diagram of an original image according to an embodiment of the present invention.

detailed description

The present invention will be further described in detail with reference to the accompanying drawings and specific embodiments.

Method embodiment:

As shown in FIG. 1 , this embodiment provides a white balance processing method, where the method includes:

Step S110: acquiring reference color information of the reference image;

Step S120: Acquire acquisition color information of the collected image;

Step S130: determining, according to the reference color information and the collected color information, a white balance calibration parameter of the collected image;

Step S140: Perform white balance calibration on the acquired image according to the white balance calibration parameter.

The reference image in this embodiment may be an image predetermined in advance, such as a user-specified image. The reference image may be an image that may include the same graphical object as the captured image. The reference image is a completely different graphic from the graphic object included in the captured image. In the embodiment, the reference image and the captured image are images formed by the same person or the scene, and may also be images collected by different characters or scenes. For example, the reference image is an image formed by image acquisition by the user A, and the captured image is an image formed by image capture of the user B, but the user B feels the effect of white balance processing on the image formed by the user A. Preferably, an image formed by the user A is designated as the reference image. Of course, the captured image formed by the user A is currently collected by the user as the reference image. Obtaining the reference color information of the reference image in the embodiment may include parsing the reference image, and acquiring color component values of each color in the reference image, for example, acquiring red, The average color component value of the three colors of green and blue. Of course, the reference color information is obtained in step S110 in this embodiment, and the reference color information of the reference image may be received from other devices. For example, the user may store the reference image in the cloud platform in order to reduce the information storage in the local device, and the step S110 may include acquiring the reference color information from the cloud platform.

In step S120, the collected color information in the acquired image is obtained, that is, the color component value corresponding to the reference image in the currently acquired image.

The reference color information can reflect the white balance in the reference image. The color information collected can reflect the white balance in the captured image.

In step S130 in this embodiment, the white balance calibration parameter of the adjusted image is determined according to the reference color information and the collected color information. The white balance calibration parameter is used to make the difference between the white balance parameter of the acquired image and the white balance parameter of the reference image within a specified threshold range, so that the captured image can be made to the captured image through step S140. After the white balance calibration, the white balance of the acquired image is similar to the white balance of the reference image (ie, within the specified threshold range), which results in an ideal image acquisition effect. It also avoids the problem of color shift or chromatic aberration caused by different white balance of the acquired images obtained by the same acquisition object in different ambient light, thereby improving the image collection effect and user satisfaction.

The step S110 may include:

The graphic objects in the reference image may include a person object and a scene object. Specifically, the character object may include a user's avatar, a user's face, a portrait of the user's child, and the like. The scene object may include a scene object such as a building or a plant such as a flower or a grass. Of course in the captured image The collection objects can also be divided into person objects and scene objects.

In the embodiment, in order to provide accurate white balance calibration parameters, when extracting the reference color information, color information of the collection object and the graphic object are first extracted separately. In a specific implementation process, if the processing rate is increased, the reference color information of the graphic object in the reference object may also be extracted and stored in advance. In this way, when the user performs image collection by using an electronic device such as a camera or a mobile phone, the reference color information may be directly extracted from the pre-stored data in step S110 to improve the response rate.

The matching of the graphic object and the acquisition object will also be performed in step S110. Regardless of the collection object and the graphic object, the shape information is included in addition to the color information. In the present embodiment, the matching in step S110 is to match the shape information of the collection object and the graphic object. The shape information herein may specifically include various contour-related information such as contour information, a spacing between respective shapes within the contour, and the like. Specifically, the face object in the matched character object may include an outline matching of the face facial features, and matching between any two organ spacings in the facial features. By matching the shape information of the graphic object and the collected object in the embodiment, if it is determined that the captured object in the captured image and the graphic object in the reference image originate from the same person or the scene, the collected object and the graphic object are considered to match.

If it matches, the color information of the graphic object in the reference image is extracted as the reference color information for determining the white balance.

In this case, if the user likes to take a selfie or the mother takes a picture of the baby. The user's own image or an image of the baby may be included in the reference image. In this case, when performing the white balance calibration of the current captured image, the previously acquired image can be used as the reference image for white balance calibration, so that the photos obtained by the user under different scenes and different ambient illuminations are not well The calibration of the white balance leads to problems such as chromatic aberration and color shift, thereby improving the intelligence of the electronic device, the image collection effect, and the user satisfaction.

The graphic object in the reference image includes a character object;

The matching the graphic object in the reference image with the collected object in the collected image comprises: determining whether the collected object is a graphic object in the reference image by using face recognition and/or graphic matching; If the collection object is the graphic object, the graphic object is considered to match the collection object. In this embodiment, the face recognition feature is used to extract the reference image and the facial features of the person object in the captured image, and the similarity of the facial features is determined by feature matching. If the similarity is greater than the specified matching threshold, the corresponding feature is considered to be matched. Then, using the principle of biostatistics to statistically process whether multiple facial features match, etc., determine whether the collected object is a graphic object. In this way, the face recognition technology can quickly realize the matching and processing of the graphic object and the collected object, and is especially suitable for the white balance calibration of the human shooting.

Of course, in a specific implementation process, the reference image object is extracted and processed in advance, and the facial features are extracted to form a feature template. When the comparison is performed, the facial features are not re-extracted from the graphic object, and the time can be reduced. Delay. And a reference image may be used in multiple image acquisition processes afterwards, so that the method of forming the feature template can reduce the amount of data processing of the device.

In this embodiment, the mobile phone or the camera is determined by itself to determine the graphic object in the reference image that matches the collected object, and may also be determined based on the user input during the specific implementation process. For example, the user selects a specified graphic object pre-stored in a mobile phone or a camera or a specified image acquired by a current mobile phone or a camera from another electronic device by means of touch or hovering touch, etc., in such a manner, allowing the user Manual input specification is added, which increases user control.

The step S110 may include: acquiring a first mean value of the three primary color components of the reference image specifying graphic object.

The step S120 may include: acquiring a second mean value of the three primary color components of the specified collection object in the acquired image.

The step S130 may include: obtaining gain coefficients of the three primary color components respectively according to the first mean value and the second mean value; and acquiring the white balance calibration parameter according to the gain coefficient;

Wherein, the gain coefficient is the white balance calibration parameter.

The three primary colors described in this embodiment may include red, green, and blue. The various color information of the character and the scene can be combined by the three colors. Therefore, when extracting the reference color information and the collected color information in the embodiment, the component values of the three primary colors can be extracted. Obtain.

The gain coefficient can be obtained by comparing the reference image with the corresponding three primary colors in the acquired image. The gain factor or the inverse of the gain factor will be used directly as the white balance calibration parameter to calibrate the white balance of the acquired image.

Obtaining the white balance calibration parameters in this way has the characteristics of being simple and quick to implement.

Obtaining the white balance calibration parameter according to the gain coefficient, including:

The preset function relationship in this embodiment may be an average function relationship of at least two of the gain coefficients. Thus the function value is the mean of at least two of the gain coefficients.

Of course, in a specific implementation process, weights may also be set for the gain coefficients corresponding to each reference image, such that the functional relationship may be the mean of the sum of the products of the corresponding weights multiplied by the respective weight coefficients.

The weight here may be set for the user for each of the reference images, or may be intelligently allocated according to the time formed by the reference image by a device such as a mobile phone or a camera. Generally, the closer the time the reference image is formed in the device to the current time, the larger the weight can be allocated, and the longer the current time is, the smaller the weight is assigned. Thus, when a plurality of reference images are stored in the mobile phone and the camera, the white balance of the currently formed captured image can be made closer to the white balance of the newly formed reference image. This way of intelligently assigning weights, considering the continuity of time, considering the changes in the characters and scenes over time, so that one can It is convenient to accurately calibrate the white balance, and on the other hand, it can accurately reflect the variability of the collected objects, further improving the intelligence of the electronic device and the satisfaction of the user.

As shown in FIG. 2, the method further includes:

Step S100: Before acquiring the reference color information of the reference image, setting the reference image according to a user instruction or according to a built-in indication.

In the embodiment, the user indicates that the user may indicate that an image in the sub-mobile phone or the camera has been stored as a reference image by means of touch or hovering touch, keyboard, mouse, or voice. The built-in instruction sets a corresponding built-in instruction to intelligently determine the reference image before the mobile phone or the camera is shipped from the factory. The images taken by themselves are stored in an image database, and all images in the image database are taken as the reference images. It is of course also possible to use the acquired image within the most recent specified period of time as the reference image. In summary, the present embodiment provides two methods for setting the reference image, which are specific to implementation and simple to implement.

Generally, the step S100 described in this embodiment only needs to ensure that it is performed before step S110, but in order to further improve the response speed, before the image acquisition before or before the mobile phone or camera image acquisition application is closed, the automatic or The corresponding captured image is set to the corresponding reference image based on the user indication, so that the next time the image is captured, the appropriate reference image can be directly found from the determined reference image, and the white balance calibration can be performed.

Equipment embodiment:

As shown in FIG. 3, this embodiment provides a white balance processing apparatus, where the apparatus includes:

The first obtaining unit 110 is configured to acquire reference color information of the reference image;

The second obtaining unit 120 is configured to acquire the collected color information of the collected image.

The determining unit 130 is configured to determine a white balance calibration parameter of the collected image according to the reference color information and the collected color information;

The calibration unit 140 is configured to perform white balance calibration on the acquired image according to the white balance calibration parameter.

The first obtaining unit 110 in this embodiment may include a communication interface, and the communication interface may be a wireless interface or a wired interface, and configured to receive the reference image information from other electronic devices. The specific structure of the first obtaining unit 110 may further include a structure having an information processing structure and a processing circuit, which can be used to extract the reference color information from its reference image, or query the reference color information that has been extracted in advance, and the like. . The processor may include an application processor AP, a digital signal processor DSP, a central processing unit CPU, a microprocessor MCU, or a programmable circuit PLC. The processing circuit can include a structure such as an application specific integrated circuit ASIC.

The specific structure of the second obtaining unit 120, the determining unit 130, and the calibration unit 140 includes the processor or the processing circuit, and the processor or the processing circuit can perform the obtaining by performing the execution of the executable code. Collect color information, white balance calibration parameter determination, and white balance calibration.

These units may correspond to one processor or processing circuit, or may correspond to different processors or processing circuits, respectively. When any two of the foregoing unit integrations correspond to the same processor or processing circuit, the processor or processing circuit performs a corresponding operation by means of time division multiplexing or concurrent threads.

The graphic objects in the reference image in this embodiment include a character object and a scene object. The reference image and the acquired image related definition can be referred to the foregoing method embodiment, and will not be repeated here.

The white balance processing device according to the embodiment may be a device capable of performing white balance processing such as a mobile phone, a camera, or a tablet, and is further preferably an electronic device further including an image acquisition unit. The image acquisition unit may include a structure capable of image and/or video capture, such as a camera.

In summary, the embodiment provides a white balance processing device capable of performing white balance processing according to a reference image, so that the acquired image does not use different white balances due to different collection environments, resulting in collection of the same person or scene. This leads to problems such as large color shift, which improves the effect of image acquisition, the intelligence of electronic devices, and user satisfaction.

The first acquiring unit 110 is configured to match the graphic object in the reference image with the collected object in the collected image; if the collected object matches the graphic object in the reference image, the extraction device The color information of the graphic object is used as the reference color information.

In the embodiment, when the first acquiring unit 110 extracts the reference color information, in order to improve the accuracy of the white balance calibration, in the embodiment, the graphic object in the reference image including the matching image is to be included. In this case, it can ensure that the white balance calibration is the white balance that the user wants to use, so that the acquisition and acquisition images obtained by the user can be used to improve the user satisfaction.

The graphic object in the reference image includes a character object;

The first acquiring unit 110 is configured to determine, by using face recognition and/or graphic matching, whether the collected object is a graphic object in the reference image; if the collected object is the graphic object, the A graphical object matches the captured object.

The specific structure of the first acquiring unit 110 in this embodiment may include an information processor or a processing circuit having a face recognition technology, and the graphic object corresponding to the collected object in the collected image is determined by the face recognition technology. It has the characteristics of simple structure and simple implementation.

The first obtaining unit 110 is configured to acquire a first average value of three primary color components of the reference image specifying graphic object;

The second acquiring unit 120 is configured to acquire a second average value of the three primary color components of the specified collection object in the collected image;

The determining unit 130 is configured to obtain a gain coefficient of the three primary color components according to the first mean value and the second mean value, and obtain the white balance calibration parameter according to the gain coefficient;

Wherein, the gain coefficient is the white balance calibration parameter.

In the embodiment, the first obtaining unit 110 may include a counter and a calculator, and the counter may be used to count the number of pixels corresponding to the specified graphic in the reference image; the calculator may be used to calculate based on the calculator. The number of pixels determines the three primary color components of the specified graphic object The first mean.

The first mean value herein is used to refer to the mean values corresponding to the three primary colors, such as the mean of red, the mean of blue, and the mean of green.

The second acquiring unit may also include a counter and a calculator to obtain a second average value of the specified object in the captured image. Similarly, the second mean value here is used to refer to the mean values corresponding to the three primary colors respectively.

The determining unit 130 may include a computing structure such as a multiplier, and the white balance calibration parameter is determined by calculating the gain coefficient.

In summary, the embodiment provides a structure for determining the white balance calibration parameter, which has the characteristics of simple structure and simple implementation, and can be used to implement the white balance processing method described in the method embodiment.

The determining unit 130 is configured to: when the reference image is at least two, obtain a function value according to a preset function relationship and at least two gain coefficients formed according to the collected image and each of the reference images respectively, The white balance calibration parameter. The preset function relationship herein may include an average function relationship, and the specifically calculated function value and the white balance calibration parameter may be an average of the gain coefficients.

Of course, in the specific implementation process, the function relationship also provides different weights to the gain coefficients corresponding to different reference images, so that the white balance of the acquired image obtained by the device according to the embodiment and at least two reference images are obtained. One of the reference images in the picture is closer. In this case, it is obvious that the captured image of the user's specified demand can be provided. If the weight corresponding to the reference image is determined according to the acquisition time formed by the reference image acquisition, and the reference image is formed near the current time, and may correspond to a higher weight, the white balance calibration can be easily realized. At the same time, it is also possible to save as much as possible the color information in the captured image that truly reflects the collected person or scene, and improve the fidelity of the information.

The device further includes:

a setting unit configured to use the reference color information of the reference image before The reference image is set by the user or according to the built-in indication.

The setting unit described in this embodiment may include a storage medium and a processor or a processing circuit, and the storage medium is configured to store the reference image or the identification information of the reference image to facilitate subsequent determination of the reference image. The processor or processing circuitry can be operative to perform a corresponding operation in accordance with the user indication or built-in indication.

When the setting unit is specifically configured to set a reference image according to a user instruction, the setting unit may include a human-computer interaction interface, such as a touch screen, a floating touch screen or a keyboard, a mouse, a voice input interface, and the like. structure. These structures can all receive the user indication.

In summary, the apparatus described in this embodiment provides a hardware implementation method for implementing the white balance processing method described in the method embodiment, and has the characteristics of simple structure and white balance calibration to meet user requirements, and improves user satisfaction.

Several specific examples are provided below in connection with any of the above embodiments.

Example 1:

As shown in FIG. 4, the method for performing precise white balance processing on an image implemented by the following technical solution is divided into the following steps:

Step S1: Reference image setting.

Step S2: Acquiring an original image. Specifically, as shown in FIG. 4, an image is acquired by optical sensing, where the photosensor may include a complementary metal oxide semiconductor (CMOS) or an electrical coupling element (Charge- Coupled Device, CCD) sensor and other structures.

Step S3: face recognition and/or image matching processing. Specifically, the original image is processed by the image signal processor ISP, so that the matching of the captured object in the captured image and the graphic object in the reference image can be determined.

Step S4: It is judged whether the original image and the reference image have a match. The original image here is the acquired image in the foregoing embodiment. If yes, go to step S5, if no, go to step S8.

Step S5: Calculate the gain coefficient according to the matching.

Step S6: performing white balance correction using the gain coefficient.

Step S7: The corrected image display is adjusted.

Step S8: Calling other white balance processing to perform white balance processing of the original image. Other white balance processing herein may include white balance processing such as grayscale world algorithm, Weng algorithm, and perfect reflection algorithm. This and other white balance algorithms can be seen in the prior art and will not be listed here.

In step S1: some object images having a specific color are set as reference images. Among them, these reference images are often photographed in normal photos. For example, setting your own face image can meet the needs of frequent self-portraits; setting the child's face image can meet the needs of the photo shooter; and setting the green leaf image can meet the needs of the love scene; You can set face images of family and friends, and more. These reference images can be divided into two groups according to the image content, a face image group and other groups. These reference images can also be set as much as possible according to the specific personal photographing situation; as shown in the following table.

参考图像组1Reference image group 1	描述description
图像1Image 1	自拍Selfie
图像2Image 2	宝宝baby
图像3Image 3	另一半The other half
图像4Image 4	老妈Mom
图像5Image 5	老爸Dad
图像6Image 6	闺蜜Girlfriend
图像7Image 7	宝宝玩伴Baby playmate
……......	……......

......

参考图像组2Reference image group 2	描述description
图像1Image 1	绿萝Green radish
图像2Image 2	雪snow
图像3Image 3	爱车Car
图像4Image 4	黑板/白板Blackboard/whiteboard
图像5Image 5	关闭电源的显示器Power off display
图像6Image 6	天安门Tiananmen Square
图像7Image 7	人民币Renminbi
……......	\|……\|......

Step S2 may specifically include: acquiring an original image by using a CMOS or CCD sensor, and calculating an average value of three color components of the original image R, G, and B by the following method:

It is worth noting that in this example red is indicated by the letter R, green is indicated by the letter G, and blue is indicated by the letter B.

In the above formula, N is the total number of pixels of the original image, and Ri, Gi, and Bi are the three colors of red, green, and blue of the i-th pixel, respectively.

Step S3: The image signal processor processes the original image obtained by the CMOS or CCD sensor. Through the face recognition technology, all the images in the original image and the reference image group 1 are sequentially subjected to face recognition, and it is judged whether the original image and the reference image can be matched. When it is judged that the original image and the reference image group 1 have a match, the number x of each matching image is recorded, and the average value of the red, green and blue components of the matched reference image is calculated by the following method.

In the above formula, n is the total number of pixels of the reference image group 1 that match the original image, and Ri, Gi, and Bi are respectively the red, green, and blue color components of the ith pixel.

Similarly, all the images in the original image and the reference image group 2 are sequentially image-matched, and it is judged whether or not the original image and the reference image can be matched. When it is judged that there is a match between the original image and the reference image group 2, the number of matching images y is recorded, and the average value of the red, green and blue components of the matched reference image is calculated by the following method.

In the above formula, m is the total number of pixels of the reference image group 2 that match the original image, and Rj, Gj, and Bj are respectively the red, green, and blue color components of the jth pixel.

The step S4 may specifically include: analyzing the matching situation of the original image and the reference image, and the gain coefficients of the original image and the reference image in the three colors of the red, green, and blue components are as follows:

First: When the original image only matches the reference image group 1, the gain coefficients of the original image and the reference image in the three colors of red, green and blue are:

When the original image matches one of the reference image groups 1 (x=1),

Then Kr=Krx=Kr1, Kg=Kgx=Kg1, Kb=Kbx=Kb1;

When the original image matches 2 images of the reference image group 1 (x=2),

Then there is

When the original image matches 3 images of the reference image group 1 (x=3),

Then there is

And so on, when the original image matches the reference image group 1 with x images,

Then there is

Second:

When the original image only matches the reference image group 2, the gain coefficients of the original image and the reference image in the three colors of red, green and blue are:

When the original image matches one of the reference image groups 2 (y=1),

Then Kr=Krx=Kr1, Kg=Kgx=Kg1, Kb=Kbx=Kb1;

When the original image matches 2 images of the reference image group 2 (y=2),

Then there is

When the original image matches 3 images of the reference image group 2 (y=3),

Then there is

And so on, when the original image has y images matching the reference image group 2, then there is

The third type:

When the original image matches the reference image group 1 and the reference image group 2, the gain coefficients of the original image and the reference image in the three colors of red, green, and blue are:

Fourth: When the original image does not match the reference image group 1 and the reference image group 2, other white balance processing methods are called for processing.

Step S5: Perform white balance correction on the gain values of the red, green and blue components of the original image and the reference image analyzed according to the step (4).

That is:

Ri’=Kr*Ri, Gi’=Kg*Gi, Bi’=Kb*Bi

In the above formula, Ri, Gi, and Bi are the red, green, and blue components of the ith pixel in the original image, respectively, and Ri', Gi', and Bi' are respectively red, green, and blue of the corrected i-th pixel. Three color components.

Step (6): Adjust each pixel Ri', Gi', Bi' of the image to within the displayable range [0, 255]. For example, the maximum value of the 24-bit true color map is the maximum of all the Ri', Gi', and Bi' components in the image.

If adj>1, then red, green, and blue components are re-adjusted for each pixel in the image,

Example two:

As shown in FIG. 5, the white balance processing apparatus provided in this example includes a reference image to the setting module 11, the original image obtaining module 12, the identification matching module 13, the gain calculating module 14, and the white balance correcting module 15 in the apparatus. . The reference image setting module 11 in this embodiment is equivalent to the setting module in the foregoing embodiment. The original image acquisition module 12 is equivalent to an image acquisition unit for acquiring a captured image to acquire other structures. The identification matching module 13 and the gain calculation module 14 correspond to the determination unit in the foregoing embodiment. The white balance correction module 15 may be the structure of the calibration unit 140 in the foregoing embodiment.

The reference image setting module 11 is configured to set an image of an object having a specific color as a reference image after the white balance function is turned on, and is used as a basis for performing white balance contrast adjustment with the original image.

The original image obtaining module 12 is configured to acquire an original image by using a CMOS or CCD sensor, and calculate an average value of three colors of the original image, such as red, green, and blue.

The identification matching module 13 is configured to sequentially identify and match the original image and all the images in the reference image group 1/group 2 by face recognition technology and/or image matching technology. When it is judged that the original image and the reference image have a match, the average value of the three colors of the red, green and blue of the reference image that can be matched is calculated.

The gain calculation module 14 calculates the gain coefficients of the original color of the original image and the reference image in the red, green and blue components according to the analysis of the matching conditions of the original image and the reference image.

The white balance correction module 15 is configured to perform white balance correction according to the gain values of the red, green, and blue components of the original image and the reference image analyzed in the gain calculation module.

The identification matching module 13 may specifically include:

Face recognition sub-module: used to sequentially perform face recognition on all images in the original image and the reference image group 1, and determine the matching of the original image and the reference image. The face recognition uses computer image processing technology to extract portrait feature points from images or videos, and uses the principle of biostatistics to analyze and establish a mathematical model, that is, a face feature template. Utilize the built face feature template A feature analysis is performed on the face of the person to be tested, and a similar value is given based on the result of the analysis. Use this value to determine if it is the same person. Face recognition technology consists of three parts: face detection, face tracking, and face contrast. The greatest advantage of face recognition is its convenience, speed, and non-intrusiveness. Face recognition does not need to interfere with people's behavior to achieve recognition, without having to argue for whether they are willing to put their hands on the fingerprint collection device, or speak into the microphone, or point their eyes at the laser scanning device. You just have to walk past a camera or camera and you have been tested quickly.

The image matching sub-module is configured to sequentially perform image matching on all the images in the original image and the reference image group 2, and determine the matching of the original image and the reference image. The process of image matching is to spatially align all or part of the known image with the unfamiliar image, and find the sub-image corresponding to the pattern in a strange image according to the image of the known pattern. Image matching can be mainly divided into gray-based matching and feature-based matching. Image matching technology has become an extremely important technology in the field of image information processing, and is widely used in many fields such as image recognition, image analysis and computer vision.

The gain calculation module 14 includes:

A classification judgment sub-module for classifying according to matching of different objects.

The gain sub-module may include a calculator for calculating a gain coefficient.

Example three:

This example provides an automatic white balance processing method for an image of a mobile phone camera. When the automatic precision white balance function is turned on when the camera image is taken, the related reference image setting is performed. The set reference images are divided into two groups, group 1 is a face reference image, and group 2 is another image. It is assumed that one of the reference images set is the face image shown in FIG. 6 below. The area enclosed by the dotted line in Fig. 6 is the graphic object in the reference.

Taking a picture with a mobile phone, the original image obtained is shown in Figure 7 below, and the average of the red, green and blue components of the original image in Figure 7 below is obtained.

Statistics, that is, the original image is counted according to the red, green, and blue color storage modes of the image.

It should be noted that the original image described in this example is the captured image described in the foregoing embodiment, and the face of the picture is the collected object.

The original image and the images in the reference image group 1 and the group 2 are sequentially identified and matched, respectively. When it is judged that the original image and the reference image have a match, the average values of the red, green and blue components of the reference image that can be matched are calculated.

In this embodiment, the original image has a matching with the above-mentioned 1 human face image in the reference image group 1, then x=1, and the average value of the red, green and blue components of the reference image matching the original image can be counted.

Then the gain coefficients of the original image and the reference image in the three colors of red, green and blue are:

Then, the white balance correction is performed according to the gain values of the red, green, and blue components of the original image and the reference image calculated as described above. That is:

Ri’=Kr1*Ri=1.27*Ri

Gi’=Kg1*Gi=0.95*Gi

Bi'=Kb1*Bi=0.84*Bi

In the above formula, Ri, Gi, and Bi are the red, green, and blue components of the ith pixel in the original image, respectively, and Ri', Gi', and Bi' are respectively the red of the ith pixel after the automatic precision white balance correction. , green, blue three color components. That is, when the i-th pixel point Ri, Gi, and Bi three-color components (Ri, Gi, Bi) = (108, 155, 130) in the original image, the pixel is automatically corrected by the white balance corrected red, green, The blue three-color component is (Ri, Gi, Bi) = (Kr1*Ri, Kg1*Gi, Kb1*Bi) = (137, 147, 109). Each pixel in the original image is corrected by analogy. If the pixel in the image exceeds the displayable range [0, 255], the pixel is quantized to between the displayable range [0, 255]. The above process completes the automatic precise white balance processing of the original image.

The embodiment of the invention further provides a computer storage medium, wherein the computer storage medium stores computer executable instructions, and the computer executable instructions are used in at least one of the white balance processing methods provided by the foregoing embodiments; for example, At least one of the methods shown in FIGS. 1, 2, 4, and 5 is performed.

The computer storage medium described in this embodiment may be various types of storage media such as an optical disk, a hard disk, or a magnetic disk, and may be a non-transitory storage medium.

In the several embodiments provided by the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The device embodiments described above are merely illustrative. For example, the division of the unit is only a logical function division. In actual implementation, there may be another division manner, such as: multiple units or components may be combined, or Can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, or direct coupling, or communication connection of the components shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, and may be electrical, mechanical or other forms. of.

The units described above as separate components may or may not be physically separated, and the components displayed as the unit may or may not be physical units, that is, may be located in one place or distributed to multiple network units; Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing module, or each unit may be separately used as one unit, or two or more units may be integrated into one unit; the above integration The unit can be implemented in the form of hardware. It can also be implemented in the form of hardware plus software functional units.

A person skilled in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by using hardware related to the program instructions. The foregoing program may be stored in a computer readable storage medium, and the program is executed when executed. The foregoing storage device includes the following steps: the foregoing storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk. A medium that can store program code.

The above is only the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and modifications made in accordance with the principles of the present invention should be understood as falling within the scope of the present invention.

Claims

A white balance processing method, the method comprising:

Obtaining reference color information of the reference image;

Obtaining the collected color information of the captured image;

Determining a white balance calibration parameter of the acquired image according to the reference color information and the collected color information;

Performing white balance calibration on the acquired image according to the white balance calibration parameter.
The method of claim 1 wherein

The reference color information of the reference image is obtained by:

Matching a graphic object in the reference image with an acquisition object in the captured image;

And if the collection object matches the graphic object in the reference image, extracting color information of the graphic object as the reference color information.
The method of claim 2, wherein

The graphic object in the reference image includes a character object;

The matching the graphic object in the reference image with the collected object in the collected image comprises:

Determining whether the acquisition object is a graphic object in the reference image by using face recognition and/or graphic matching;

If the collection object is the graphic object, the graphic object is considered to match the collection object.
The method of claim 1 wherein

The obtaining reference color information of the reference image includes:

Obtaining, by the reference image, a first mean value of three primary color components of the graphic object;

The acquiring color information of the collected image includes:

Obtaining a second mean value of three primary color components of the specified collection object in the acquired image;

Determining, according to the reference color information and the collected color information, a white balance calibration parameter of the collected image, including:

Obtaining gain coefficients of the three primary color components respectively according to the first mean value and the second mean value;

Obtaining the white balance calibration parameter according to the gain coefficient;

Wherein, the gain coefficient is the white balance calibration parameter.
The method of claim 4, wherein

Obtaining the white balance calibration parameter according to the gain coefficient, including:

When the reference image is at least two, the white balance calibration parameter is obtained according to a preset function relationship and at least two gain coefficient calculation function values respectively formed according to the acquired image and each of the reference images.
The method of claim 1 wherein

The method further includes:

The reference image is set according to a user instruction or according to a built-in indication before acquiring the reference color information of the reference image.
A white balance processing device, the device comprising:

a first acquiring unit configured to acquire reference color information of the reference image;

a second acquiring unit configured to acquire collected color information of the collected image;

a determining unit, configured to determine a white balance calibration parameter of the collected image according to the reference color information and the collected color information;

And a calibration unit configured to perform white balance calibration on the acquired image according to the white balance calibration parameter.
The apparatus according to claim 7, wherein

The first acquiring unit is configured to map the graphic object in the reference image with the collecting The collected objects in the image are matched; if the collected objects match the graphic objects in the reference image, the color information of the graphic objects is extracted as the reference color information.
The device according to claim 8, wherein

The graphic object in the reference image includes a character object;

The first acquiring unit is configured to determine, by using face recognition and/or graphic matching, whether the collected object is a graphic object in the reference image; if the collected object is the graphic object, the graphic is considered The object matches the collection object.
The apparatus according to claim 7, wherein

The first acquiring unit is configured to acquire a first average value of three primary color components of the reference image specifying graphic object;

The second acquiring unit is configured to acquire a second average value of three primary color components of the specified collection object in the collected image;

The determining unit is configured to obtain gain coefficients of the three primary color components respectively according to the first mean value and the second mean value; and obtain the white balance calibration parameter according to the gain coefficient;

Wherein, the gain coefficient is the white balance calibration parameter.
The device according to claim 10, wherein

The determining unit is configured to: when the reference image is at least two, calculate a function value according to a preset function relationship and at least two gain coefficients formed according to the collected image and each of the reference images respectively, and obtain a Describe the white balance calibration parameters.
The apparatus according to claim 7, wherein

The device further includes:

And a setting unit configured to set the reference image according to a user indication or according to a built-in indication before acquiring reference color information of the reference image.
A computer storage medium having stored therein computer executable instructions for performing the white balance processing of claims 1 to 6 At least one of the methods.