CN103327345A - White balance processing method and processing device thereof - Google Patents

Abstract

The invention provides a white balance processing method and a processing device thereof, comprising the following steps: after an input image is divided into a plurality of divided block images, a white point identification is carried out on each divided block image to generate a white point identification result. Then, according to the position of each of the divided block images in the input image and the continuity of the plurality of divided block images around the input image, a weight of the white point identification result corresponding to each of the divided block images is determined. And calculating a white balance estimated value of the input image according to the white point identification results and the weights of the segmented block images, and performing white balance processing on the input image according to the white point estimation value. The invention carries out white balance processing by adding the weight analysis of the input image light source types, the positions of the segmentation block images and the continuity calculation of the segmentation block images and the surrounding segmentation blocks, and can accurately obtain the color of the input image on the premise of highlighting the color of a specific scene.

Description

White balance processing method and processing device thereof

technical field

The present invention relates to a white balance processing method and a processing device thereof, in particular to a method for dividing an image into a plurality of divided blocks, and analyzing the image data of each divided block individually, so as to reduce white point errors in the image scene. A white balance processing method and processing device for judging situations.

Background technique

With the successive introduction of digital cameras, image sensors such as charge-coupled devices (Charge-coupled devices, CCD) or complementary metal-oxide-semiconductor sensors (CMOS sensors) are increasingly important, because image sensors cannot distinguish The color of ambient light, so digital cameras usually need to perform white balance processing.

Generally speaking, the color of the light reflected by an object depends on the color of the light source. The human brain can detect and correct such color changes. Therefore, no matter in sunlight, hazy weather, indoor lighting or fluorescent light, the white objects people see are still the same. ; But because the sensing device cannot distinguish the color of the external ambient light, the white objects seen by the sensing device have different colors under different light sources, some are yellowish or red, and some are blue. In order to be close to human vision, digital cameras must adjust the color according to the light source, so that the final image can present the white color seen by human eyes. This adjustment is called white balance.

In the traditional white balance technology, facing a composition with a blue sky background, because all input images are included in the white balance statistical range, a large area of blue image data with high color temperature is used for white balance statistical analysis data, so that the image can be corrected. After balance correction, the corrected image will have a lower color temperature than the original input image.

Therefore, the known technology proposes to divide the input image into multiple frames, then conduct white point identification on the image data in each segmented frame, and evaluate the white point detection range of high color temperature pixels for the image data located above the position of the input image. Higher limit, exclude high color temperature pixels into the statistical analysis of white point detection, to avoid misjudgment of scenes with large blue sky when performing white balance, such white balance results can show the user's desire to highlight colored blue.

However, the white balance analysis and correction method of the segmented image in the prior art only emphasizes the up-down relationship based on the position of the segmented block, and uses the color temperature value as the limit of the white point detection range, but ignores the overall composition content and image For further analysis of the light source, that is, when the upper block of the image is mostly a high color temperature image, it is still necessary to accurately judge whether the scene is an outdoor blue sky scene based on the continuity of the images of each segmented block, so as to avoid multiple high color temperature images in the input image. When the color temperature object is located above the image, misjudgment of the image content of the segmented block occurs, which leads to a deviation in the white balance correction result.

In addition, the reason for the previous case is to limit the white point detection range of the image data within the segmented range based on the position of each segmented block. If there is no white point information in the segmented block for white point identification, it is also It will cause deviation to the white balance correction result.

In view of this, the present invention addresses the above problems, and proposes a white balance processing method after dividing the captured image into multiple frames, analyzing the image data content, light source, and position of each frame, so as to effectively overcome the existing white balance technology. missing.

Contents of the invention

In order to solve the existing white balance technology, further analysis of the overall composition content and image light source is ignored, and the continuity of the images of each segmented block is not analyzed, so as to accurately judge the scene content at the time, and it is easy to misjudge the image content of the segmented blocks , and make the white balance correction result deviate, the object of the present invention is to provide a white balance processing method and device.

To achieve the above object, the present invention adopts the following technical solutions:

A white balance processing method, the method includes the following steps: capturing an input image, dividing the input image into a plurality of segmented block images, and then performing a white point identification on each segmented block image to generate each segmented image A white point identification result of the block image. Then, according to the position of each segmented block image in the input image and the continuity with multiple segmented block images around it, the weight of the white point identification result corresponding to each segmented block image is determined. Afterwards, an estimated white balance value of the input image is calculated based on the white point identification results and weights of the divided block images.

Wherein, the aforesaid continuity means that the color temperature difference between each divided block image and multiple surrounding divided block images conforms to a preset value.

Wherein, the aforementioned step of determining the weight includes: dynamically determining the weight of a first light source type of the input image according to a brightness value of the input image; Determine whether the color temperature difference of the surrounding block images meets a preset value to determine a continuity judgment result; determine the input according to the weight of the first light source type, the position of each divided block image, and the continuity judgment result of each block image A second light source type weight of the image, wherein the second light source type weight includes a plurality of outdoor light source type weights and a plurality of indoor light source type weights; and according to the position of each segment image, the continuity of each segment image The judgment result and the weight of the second light source type determine the weight of the white point identification result of each segmented block image.

The dynamically determining the weight of a first light source type of the input image according to a brightness value of the input image further includes: determining a weight of the input image based on the brightness value of the input image and a plurality of color blocks in the input image that match preset data. The weight of the first light source type.

The white balance processing method of the present invention further includes: in each divided block image, according to an image characteristic of the divided block, selecting more than one white point identification block, and using more than one white point identification block image data For white spot identification. Wherein, the image characteristic is a neutral color or a flat block.

The white balance processing method of the present invention further includes: determining an input image color temperature threshold value according to the input image, and correcting the white balance estimated value according to the input image color temperature threshold value if the estimated white balance value exceeds the input image color temperature threshold value. If the estimated white balance value does not exceed the threshold value of the color temperature of the input image, the estimated white balance value is used for white balance processing.

The white balance processing method of the present invention further includes: if the weights of the plurality of outdoor light sources of the second light source type weight of the input image are higher than a preset ratio. In addition, the color temperature of each divided block image located above the input image is higher than a threshold value. The color temperature values of the segmented block image and the plurality of segmented block images around it have continuity, wherein the continuity means that the color temperature difference between the segmented block images conforms to a preset value. Then reduce the weight of the white point identification result of at least one segmented block image located above the input image.

The white balance processing method of the present invention further includes: if the weights of the plurality of outdoor light sources of the second light source type weight of the input image are lower than a preset ratio. In addition, the color temperature of each divided block image located above the input image is higher than a threshold value. The color temperature values of the segmented block and the images of the plurality of segmented blocks around it have continuity, wherein the continuity means that the color temperature difference between the images of the segmented block conforms to a preset value. Then the weight of the white point identification result of at least one segment image located above the input image is not adjusted.

The white balance processing method of the present invention further includes: if the color temperature of the plurality of segmented block images in the center of the input image is lower than a threshold value, and the color temperature value of each segmented block image and the plurality of segmented block images around it is continuous , wherein, continuity means that the color temperature difference between the images of the divided blocks conforms to a preset value. Then reduce the weight of the white point identification result of at least one segmented block image located in the center of the input image.

The white balance processing method of the present invention further includes: if the weights of the plurality of outdoor light sources of the second light source type weight of the input image are higher than a preset ratio. The segmented block image above the input image has a first color temperature, and each segmented block image has continuity with the color temperature values of multiple segmented block images around it, wherein the continuity is that the color temperature difference between the segmented block images conforms to a default value. The segmented image below the input image is a second color temperature, and the color temperature values of each segmented image and multiple segmented images around it are continuous, wherein the continuity is that the color temperature difference between the segmented blocks conforms to a default value. Then the weights of the white point identification results of some segmented block images are not adjusted.

Compared with the effect of the prior art: the present invention performs white balance processing by adding the weight analysis of the input image light source type, the position of the segmented block image and the continuity calculation with the surrounding segmented blocks, and highlights the color of a specific scene. Under the premise, the color rendering of the input image is accurately obtained.

The present invention also provides a white balance processing device, which includes:

An image capturing unit is used for capturing an input image. A division unit, coupled to the image capture unit, is used to divide the input image into a plurality of divided block images. A white point identification unit, coupled to the segmentation unit, conducts a white point identification for each segmented block image, and generates a white point identification result for each segmented block image. A white point weight determination unit, coupled to the white point identification unit, which can determine each segmented block according to the position of each segmented block image in the input image and the continuity of multiple segmented block images around it A weight for the white point identification result of the image. A calculation unit, coupled to the white point weight unit, calculates an estimated value of white balance for the input image by using the white point identification results and weights of the divided block images. A white balance processing unit, coupled to the computing unit, is used for performing a white balance process on the input image by using a white balance estimated value.

The white balance processing device of the present invention further includes: a white point selection unit, coupled to the segmentation unit and the white point identification unit, which can be used in each segmented block image according to an image of the segmented block characteristics, determine more than one white point identification block, and use its image data for white point identification. Wherein, the image characteristic is a neutral color or a flat area.

The white balance processing device of the present invention further includes: a correction unit, coupled to the calculation unit and the white balance processing unit, used for estimating the white balance according to the input image color temperature threshold when the white balance estimation value exceeds the input image color temperature threshold If the estimated value of white balance does not exceed the threshold value of the color temperature of the input image, the estimated value of white balance will be used for white balance processing.

Wherein, the white point weight determination unit further includes: a light source type weight calculation unit for determining a first light source type weight. A position determination unit, coupled to the light source type weight determination unit, is used to determine the position of the segmented block image in the input image. A continuity determination unit, coupled to the position determination unit, is used to determine whether the color temperature difference between each divided block image and a plurality of surrounding divided block images conforms to a preset value, so as to determine whether the divided block and its surrounding Continuity of color temperature values around the segmented blocks. In addition, the light source type weight determining unit can determine a second light source type weight of the input image according to the first light source type weight, the position of each divided block image, and the continuity of the color temperature difference between each block image and its surrounding block images. In addition, the white point weight determination unit can determine the weight of the white point identification result of each divided block image according to the position of each divided block image, the continuity of the color temperature difference between each block image and its surrounding block images, and the weight of the second light source type. .

The advantages and spirit of the present invention can be further understood through the following specific embodiments and accompanying drawings.

Description of drawings

Fig. 1 is a flow chart of the white balance processing method of the present invention;

Fig. 2 is a flow chart of the color temperature threshold correction mechanism of the present invention;

Fig. 3 is a block diagram of the white balance processing device of the present invention;

FIG. 4A and FIG. 4B are schematic diagrams of the present invention performing white balance processing on a background with a blue sky theme;

FIG. 5A and FIG. 5B are schematic diagrams of the present invention performing white balance processing on a scene with multiple objects of a specific color temperature arranged side by side;

6A and 6B are schematic diagrams of the present invention performing white balance processing on scenes with different continuous color temperature blocks; and

7A and 7B are schematic diagrams of the present invention performing white balance processing on a scene with a human face in the center.

Detailed ways

The detailed features and advantages of the present invention are described in detail below in the embodiments, the content of which is sufficient to enable any person skilled in the art to understand the technical content of the present invention and implement it accordingly, and according to the content disclosed in this specification, the patent scope of the application and the accompanying drawings , any person skilled in the art can easily understand the related objects and advantages of the present invention.

Most cameras are equipped with a white balance device. The present invention utilizes the existing white balance device of the camera, supplemented by segmentation, position identification, and methods for judging the surrounding relationship of the segmented blocks to achieve accurate white balance. The white balance correction method of the present invention uses an image capture device such as a digital camera to perform white balance processing. Please refer to FIG. 1 , which shows a flow chart of the white balance processing method of the present invention. As shown in the figure, this method includes the following steps.

及平均值(Mean value)

在各分割区块影像中，选择接近中性色及属平坦区的区块做白点鉴定。First, in step S10, an input image is captured, and a color temperature threshold value and a brightness value of the input image are determined; in step S13, the input image is divided into a plurality of divided block images, and each divided block image In step S15, according to the image characteristics, detect the image data source of more than one partial block in each segmented block that can be used as white point analysis, and select it as the white point identification of each segmented block image data in step S16 , that is, select the reference point that best represents its color in each segmented block, which is based on a preset characteristic curve, and uses the standard deviation (Standard deviation)

and mean value

In each segmented block image, select a block that is close to the neutral color and belongs to the flat area for white point identification.

After the input image is divided into a plurality of divided block images in step S13, a first light source type weight of the input image is determined according to the brightness value of the input image in step S14; each divided block image is detected in the input image by step S17 position, step S18 detects the continuity of each divided block image and the surrounding divided block images, step S19 determines a second light source type weight of the input image according to the composition analysis results of S17 and S18, and steps S20 with S17 The information in ~ S19 calculates the weight of the white point identification result of each segmented block image.

Next, in step S21, an estimated white balance value is determined according to the white point identification result of each divided block image and the weight obtained in step S20; if the estimated white balance value obtained in step S21 exceeds the color temperature threshold of the input image, execute Step S22, correct the white balance estimation value according to the input image color temperature threshold value, as shown in Figure 2, capture the image in step S201 and determine a color temperature threshold value, then calculate the block image white point identification weight in step S203, S205 obtains an estimated white balance value. If the estimated white balance value meets the color temperature threshold in step S207 , then directly perform white balance processing with the estimated white balance value in step S211 . If it is determined in step S207 that the white balance estimated value does not match the color temperature threshold value, then in step S209 the color temperature threshold value of the input image is corrected, and in step S211 the corrected white balance estimated value is used. The white balance handles. Wherein, the correction method can be a dynamic interpolation, if the difference between the color temperature threshold value and the estimated value of white balance increases, the correction value increases; if the difference between the color temperature threshold value and the estimated value of white balance decreases, the correction value performed reduce. On the contrary, if the estimated white balance value does not exceed the color temperature threshold of the input image, then in step S23 , a white balance process is directly performed on the input image by using the estimated white balance value. Wherein, the above white point identification technology is a well-known technology in the art, so it will not be repeated here.

The above is only a preferred example of the present invention, wherein the image can be a digital video camera, a mobile phone with a camera, or a notebook computer with a camera, etc., which have the same function, and should not be limited thereto.

So far, the technical characteristics of the present invention have been described, and the method for verifying the present invention will be described in detail below with specific examples, so that those skilled in the art can acquire sufficient knowledge to implement it.

Please refer to FIG. 4A , in the blue sky landscape composition example on a sunny day, the blue sky occupies most of the image, and the steps of using the white balance processing method of the present invention are as follows:

After capturing an input image, determine the color temperature threshold value and brightness value of the input image; cut the input image into a plurality of segmented block images, and perform white point identification on each of the segmented block images; according to the A brightness value of the input image dynamically determines the weight of a first light source type of the input image, specifically, according to the brightness value of the input image and a plurality of color blocks in the input image that match preset data, determine the first light source type weight of the input image. A light source type weight; according to the weight of the first light source type, the position of each of the divided block images, and whether the color temperature difference between each block image and its surrounding block images conforms to a preset value, determine a value of the input image The weight of the second light source type is determined to correspond to each of the segmented regions according to the position of each segmented block image in the input image, its continuity with the surrounding images of multiple segmented blocks, and the weight of the second light source type. A weight of the white point identification result of the block image, wherein, the second light source type weight includes a plurality of outdoor light source type weights and a plurality of indoor light source type weights; at this time, the upper half of the input image is blue sky image data with high color temperature (Fig. 4A, 51 blocks), wherein, the color temperature values of each segmented block image located above the input image and its surrounding multiple segmented block images are continuous (in FIG. The segmented block images, which have continuous color temperature values with the surrounding segmented block images 5001, 5002, 5003, 5004, 5005, 5006, 5007, 5008, 5009, 5010 and 5011), and the second input image The weight of the outdoor light source in the weight of the light source type is higher than the preset ratio, so the weight of the white point identification results of the above-mentioned blocks is lowered, so as to avoid including large blue images in the white balance statistical range, resulting in color cast in large blue sky images white, and accurate white balance can be obtained, so that the color rendering of the image with a large blue sky is more in line with what the human eye sees; the difference between the white balance estimation result and the color temperature value of the input image is low, and does not exceed the color temperature threshold of the input image. There is no need to adjust the white balance estimation result, and the white balance correction is performed directly based on it.

Please refer to FIG. 5A , in an example of indoor scene composition with multiple high color temperature objects juxtaposed above the image, the steps of using the white balance processing method of the present invention are as follows:

After capturing an input image, determine the color temperature threshold value and brightness value of the input image; cut the input image into a plurality of segmented block images, and perform white point identification on each of the segmented block images; according to the A brightness value of the input image dynamically determines a first light source type weight of the input image; according to the first light source type weight, the position of each of the divided block images, and the relationship between each of the block images and its surrounding block images Determine whether the color temperature difference value conforms to a preset value, and determine a second light source type weight of the input image; and then according to the position of each segmented block image in the input image, and a plurality of segmented block images around it The continuity and the weight of the second light source type determine a weight corresponding to the white point identification result of each of the divided block images; at this time, although the upper half of the input image is a plurality of image data with high color temperature (Fig. 5A, 612, 613, 614, 615, 616, 617, 618), but the color temperature values of each segmented block image above the input image and its surrounding multiple segmented block images are discontinuous (in FIG. 5B, represented by "600" is a segment image selected as a judgment, and its color temperature value is different from that of the surrounding segment images 6001, 6002, 6003, 6004, 6005, 6006, 6007, 6008, 6009, 6010, and 6011. Continuous), and the outdoor light source weight of the second light source type weight of the input image is lower than the preset ratio, it is judged that it is an unusual large blue sky outdoor scene, and the overall image color balance needs to be taken into account, so it is not lowered in the input image The weight of the white point identification results of the multiple segmented block images above is used to estimate the white balance; the difference between the white balance estimation result and the color temperature value of the input image is low, does not exceed the color temperature threshold of the input image, and does not need to be adjusted The white balance estimation result is directly used as a basis for white balance correction.

Please refer to FIG. 6A , which is an example of outdoor landscape composition with blue sky and green space. The steps of using the white balance processing method of the present invention are as follows:

After capturing an input image, determine the color temperature threshold value and brightness value of the input image; cut the input image into a plurality of segmented block images, and perform white point identification on each of the segmented block images; according to the A brightness value of the input image dynamically determines a first light source type weight of the input image; according to the first light source type weight, the position of each of the divided block images, and the relationship between each of the block images and its surrounding block images Determine whether the color temperature difference value conforms to a preset value, and determine a second light source type weight of the input image; and then according to the position of each segmented block image in the input image, and a plurality of segmented block images around it Continuity, the weight of the second light source type, determine a weight corresponding to the white point identification result of each of the divided block images; at this time, because the upper half of the input image is a large piece of image data with high color temperature (Fig. 6A 71 blocks), wherein, the color temperature values of each of the segmented block images located above the input image and a plurality of segmented block images around it are continuous (in FIG. The segmented block image of the input image, whose color temperature value is continuous with the surrounding segmented block images 7011, 7012, 7013, 7014, 7015, 7016, 7017), and the second light source type weight of the input image has a high outdoor light source weight In the preset ratio, the scene judged to be an outdoor blue sky scene should reduce the weight of the white point identification results of the plurality of segmented block images located above the input image to perform white balance estimation. However, the segmented block image below the input image is the second color temperature (block 72 in FIG. 6A ), and the color temperature values of each segmented block image and its surrounding multiple segmented block images are continuous (in FIG. 6B, "702" is selected as a selected block image, and its color temperature value is continuous with the surrounding block images 7021, 7022, 7023, 7024, 7025, 7026, and 7027), and the judgment is not adjusted at this time The weight of the white point identification results of all multiple divided block images, so as to avoid emphasizing the color rendering of a certain block of 71 or 72 in FIG. 6A and causing the color temperature imbalance of another block; the white balance estimation result and input The color temperature value difference of the image is low and does not exceed the color temperature threshold value of the input image, and the white balance estimation result does not need to be adjusted, and the white balance correction is performed directly based on it.

Please refer to FIG. 7A , an example of a general composition with a human face appearing in the center, the steps of using the white balance processing method of the present invention are as follows:

After capturing an input image, determine the color temperature threshold value and brightness value of the input image; cut the input image into a plurality of segmented block images, and perform white point identification on each of the segmented block images; A brightness value of the input image dynamically determines a first light source type weight of the input image; according to the first light source type weight, the position of each of the divided block images, each of the block images and its surrounding block images Determine whether the color temperature difference value of the input image conforms to a preset value, and determine a second light source type weight of the input image; then, according to the position of each segmented block image in the input image, and a plurality of segmented blocks around it Image continuity and the weight of the second light source type determine a weight corresponding to the white point identification result of each of the divided block images; at this time, because the center of the input image is a large piece of low-temperature image data (the 82 area of FIG. 7A block), wherein, the color temperature values of each of the divided block images located in the center of the input image and its surrounding multiple divided block images are continuous (in FIG. 7B, 801 is a selected divided block image, its color temperature value is continuous with the surrounding segmented block images 8011, 8012, 8013, 8014, 8015, 8016, 8017, and 8018), so it is judged that the scene is dominated by human faces, and the background located in the center of the input image is lowered. The weight of the white point identification result of each segmented block image is used to perform white balance estimation to avoid the pale color rendering of the face block; the difference between the white balance estimation result and the color temperature value of the input image is low, and does not exceed the stated value. The color temperature threshold value of the input image is directly used as a basis for white balance correction without adjusting the white balance estimation result.

In a scene where the input image has an extremely low color temperature, such as an input image in a tunnel, the steps of using the white balance processing method of the present invention are as follows:

After capturing an input image, it is determined that the color temperature threshold value of the input image is a low color temperature and its brightness value is lower than a threshold value, which belongs to an indoor scene; the input image is divided into a plurality of segmented block images; for each Identifying the white point of the segmented block image; dynamically determining a first light source type weight of the input image according to a brightness value of the input image; according to the first light source type weight, the position of each segmented block image , whether the color temperature difference between each of the block images and its surrounding block images meets a preset value, and determine a second light source type weight of the input image; and then according to each of the divided block images in the input The position of the image, the continuity of the images of multiple division blocks around it, and the weight of the second light source type determine a weight corresponding to the white point identification result of each of the division block images; because the plurality of division block images are all Low color temperature images, so the weights of the white point identification results of each block are not adjusted; an estimated white balance value for the input image is calculated based on the white point identification results and the weights of the divided block images; The difference between the white balance estimation result (correcting the image to be whiter) and the color temperature value of the input image (lower color temperature) is high, exceeding the color temperature threshold of the input image, so the dynamic interpolation method is used to move towards the color temperature threshold Correct the color temperature block to get close to the result observed by human eyes.

Please refer to FIG. 3 , which shows a block diagram of a white balance processing device of the present invention. In the figure, the white balance processing device includes an image capture unit 101, a segmentation unit 103, a white point selection unit 104, a white point identification unit 105, a white point weight determination unit 107, a light source type weight calculation unit 108, a A position determination unit 109 , a continuity determination unit 110 , a calculation unit 111 , a correction unit 113 and a white balance processing unit 115 .

The image capture unit 101 is used to capture an input image, and the segmentation unit 103 is coupled to the image capture unit and used to divide the input image into a plurality of divided block images. The white point selection unit 104 is coupled to the segmentation unit and the white point identification unit, and can determine more than one white point identification in each segmented block image according to the image characteristics of the segmented block (neutral color or flat block) For the block, use the white point identification block image data for the white point identification. The white point identification unit 105 is coupled to the segmentation unit, and performs a white point identification on each segmented block image to generate a white point identification result for each segmented block image. The white point weight determination unit 107 is coupled to the white point identification unit, and can determine each of the segmented regions according to the position of each segmented block image in the input image and the continuity of the plurality of segmented block images around it. The weight of the white point identification result of the block image; wherein the white point weight determination unit 107 also includes: a light source type weight calculation unit 108, used to determine a first light source type weight; a position determination unit 109, coupled to The light source type weight determination unit is used to determine the position of the divided block image in the input image; a continuity determination unit 110 is coupled to the position determination unit and used to determine each of the whether the color temperature difference between the segmented block image and multiple segmented block images around it meets a preset value; position, whether the color temperature difference between each block image and its surrounding block images conforms to a preset value, determines a second light source type weight of the input image; and the white point weight determination unit 107 can be based on each The position of the divided block image, whether the color temperature difference between each said block image and its surrounding block images conform to a preset value, and the weight of the second light source type determine the white point identification result of each said divided block image the weight of.

A calculation unit 111, coupled to the white point weight determination unit, can calculate a white balance estimation value for the input image based on the white point identification results and weights of the divided block images; a correction unit 113, coupled to In the calculation unit and the white balance processing unit, its function is: if the white balance estimated value exceeds the input image color temperature threshold value, the white balance estimated value is corrected according to the input image color temperature threshold value; if the If the estimated white balance value does not exceed the color temperature threshold of the input image, the white balance process is performed with the estimated white balance value. A white balance processing unit 115 , coupled to the calculation unit, is used for performing a white balance process on the input image by using the estimated white balance value.

The above is to illustrate the characteristics of the present invention through the embodiments, and its purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly, rather than to limit the patent scope of the present invention. Equivalent modifications or modifications accomplished by the disclosed spirit should still be included in the scope of patents described below.

Claims (14)

1. a white balancing treatment method is characterized in that, comprising:

Acquisition one input image;

Be a plurality of block images of cutting apart with described input Image Segmentation;

Each described block image of cutting apart is done white point evaluation, produce each described white point qualification result of cutting apart the block image;

According to each described block image of cutting apart in the position of described input image, and with described a plurality of continuitys of cutting apart the block image around it, determine corresponding each described weight of cutting apart the described white point qualification result of block image;

With more described described white point qualification result and the described weight of cutting apart the block image, calculate the white balance estimated value to described input image; And

With described white balance estimated value described input image being carried out a white balance processes.

2. white balancing treatment method as claimed in claim 1 is characterized in that, described continuity cuts apart the block image for each and a plurality of colour temperature differences of cutting apart between the block image around it meet a preset value.

3. white balancing treatment method as claimed in claim 1 is characterized in that, determines that the step of described weight also comprises:

Dynamically determine one first light source kind weight of described input image according to described input image one brightness value;

Judge each described position of cutting apart the block image above the described input image, central authorities or below;

Judge that whether each described colour temperature difference of cutting apart the block image around block image and its of cutting apart meets a preset value, obtains a continuity judged result;

According to described the first light source kind weight, each described position and described continuity judged result of cutting apart the block image, determine a secondary light source kind weight of described input image, wherein said secondary light source kind weight comprises a plurality of outdoor optical source category weights and a plurality of room light source category weight; And

According to each the described position of block image, described continuity judged result and described secondary light source kind weight cut apart, determine each described described weight of cutting apart block image white point qualification result.

4. white balancing treatment method as claimed in claim 3 is characterized in that, described weight dynamically determines the described first light source kind weight of described input image according to the described brightness value of described input image, also comprise:

With a plurality of color lumps that conform to preliminary setting data in the brightness value of described input image and the described input image, determine one first light source kind weight of described input image.

5. white balancing treatment method as claimed in claim 1 is characterized in that, also comprises:

Each described cutting apart in the block image, according to a described image characteristics of cutting apart the block image, select more than one white point to identify block, make white point with described more than one white point evaluation block image data and identify; Wherein, described image characteristics is a muted color or a smooth block.

6. white balancing treatment method as claimed in claim 1 is characterized in that, also comprises:

Determine an input photo dazzle color temperature threshold value according to described input image, if described white balance estimated value surpasses described input photo dazzle color temperature threshold value, then according to described input photo dazzle color temperature threshold value described white balance estimated value is revised;

If described white balance estimated value does not surpass described input photo dazzle color temperature threshold value, then carry out described white balance with described white balance estimated value and process.

7. white balancing treatment method as claimed in claim 1 is characterized in that, also comprises:

If a plurality of outdoor light source weight of a secondary light source kind weight of described input image is higher than a preset ratio;

And each the described colour temperature of cutting apart the block image that is positioned at described input image top is higher than a threshold value;

Describedly cut apart a plurality of block photo dazzle color temperature value tool continuitys of cutting apart around block image and its, wherein, described continuity is that the colour temperature difference of cutting apart between the block image meets a preset value;

Then reduce the described at least one described weight of cutting apart the described white point qualification result of block image that is positioned at described input image top.

8. white balancing treatment method as claimed in claim 1 is characterized in that, also comprises:

If a plurality of outdoor light source weight of a secondary light source kind weight of described input image is lower than a preset ratio;

And each the described colour temperature of cutting apart the block image that is positioned at described input image top is higher than a threshold value;

Describedly cut apart a plurality of block photo dazzle color temperature value tool continuitys of cutting apart around block and its, wherein, described continuity is that the colour temperature difference of cutting apart between the block image meets a preset value;

Then do not adjust the described at least one described weight of cutting apart the described white point qualification result of block image that is positioned at described input image top.

9. white balancing treatment method as claimed in claim 1 is characterized in that, also comprises:

If a plurality of colour temperatures of cutting apart the block image of described input image central authorities are lower than a threshold value, and each is described cuts apart the block image and a plurality of block photo dazzle color temperature value tool continuitys of cutting apart around it, wherein, described continuity is that the colour temperature difference of cutting apart between the block image meets a preset value;

Then reduce the described at least one described weight of cutting apart the described white point qualification result of block image be positioned at described input image central authorities.

10. white balancing treatment method as claimed in claim 1 is characterized in that, also comprises:

If a plurality of outdoor light source weight of a secondary light source kind weight of described input image is higher than a preset ratio;

The described block image of cutting apart of described input image top is one first colour temperature, and each is described cuts apart the block image and a plurality of block photo dazzle color temperature value tool continuitys of cutting apart around it, wherein, described continuity is that the colour temperature difference of cutting apart between the block image meets a preset value;

And the described block image of cutting apart of described input image below is one second colour temperature, and each is described cuts apart the block image and a plurality of block photo dazzle color temperature value tool continuitys of cutting apart around it, wherein, described continuity is that the colour temperature difference of cutting apart between the block image meets a preset value;

Then do not adjust the more described described weight of cutting apart the described white point qualification result of block image.

11. a white balance processing equipment is characterized in that, comprising:

One image acquisition unit is used for acquisition one input image;

One cutting unit is coupled to described image acquisition unit, and being used for described input Image Segmentation is a plurality of block images of cutting apart;

One white point is identified the unit, is coupled to described cutting unit, and each described block image of cutting apart is done white point evaluation, produces each described white point qualification result of cutting apart the block image;

One white point weight determining means, be coupled to described white point and identify the unit, it can be according to each described block image of cutting apart in the position of described input image, and determines corresponding each described weight of cutting apart the described white point qualification result of block image with described a plurality of continuitys of cutting apart the block image around it;

One computing unit is coupled to described white point weight determining means, with more described described white point qualification result and the described weight of cutting apart the block image, calculates the white balance estimated value to described input image; And

One white balance processing unit is coupled to described computing unit, is used for described white balance estimated value described input image being carried out a white balance and processes.

12. white balance processing equipment as claimed in claim 11 is characterized in that, also comprises:

One white point is chosen the unit, be coupled to described cutting unit and described white point and identify the unit, it is used at each described block image of cutting apart, according to a described image characteristics of cutting apart block, determine more than one white point evaluation block, make described white point with described more than one white point evaluation block image data and identify; Wherein, described image characteristics is a muted color or a smooth block.

13. white balance processing equipment as claimed in claim 11 is characterized in that, also comprises:

One amending unit is coupled to described computing unit and described white balance processing unit, if described white balance estimated value surpasses described input photo dazzle color temperature threshold value, then according to described input photo dazzle color temperature threshold value described white balance estimated value is revised; And

If described white balance estimated value does not surpass described input photo dazzle color temperature threshold value, then carry out white balance with described white balance estimated value and process.

14. white balance processing equipment as claimed in claim 11 is characterized in that, described white point weight determining means also comprises:

One light source kind weight calculation unit is used for determining one first light source kind weight;

One determining positions unit is coupled to described light source kind weight determining means, is used for determining that the described block image of cutting apart is in the position of described input image;

One continuity determining means, be coupled to described determining positions unit, be used for to determine whether each described block image of cutting apart meets a preset value with a plurality of colour temperature differences of cutting apart the block image around it, to determine described block image and its continuity judged result of cutting apart the block color temperature value on every side cut apart;

Described light source kind weight determining means can according to described the first light source kind weight, each described position, each described continuity judged result of cutting apart the block image of cutting apart the block image, determine a secondary light source kind weight of described input image again; And

Described white point weight determining means can according to each described position, each the described continuity judged result of block image, described secondary light source kind weight cut apart of cutting apart the block image, determine each described weight of cutting apart block image white point qualification result again.

Images