WO2022262997A1

WO2022262997A1 - A device and method for creating a high-dynamic range colour image

Info

Publication number: WO2022262997A1
Application number: PCT/EP2021/066649
Authority: WO
Inventors: Samu Koskinen; Mikko Muukki
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2021-06-18
Filing date: 2021-06-18
Publication date: 2022-12-22
Also published as: CN117203974A

Abstract

The disclosure relates to a radiation detection device comprising a pixel array superposed with a color filter array, wherein each pixel comprises one or more pixel parts. The pixel array comprises one or more blocks of pixels. Each block comprises at least four subblocks of pixels, each subblock of pixels being superposed by white color filters and colored color filters of a same respective color. For each subblock, all pixel parts superposed with white color filters are associated with a first binning group, wherein for each subblock all pixel parts superposed with colored color filters are associated with a second binning group, wherein in each subblock the positions of the pixel parts associated with the second binning group are arranged in a binning group pattern that is point-symmetric with respect to a geometrical center of the respective subblock.

Description

A DEVICE AND METHOD FOR CREATING A HIGH-DYNAMIC RANGE COLOUR

IMAGE

TECHNICAL FIELD

The disclosure relates to wavelength- sensitive detection of radiation. In particular, the disclosure proposes a radiation detection device and a corresponding method for operating the radiation detection device. The device comprises a pixel array superposed with a new kind of color filter array.

BACKGROUND

Image quality is influenced by many factors, but especially sharpness and the amount of details are important parameters of the image quality. In addition, the amount of noise also influences the image quality. These parameters are all impacted by the color filter array (CFA) pattern superposing an image sensor used to take the image, and by the subsequent image processing.

The well-known Bayer pattern has been the most popular CFA pattern in cameras. Lately, the so-called Quad Bayer pattern has been proposed and taken into use. However, it has a worse performance compared to the normal Bayer pattern in terms of sharpness and details in full resolution mode. In some devices, also the so-called Nona pattern has been used.

A Bayer pattern consists of repetitions of identical groups of 2x2 pixels of the image sensor superposed with color filters. The groups include 2 pixels superposed with green color filters, 1 pixel superposed with red color filters, and 1 pixel superposed with blue color filters. The 2 green pixels are included to achieve better sharpness. Using twice as many green elements as red or blue elements helps to mimic the physiology of the human eye.

A Quad Bayer CFA pattern consists of repetitions of 4x4 pixels of the image sensor superposed with color filters comprising four groups. Each group consists of 2x2 i.e. 4 pixels, wherein all pixels within a group are superposed with color filters of a same color. The four groups include 8 pixels superposed with green color filters, 4 pixels superposed with red color filters and 4 pixel superposed with blue color filters.

Pixel binning describes combining the information detected by pixels of the image sensor within a group (e.g., 4 pixels superposed with red color filters are combined into one pixel). Binning pixels according to a Quad Bayer CFA pattern, which consists of smaller pixels, will lead to an effective color detection pattern that is similar to a normal Bayer pattern with bigger pixels. However, in very dark conditions a sensor with bigger pixels produces brighter images than a same sized sensor with smaller pixels.

A Nona Bayer CFA pattern consists of repetitions of 6x6 pixels of the image sensor superposed with color filters comprising four groups. Each group consists of 3x3 i.e. 9 pixels, wherein all pixels within a group are superposed with color filters of a same color. The four groups include 18 pixels superposed with green color filters, 9 pixels superposed with red color filters, and 9 pixels superposed with blue color filters.

So-called RGBJ (red, green, blue, jade) color filters in a CFA pattern generally produce a better color performance compared to using one red, one blue and two green color filters in a CFA pattern.

Using a radiation detection system comprising two cameras can increase the total light sensitivity of the detection device while also collecting wavelength-sensitive information. Hereby, pixels in one camera are superposed by white color filters, and pixels in another camera are superposed by colored color filters. However, this solution requires the registration of images from different cameras.

Further, split pixels exist in some sensors and are currently used for improving the autofocus.

SUMMARY

In view of the above, embodiments of this disclosure aim to provide an improved CFA pattern and an effective color detection pattern, i.e. binning pattern. In particular, an object is to obtain an improved color performance, but also an improved sensitivity without requiring a multi camera system.

Further, embodiments of this disclosure intend to provide point-symmetric CFA and binning patterns that include pixels superposed with white color filters to increase the sensitivity while also balancing the binning gravity. Simultaneously, the embodiments intend to balance the color channel sensitivity, align the output after binning, and allow independent adjustments of the exposure times of different binning sub-/groups in a same subblock of pixels of an image sensor. In other words, the embodiments particularly intend to improve the color performance and sensitivity without requiring a multi-camera system.

These and other objectives are achieved by the embodiments of this disclosure as described in the enclosed independent claims. Advantageous implementations of the embodiments are further defined in the dependent claims.

In particular, a CFA pattern and binning pattern according to the embodiments of this disclosure combines some or all of the following benefits:

Increased sensitivity by using pixels superposed with white color filters;

- No phase shift after binning, i.e. the binning gravity may be balanced, or the center of gravity of each color channel may be the same with a constant interval;

The color channel sensitivities may be balanced (on average the white balance compensation gains are close to one);

The output after binning may form multiple frames that can be perfectly aligned;

- Easier high-dynamic-range capture, i.e. using short and long exposure times in the same image sensor for the same color channels may be possible without destroying the benefit of having no phase shift after binning.

A first aspect of this disclosure provides a radiation detection device comprising a pixel array superposed with a color filter array, wherein each pixel of the pixel array comprises one or more pixel parts, wherein each pixel part is superposed by one color filter of the color filter array, and wherein each color filter is of a red, blue, green, or white color filter type of a respective color. The pixel array comprises one or more blocks of pixels, each block of pixels being superposed with an identical pattern of color filters of the color filter array. Each block comprises at least four subblocks of pixels, each subblock of pixels being superposed by the color filters of the white color filter type and the color filters of either the red, blue, or green color filter type of a same respective color. For each subblock, all pixel parts superposed with the color filters of the white color filter type are associated with a first binning group, wherein for each subblock, all pixel parts superposed with color filters of the red, blue, or green color filter type are associated with a second binning group, wherein in each subblock the positions of the pixel parts associated with the second binning group are arranged in a binning group pattern that is point-symmetric with respect to a geometrical center of the respective subblock, wherein the binning group patterns are the same in each block, and wherein no color filter or a color filter that is without wavelength- selectivity is a color filter of the white color filter type.

According to the above, a binning gravity of each binning group may be balanced, as calculating a spatial average over a point-symmetric binning group pattern equates to the geometrical center of the subblock and binning group. Additionally, the color filters of the white color filter type increase the total sensitivity compared to only using colored color filters.

A binning gravity of a binning group or binning subgroup may still be considered to be balanced, if the discrete size of a pixel part leads to a spatial average that does not perfectly align with the geometrical center of the subblock in both dimensions, but only in one. For example, if in a subblock of 3x3 pixels the center left and center right pixel are in a second binning group, and the residual pixels are in a first binning group, then the spatial average of the second binning group would equate to a vertical line through the geometrical center of the subblock and not an infinitesimally small point in the geometrical center. It may be beneficial for the balance of the binning gravity to only use binning groups and/or binning subgroups with a respective spatial average that perfectly aligns with the geometrical center of the subblock. For example, if in a subblock of 3x3 pixels the corner pixels and the center pixel are in a second binning group, and the residual pixels are in a first binning group, then the spatial average of the second binning group would perfectly align with the geometrical center of the subblock.

Notably, each subblock comprises exactly two different color filter types, as each subblock comprises a first binning group and a second binning group. Additionally, different binning subgroups in a subblock may comprise color filters of the same type, but may also comprise color filters of different colors.

Further, a variety of subblock structures may be used. A subblock may comprise N x M pixels, wherein N and M are positive integers. For example, subblocks of 1x2, 3x3, 4x4, 2x3, or 3x4 pixels may be used.

Further, a variety of block structures may be used. A block may comprise N’ x M’ subblocks, wherein N’ and M’ are positive integers, and the product of N’ and M’ is larger than 3. For example, blocks of 3x3, 4x4, 2x3, or 3x4 subblocks may be used. In particular, blocks of 2x2 subblocks may be used. Other block structures with a product of N’ and M’ smaller than 4 may be used as well.

In a further implementation form of the first aspect, each block comprises four subblocks of pixels. Two of the subblocks in each pixel block comprise color filters of the green color filter type, one of the subblocks in each pixel block comprises color filters of the red color filter type, and one of the subblocks in each pixel block comprises color filters of the blue color filter type.

In a further implementation form of the first aspect, the subblocks in each block are arranged in a grid of 2x2 subblocks, and the two subblocks comprising color filters of the green color filter type are diagonally arranged in the grid.

Thus, a Bayer pattern of subblocks may be formed when associating each subblock with the colored color filters of the subblock. For example, if each subblock is combined into one hypothetical pixel, the resulting four hypothetical pixels superposed with the respective colored color filters of the respective subblock form a Bayer pattern. If each respective binning group is binned together, the resulting binning outputs represent four pixels superposed with colored color filters and four pixels superposed with white color filters, wherein the four pixels superposed with colored color filters form a Bayer pattern. Thus, the data may be processed with a standard Bayer demosaicing algorithm, even though the underlying CFA pattern may have a different or more complex pattern.

In a further implementation form of the first aspect, the top left subblock comprises color filters of a green color filter type of a first color, and the bottom right subblock comprises color filters of a green color filter type of a second color, and the top right subblock comprises color filters of the red color filter type of a color, and the bottom left subblock comprises color filters of the blue color filter type of a color.

Hereby, the above mentioned positions are with respect to a geometrical center of a block comprising a grid of 2x2 subblocks. Two subblocks are positioned in the top half of the block, and two subblocks are positioned in the bottom half of the block. Likewise, two subblocks are positioned in the left half of the block, and two subblocks are positioned in the right half of the block. Similar block patterns, i.e. binning group patterns, binning subgroup patterns and CFA patterns, may also be used, for example, rotated or flipped versions of the above described pattern.

In a further implementation form of the first aspect, each second binning group comprises one binning subgroup or each second binning group comprises two binning subgroups, wherein in each subblock the positions of the pixel parts associated with each binning subgroup are arranged in a binning subgroup pattern that is point-symmetric with respect to the geometrical center of the respective subblock, wherein the binning subgroup patterns are the same in each block.

Thus, the binning gravity is balanced for each binning group, as well as for each binning subgroup. In other words, all binning patterns, i.e. all binning group patterns and binning subgroup patterns, are point-symmetric with respect to the geometrical center of the respective subblock.

In a further implementation form of the first aspect, each binning subgroup in each subblock in a block corresponds mutually to one binning subgroup of each other subblock in the same block, and each first binning group in a subblock corresponds mutually to the first binning group in each other subblock in the same block, wherein all mutually corresponding first binning groups and binning subgroups are grouped into respective binning frames for processing.

In a further implementation form of the first aspect, the device comprises a processor for jointly processing each binning frame.

Thus, for each block there is one binning frame for the first binning group and one or more binning frames equal to the number of binning subgroups in each subblock. Each binning frame may consist of a number of binning outputs/binning sub-/groups that is equal to the number of subblocks in a block. If, for example, the block comprises four subblocks forming a grid of 2x2 subblocks, then the binning frames consist of 2x2 binning outputs, and the binning frames associated with the binning subgroups may be processed with a standard Bayer demosaicing algorithm, even though the underlying CFA pattern may have a different or more complex pattern. In other words, the binning frames may be perfectly aligned with regards to the Bayer demosaicing algorithm. In a further implementation form of the first aspect, each binning subgroup corresponds to radiation of a wavelength range depending on the bandpass of the color filters of a color filter type of a color superposing the pixel parts of the respective binning subgroup.

For example, an orange color filter of a red color filter type would typically have a bandpass in the 590-625 nm wavelength range. A binning subgroup superposed with such an orange color filter of a red color filter type would accordingly correspond to radiation in the 590-625 nm wavelength range.

In a further implementation form of the first aspect, in each subblock of the block, a total sensitivity to radiation of a first respective corresponding wavelength range of a binning subgroup, is approximately equal to a total sensitivity to radiation of a second respective corresponding wavelength range of each respective corresponding binning subgroup in the other subblocks of the block.

Thus, the total sensitivity to the respective corresponding wavelength range of each binning output in a same binning frame is approximately equal. Thus, each binning output of the same binning frame can be equally compared even though each binning output may correspond to different wavelength ranges. Accordingly, the color channel sensitivities are balanced. However, binning groups superposed with white color filters may have a different total sensitivity compared to the corresponding first binning groups in the same block. This is less important, as these first binning groups only convey intensity information and not wavelength information.

In a further implementation form of the first aspect, a total sensitivity of a binning subgroup superposed with color filters of a color filter type of a color to radiation of a corresponding wavelength range is the product of a sensitivity to radiation of the corresponding wavelength range of the color filters of the color filter type of the color superposing the pixel parts of the binning subgroup and a combined active surface area of the pixel parts of the binning subgroup.

Thus, a total sensitivity of a binning subgroup describes the total power a binning subgroup detects relative to the total incoming power of the corresponding wavelength range. Assuming equal intensity and equal wavelength of the incoming radiation over the entire binning subgroup, the total sensitivity increases linearly with an increasing active surface area of the pixel parts. Additionally, the total sensitivity depends linearly on how much radiation of the corresponding wavelength range each color filter transmits. Notably, all color filters in each binning group and binning subgroup are of the same color and type.

In a further implementation form of the first aspect, the pixel parts of each binning subgroup and each first binning group in each subblock of the block are configured to collect radiation with a different first exposure time compared to other binning subgroups and/or another first binning group in the respective subblock, wherein the pixel parts of each respective corresponding binning subgroup and each respective corresponding first binning group in the other subblocks of the block are configured to collect radiation with the same first exposure time.

For example, in a subblock of a block the first binning group may collect radiation with a first exposure time, a first binning subgroup may collect radiation with a second exposure time, and a second binning subgroup may collect radiation with a third exposure time. Accordingly, the first binning groups, first binning subgroups, and second binning subgroups in the other subblocks of the block may collect radiation with the respective same first, second and third exposure time. Thus, the three binning frames of the block may each be based on different exposure times. Therefore, different exposure times can be used while still having aligned binning frames. In particular, binning groups superposed with white color filters may use different exposure times, as their total sensitivity is generally larger.

In a further implementation form of the first aspect, in each block at least one binning group forms a disjoint cross pattern.

In a further implementation form of the first aspect, in each subblock a pixel array comprising all of the pixel parts in the respective subblock has a rectangular shape.

In a further implementation form of the first aspect, at least one subblock in a block comprises less surface area of pixel parts being superposed with white color filters compared to another subblock in the block.

Thus, at least one subblock comprises different binning groups and binning group patterns compared to the other subblocks in the same block. In a further implementation form of the first aspect, each pixel part is half the size or the same size of a pixel in the pixel array.

Accordingly, two pixel halves of a same pixel may be grouped into different binning groups and binning subgroups. Thus, the flexibility of adjusting the active surface area in a binning group and binning subgroup is increased. Thus, the channel sensitivity can be balanced more accurately.

In a further implementation form of the first aspect, each block comprises a pixel array of 9x9 pixels comprising four subblocks, each subblock comprising a pixel array of 3x3 pixels.

In a further implementation form of the first aspect, the top left subblock comprises a first binning subgroup comprising the top center and the bottom center pixel, and a second binning subgroup comprising the right half of the center left pixel and the left half of the center right pixel. The bottom right subblock comprises a first binning subgroup comprising the top center and the bottom center pixel, and a second binning subgroup comprising the center left pixel and the center right pixel. The bottom left subblock comprises a first binning subgroup comprising the corner pixels, and a second binning subgroup comprising the center pixel. The top right subblock comprises a first binning subgroup comprising the corner pixels, and a second binning subgroup comprising the center pixel. The residual pixel parts in each subblock are grouped into the first binning group respectively.

Similar implementations may be used. For example, an implementation with each pixel parts being the same size as a pixel, wherein the top left subblocks second binning subgroup comprises the center left pixel and the center right pixel. As another example, respective implementations may be used, wherein each second binning group consists of only one binning subgroup. As another example, respective rotated or flipped versions of the implementations may be used.

In a further implementation form of the first aspect, in each block all of the first binning subgroups correspond to each other, and all of the second binning subgroups correspond to each other, and all of the first binning groups correspond to each other. In a further implementation form of the first aspect, the device comprises a processor or an analog binning module for binning the pixel parts of the same binning group and/or binning subgroup together.

A Bayer demosaicing algorithm may be applied to process some binning frames, if a binning frame represents a Bayer pattern. Hereby, the color of the color filters superposing a respective binning sub-/group is treated as the color of the binning output. If, for example, four binning outputs of four binning sub-/groups form a binning frame, the binning frame may form a Bayer pattern according to the colors of each binning output.

In an implementation form of the first aspect, each of the red, green and blue color filter types is a color filter type of a respective color comprising a respective red, green or blue color component with a largest magnitude compared to the other color components; and wherein each color filter type of a respective color comprising two components of the same largest magnitude is a color filter type of one of the two components according to the following hierarchy in descending order: red, blue, green, red.

The colors may be defined in accordance with the RGB color model, wherein the three color components red, green and blue are combined with a respective magnitude into a color. For example, the color of an orange color filter comprises approximately the color components red with a magnitude of 75%, green with a magnitude of 25% and blue with a magnitude of 0%. Thus, in this case the color filter is a red color filter type of the color orange, as the component with the largest magnitude is the red component. In another example, the color of a yellow color filter comprises approximately the color components red with a magnitude of 50%, green with a magnitude of 50% and blue with a magnitude of 0%. Thus, the color filter is a green color filter type of the color yellow, as the components with the largest magnitude are the red and green component, and the above listed hierarchy defines that the green color component is prioritized in this case. In the case that all three color components are of approximately the same magnitude, the color filter is of a white color filter type, as the resulting color filter is substantially without wavelength- selectivity.

The wavelengths of the color components may be approximately in the range of 625-700 nm, 500-565 nm and 450-485 nm for the red, green and blue color component respectively. A typical example of wavelengths of the color components is 650 nm, 550 nm and 450 nm for the red, green and blue color component respectively. Other reddish, greenish and blueish colors may also be suitable candidates for the color components.

A second aspect of this disclosure provides a method of operating a radiation detection device, the device comprising a pixel array superposed with a color filter array, wherein each pixel of the pixel array comprises one or more pixel parts, wherein each pixel part is superposed by one color filter of the color filter array, and wherein each color filter is of a red, blue, green, or white color filter type of a respective color. The pixel array comprises one or more blocks of pixels, each block of pixels being superposed with an identical pattern of color filters of the color filter array. Each block comprises four subblocks of pixels, each subblock of pixels being superposed by color filters of the white color filter type and color filters of either the red, blue, or green color filter type of a same respective color. The method comprises associating for each subblock all pixel parts superposed with color filters of the white color filter type with a first binning group, and associating for each subblock all pixel parts superposed with color filters of the red, blue, or green color filter type with a second binning group, wherein in each subblock the positions of the pixel parts associated with the second binning group are arranged in a binning group pattern that is point-symmetric with respect to a geometrical center of the respective subblock, wherein the binning group patterns are the same in each block, and wherein no color filter or a color filter that is without wavelength-selectivity is a color filter of the white color filter type.

In a further implementation form of the second aspect, each block comprises four subblocks of pixels. Two of the subblocks in each pixel block comprise color filters of the green color filter type, one of the subblocks in each pixel block comprises color filters of the red color filter type, and one of the subblocks in each pixel block comprises color filters of the blue color filter type.

In a further implementation form of the second aspect, the subblocks in each block are arranged in a grid of 2x2 subblocks, and the two subblocks comprising color filters of the green color filter type are diagonally arranged in the grid.

In a further implementation form of the second aspect, the top left subblock comprises color filters of a green color filter type of a first color, and the bottom right subblock comprises color filters of a green color filter type of a second color, and the top right subblock comprises color filters of the red color filter type of a color, and the bottom left subblock comprises color filters of the blue color filter type of a color.

In a further implementation form of the second aspect, the method comprises partitioning each second binning group into one binning subgroup or each second binning group into two binning subgroups, wherein in each subblock the positions of the pixel parts associated with each binning subgroup are arranged in a binning subgroup pattern that is point-symmetric with respect to the geometrical center of the respective subblock, wherein the binning subgroup patterns are the same in each block.

In a further implementation form of the second aspect, each binning subgroup in each subblock in a block is set into mutual correspondence to one binning subgroup of each other subblock in the same block, and each first binning group in a subblock is set into mutual correspondence to the first binning group in each other subblock in the same block, wherein all mutually corresponding first binning groups and binning subgroups are grouped into respective binning frames for processing.

In a further implementation form of the second aspect, the device comprises a processor and the method comprises jointly processing each binning frame.

In a further implementation form of the second aspect, each binning subgroup is set into correspondence to radiation of a wavelength range depending on the bandpass of the color filters of a color filter type of a color superposing the pixel parts of the respective binning subgroup.

In a further implementation form of the second aspect, in each subblock of the block, a total sensitivity to radiation of a first respective corresponding wavelength range of a binning subgroup, is approximately equal to a total sensitivity to radiation of a second respective corresponding wavelength range of each respective corresponding binning subgroup in the other subblocks of the block.

In a further implementation form of the second aspect, a total sensitivity of a binning subgroup superposed with color filters of a color filter type of a color to radiation of a corresponding wavelength range is the product of a sensitivity to radiation of the corresponding wavelength range of the color filters of the color filter type of the color superposing the pixel parts of the binning subgroup and a combined active surface area of the pixel parts of the binning subgroup.

In a further implementation form of the second aspect, the pixel parts of each binning subgroup and each first binning group in each subblock of the block are configured to collect radiation with a different first exposure time compared to other binning subgroups and/or another first binning group in the respective subblock, wherein the pixel parts of each respective corresponding binning subgroup and each respective corresponding first binning group in the other subblocks of the block are configured to collect radiation with the same first exposure time.

In a further implementation form of the second aspect, in each block at least one binning group forms a disjoint cross pattern.

In a further implementation form of the second aspect, in each subblock a pixel array comprising all of the pixel parts in the respective subblock has a rectangular shape.

In a further implementation form of the second aspect, at least one subblock in a block comprises less surface area of pixel parts being superposed with white color filters compared to another subblock in the block.

In a further implementation form of the second aspect, each pixel part is half the size or the same size of a pixel in the pixel array.

In a further implementation form of the second aspect, each block comprises a pixel array of 9x9 pixels comprising four subblocks, each subblock comprising a pixel array of 3x3 pixels.

In a further implementation form of the second aspect, the top left subblock comprises, a first binning subgroup comprising the top center and the bottom center pixel, and a second binning subgroup comprising the right half of the center left pixel and the left half of the center right pixel. The bottom right subblock comprises, a first binning subgroup comprising the top center and the bottom center pixel, and a second binning subgroup comprising the center left pixel and the center right pixel. The bottom left subblock comprises, a first binning subgroup comprising the corner pixels, and a second binning subgroup comprising the center pixel. The top right subblock comprises, a first binning subgroup comprising the corner pixels, and a second binning subgroup comprising the center pixel. The residual pixel parts in each subblock are grouped into the first binning group respectively.

In a further implementation form of the second aspect, in each block all of the first binning subgroups are set into correspondence to each other, and all of the second binning subgroups are set into correspondence to each other, and all of the first binning groups are set into correspondence to each other.

In a further implementation form of the second aspect, the device comprises a processor or an analog binning module and the method comprises binning the pixel parts of the same binning group and/or binning subgroup together.

In an implementation form of the second aspect, each of the red, green and blue color filter types is a color filter type of a respective color comprising a respective red, green or blue color component with a largest magnitude compared to the other color components; and wherein each color filter type of a respective color comprising two components of the same largest magnitude is a color filter type of one of the two components according to the following hierarchy in descending order: red, blue, green, red.

The method of the second aspect and its implementation forms achieve the advantages and effects described above for the device of the first aspect and its respective implementation forms.

It has to be noted that all devices, elements, units and means described in the disclosure could be implemented in the software or hardware elements or any kind of combination thereof. All steps which are performed by the various entities described in the disclosure as well as the functionalities described to be performed by the various entities are intended to mean that the respective entity is adapted to or configured to perform the respective steps and functionalities. Even if, in the following description of specific embodiments, a specific functionality or step to be performed by external entities is not reflected in the description of a specific detailed element of that entity which performs that specific step or functionality, it should be clear for a skilled person that these methods and functionalities can be implemented in respective software or hardware elements, or any kind of combination thereof. BRIEF DESCRIPTION OF DRAWINGS

The above described aspects and implementation forms will be explained in the following description of specific embodiments in relation to the enclosed drawings, in which

FIG. 1 shows a device according to an example of this disclosure.

FIG. 2a shows an exemplary block of pixels comprising four subblocks according to an example of this disclosure.

FIG. 2b shows a subblock comprising two binning groups and two binning subgroups according to an example of this disclosure.

FIG. 3a shows three color components according to an example of this disclosure.

FIG. 3b shows six different exemplary colors of color filters and to which color filter type the respective color filters would belong according to an example of this disclosure.

FIG. 4a shows an exemplary block of pixels comprising pixel parts of half a pixel size and two binning subgroups according to an example of this disclosure.

FIG. 4b shows a corresponding binning frame according to an example of this disclosure.

FIG. 5 shows an exemplary pixel array comprising multiple blocks of pixels and an exemplary CFA pattern according to an example of this disclosure.

FIG. 6a shows an exemplary block of pixels comprising two binning groups in each subblock according to an example of this disclosure.

FIG. 6b shows a corresponding binning frame according to an example of this disclosure.

FIG. 7a shows an exemplary block of pixels comprising two binning subgroups in each subblock according to an example of this disclosure. FIG. 7b shows a corresponding binning frame according to an example of this disclosure.

FIG. 8a shows an exemplary block of pixels comprising two binning subgroups in each subblock according to an example of this disclosure.

FIG. 8b shows a corresponding binning frame according to an example of this disclosure.

FIG. 9a shows a shows an exemplary block of pixels comprising one binning subgroup in each subblock according to an example of this disclosure.

FIG. 9b shows a corresponding binning frame according to an example of this disclosure.

FIG. 10 shows multiple exemplary blocks of pixels with respective exemplary CFA patterns and binning patterns according to an example of this disclosure.

FIG. 11 shows multiple exemplary blocks of pixels with respective exemplary CFA patterns and binning patterns according to an example of this disclosure.

FIG. 12 shows a device according to an example of this disclosure.

FIG. 13 shows a device according to an example of this disclosure.

FIG. 14 shows a method according to an example of this disclosure.

DETAILED DESCRIPTION OF EMBODIMENTS

FIG. 1 shows a radiation detection device 100 according to an embodiment of this disclosure. The device 100 comprises a pixel array 101 (here an exemplary pixel array is illustrated) comprising at least one block of pixels 108 (an exemplary block of pixels 108 is illustrated). The pixels 103 of the pixel array 101 are superposed by a CFA 102 (an exemplary CFA 102 is illustrated), wherein the CFA 102 comprises multiple different color filters 105. In particular, each pixel 103 comprises one or more pixel parts 104, wherein each pixel part 104 is superposed by one color filter 105 of the CFA 102. Each color filter 105 is of a red 106a, blue 106b, green 106c, or white 106d color filter type 106 of a respective color 107. At least a part of an incoming radiation power may be attenuated by the CFA 102, which is positioned on top of the pixels 103 of the pixel array 101 such that it is in between the incoming radiation and the pixels 103. The pixels 103 of the pixel array 101 detect the (residual) radiation that is transmitted through the CFA 102. As the color filters of the CFA 102 may attenuate radiation depending on the wavelength of the radiation, the device 100 may collect wavelength- sensitive intensity information about the incoming radiation.

Generally, each block 108 of the pixel array 101 comprises at least four subblocks 109 of pixels 103, each subblock of pixels 103 being superposed by the color filters 105 of the white color filter type 109d and the color filters of either the red 106a, blue 106b, or green 106c color filter type 106 of a same respective color 107. For each subblock 109a, 109b, 109c, 109d, all pixel parts 104 superposed with the color filters 105 of the white color filter type 106d are associated with a first binning group 110. For each subblock 109a, 109b, 109c, 109d, all pixel parts 104 superposed with color filters of the red, blue, or green color filter type 106a, 106b, 106c are associated with a second binning group 111. In each subblock 109a, 109b, 109c, 109d the positions of the pixel parts 104 associated with the second binning group 111 are arranged in a binning group pattern 112 that is point-symmetric with respect to a geometrical center 121 of the respective subblock 109a, 109b, 109c, 109d.

FIG. 2a shows an exemplary block 108 of pixels 103 according to an embodiment of this disclosure. The block 108 comprises four subblocks 109 each comprising 3x3 pixels 103. The block of pixels 108 comprises pixel parts 104 of the same size as a pixel 103 and pixel parts 104 of half the size of a pixel 103. The pixel array 101 may comprise pixel parts 104 of various sizes, for example of a third or quarter of a pixel 103. A pixel part 104 may not be larger in size than a pixel 103. The split pixels 104 in the top left subblock 109d may be implemented using a dual PD (photodiode) structure. Split pixels are known to exist in current sensors but are currently used only for autofocus and not to balance channel sensitivity.

Each pixel part 104 is superposed by a color filter 105. The device 100 may comprise four types 106 of color filters: type white 106d, type green 106c, type red 106a and type blue 106b. All color filters 105 in a same subblock 109 of the same type 106 are of a same color 107. Color filters 105 that are of the same type 106 but are located in different subblocks 109 may or may not be of a different color 107. The exemplary block 108 in FIG. 2 comprises two subblocks 109 each comprising color filters of the green color filter type 106c. The top left subblock 109d comprises green color filters 107d of the green color filter type 106c while the bottom right subblock 109c comprises jade color filters 107c of the green color filter type 106c. The block 108 further comprises a top right subblock 109a comprising red color filters 107a of the red color filter type 106a. The block 108 further comprises a bottom left subblock 109b comprising blue color filters 107b of the blue color filter type 106b. All four subblocks 109 comprise white color filters of the white color filter type 106d.

FIG. 2b shows the top right subblock 109a of the exemplary block 108 in FIG. 2a. A geometrical center 121 of the subblock 109 is in the center of the center pixel 103. Subblocks 109 of different shapes, for example of 2x2 pixels 103, may have a geometrical center 121 of the subblock 109 in between pixels 103.

FIG. 3a shows three color components 113a, 113b, 113c according to an embodiment of this disclosure. The three color components may be red 113a, green 113b and blue 113c. The colors and color components are defined in reference to the RGB (red, green, blue) color model, wherein the components 113 are combined according to their respective magnitude into a color 107. For example, a yellow color comprises 50% magnitude of the red color component 113a, 50% magnitude of the green color component 113c, and 0% magnitude of the blue color component 113b. A color filter of a color 107 is, for example, of a red color filter type 106a, if the red color component 113a has the largest magnitude compared to the blue 113b and green 113c color components. Accordingly, a color filter 105 has a type 106 and a color 107, wherein the type 106 can be different from the color 107. For example, there may be a color filter 105 of a green color filter type 106c of a color jade 107c and a color filter 105 of a green color filter type 106c of a color green 107d.

The exemplary colors 107a, 107b, 107d depicted in FIG. 3a comprise, respectively, 100% magnitude of the red color component 113a, 100% magnitude of the blue color component 113b and 100% of the green color component 113c, and are accordingly the same as the color 107 of the respective color components 113a, 113b, 113c. In contrast, the colors 107c and 107e depicted in FIG. 3b comprise a mix out of multiple color components 113. Notably, color filters 105 of the colors 107d and 107c are of the same color filter type 106c, and color filters 105 of the colors 107a and 107e are of the same color filter type 106a. A hierarchy may be used to specify which color filters 105 are of which type, if, for example, the red 113a and green 113c component are combined with a same magnitude into a color 107 of a color filter 105. As the color 107 yellow comprises the red 113a and green 113c color components with the same magnitude, it would be ambiguous if the color filters 105 of a yellow color 107 are considered as a green 106c or a red 106a color filter type. According to, for example, the following hierarchy in descending order: red 113a, blue 113b, green 113c, red 113a, a yellow color filter 105 would be of a green color filter type 106c, as green 113c is directly left to red 113a in the hierarchy. Further, a color filter 105 of the color 107 magenta (50% R and 50% B) would be of the red color filter type 106a, as red 113a is directly left of blue 113b in the hierarchy.

FIG. 3b shows six different exemplary colors 107 of color filters 105, and of what color filter type 106 the respective color filters 105 would be. A color filter 105 of a white color filter type 106d may be a color filter 105 without wavelength selectivity. In particular, using no color filter 105 is considered as using a color filter 105 of the white color filter type 106d, as there would not be wavelength selective attenuation.

FIG. 4a shows the same exemplary pixel block 108 as Fig. 2a. Fig. 4a further outlines the binning groups 110, 111 and binning subgroups 114. In each subblock 109 all pixel parts 104 superposed with color filters 105 of the same type 106 are grouped into a same respective binning group 110, 111. In each subblock 109 all pixel parts 104 superposed with color filters 105 of the white color filter type 106d are grouped into a first binning group 110. In each subblock 109 all pixel parts 104 superposed with color filters 105 of either the red 106a, green 106c or blue color 106b filter type are grouped into a respective second binning group 111. Each second binning group 111 may or may not be partitioned further into binning subgroups 114. There may be one or more binning subgroups 119, 120 in each subblock 109. The exemplary block in FIG. 4a comprises a first binning group 110, a first binning subgroup 119 and a second binning subgroup 120. Hereby, the second binning group 111 is partitioned into the first binning subgroup 119 and the second binning subgroup 120.

Notably, pixel parts 104 that are positioned in different subblocks 109 are usually not in the same binning group 110, 111. Each binning group 110, 111 and each binning subgroup 114 typically comprises only pixels 103 of a single subblock 109. However, all first binning groups 110, all first binning subgroups 119, and all second binning subgroups 120 correspond to the respective other first binning groups 110, first binning subgroups 119, and second binning subgroups 120 in the other subblocks. The correspondence is accordingly adjusted if the block 108 comprises more or less binning subgroups 114.

Corresponding binning groups 110, 111 and/or binning subgroups 114 may be grouped into binning frames 118. Binning frames 118 represent the binning output 122 when every pixel part 104 is binned according to their binning group 110, 111 or binning subgroup 114. Each binning output 122 may be grouped with other corresponding binning outputs 122, and may thus form binning frames 118.

The exemplary block 108 depicted in FIG. 4a may be binned into the three binning frames 118a, 118b, 118c depicted in FIG. 4b, as each subblock 109a, 109b, 109c, 109d comprises a total of three binning sub-/groups, i.e. one first binning group 110, one first binning subgroup 119 and one second binning subgroup 120, that are binned together. As the block 108 comprises four subblocks 109, each binning frame 118a, 118b, 118c comprises four binning outputs 122. Accordingly, the twelve binning outputs 122 of the block 108 form three binning frames 118a, 118b, 118c each comprising four binning outputs 122, wherein each binning output 122 represents a first binning group 110 or a binning subgroup 119, 120. Two of the binning frames 118a, 118b form a Bayer pattern according to the color filters 105 that superpose the pixel parts 104 of the respective binning sub-/group 110, 111, 114. As the binning frames 118 represent a Bayer pattern, a standard Bayer demosaicing algorithm can be used to process multiple binning frames 118 into a color image. Notably, the binning frame 118 comprising the binning outputs 122 of all of the first binning groups 110a, 110b, 110c, 1 lOd does not represent a Bayer pattern, as the pixel parts 104 of each first binning group 110a, 110b, 110c, l lOd are superposed by color filter 105 of the white color filter type 106d. Different implementations of the device 100 may comprise different exemplary blocks 108 comprising a different number of binning frames 108 and binning outputs 122 depending on the number of subblocks 109 and binning sub groups 110, 111, 114 of the blocks.

FIG. 4b shows a corresponding binning frame 118 according to the exemplary pixel block 108 shown in FIG. 4a.

FIG. 5 shows an exemplary pixel array 101 comprising multiple blocks 108 of pixels 103 and an exemplary CFA 102 pattern according to an embodiment of this disclosure. A pixel array 101 may comprise one or more pixel blocks 108, wherein each pixel block 108 is identical to each other pixel block 108 in the same pixel array 101 in terms of CFA 102 pattern, binning group pattern 112 and binning subgroup pattern 115. Different implementations of the device 100 may comprise pixel blocks 108 of different CFA 102 patterns, binning group patterns 112 and binning subgroup patterns 115. However, the pixel blocks 108 of a same pixel array 101 are identical. Multiple pixel block 108 candidates according to an embodiment of this disclosure for such alternative implementations are depicted in FIG. 10. Notably, FIG. 5 shows an exemplary pixel array 101 comprising four pixel blocks 108, wherein each pixel block 108 is a pixel block 108 as shown in FIG. 4a.

FIG. 6a shows an exemplary block 108 of pixels comprising two binning groups 110, 111 according to an embodiment of this disclosure. The block 108 comprises four subblocks 109 each comprising 3x3 pixels 103 similarly as in a Nona pixel array. The CFA 102 pattern comprises color filters 105 of a white color filter type 106d and color filters 105 of a respective red 107a, blue 107b, green 107d and jade 107c color. Two subblocks 109 comprise color filters 105 ofthe green color filter type 106c of a respective green 107d andjade 107c color. Generally, CFA 102 patterns comprising two differently colored 107 color filters 105 of the green color filter type 106c achieve a better color performance. However, a simplified version comprising two subblocks 109 comprising color filters 105 of the green color filter type 106c of the same color 107d may be used. The other two subblocks comprise red 107a and blue 107b color filters 105 of the red 106a and blue 106b color filter type respectively. All four subblocks 109 additionally comprise color filters 105 of the white color filter type 106d.

Each subblock 109 comprises a first binning group 110 and a second binning group 111. Thus, the block comprises four first binning groups 110a, 110b, 110c, 1 lOd and four second binning groups 111a, 111b, 111c, 11 Id, wherein each first binning group 110 corresponds to each other, and each second binning group 111 corresponds to each other. Each first binning group 110 comprises the pixels 103 superposed with color filters 105 of the white color filter type 106d of the respective subblock 109, and each second binning group 111 comprises the pixels 103 superposed with color filters 105 of the other color filter type 106a, 106b, 106c ofthe respective subblock 109. Therefore, all pixel parts 104 superposed with color filters 105 of different colors 107 are binned separately. The second binning groups 120 in each subblock form a disjoint cross binning group pattern 112, comprising the center pixel 103 and the comer pixels 103. The first binning groups 110 in each subblock 109 form a disjoint cross binning group pattern 112 in each subblock 109, comprising the residual pixels 103. Therefore, the binning gravity of each binning group 110, 111 is balanced, as the binning group pattern 112 is point symmetric with respect to the geometrical center 121 of the respective subblock 109. In other words, the center point of each binning group 110, 111 is in the geometrical center 121 of the respective subblock 109. Therefore, the gravity/phase shift after binning is optimal.

The sensitivity to radiation of the block 108 may be increased compared to the prior art, as multiple pixels 103 are superposed with color filters 105 of the white color filter type 106d. Color filters 105 of the white color filter type 106d attenuate less radiation, as they transmit radiation of all wavelengths approximately equally. The channel sensitivity of the block 108 may or may not be balanced. In particular, it may not be balanced, as the color filters 105 of different colors 107 attenuate radiation of the respective corresponding wavelength range with a different magnitude. For example, green 107d color filters 105 typically attenuate green 107d radiation less than blue 107b color filters 105 attenuate blue 107b radiation. Further, the pixels 103 of the first binning group 110 may have different exposure times than the pixels 103 of the second binning group 111.

FIG. 6b shows a corresponding binning frame 118 according to the exemplary pixel block 108 shown in FIG. 6a. As there are in total two binning groups 110, 111 and zero/one binning subgroup 114 in each subblock 109 the pixels 103 may be binned into two binning frames 118, wherein the bottom binning frame 118a represents a Bayer pattern, and the top binning frame 118c does not include color information.

FIG. 7a shows an exemplary block 108 of pixels 103 comprising two binning subgroups 119, 120 in each subblock 109 according to an embodiment of this disclosure. The block 108 is identical to the block 108 shown in FIG. 6a except that each second binning group 111 is partitioned into two binning subgroups 119, 120, wherein each first binning subgroup 119 comprises the center pixel 103, and each second binning subgroup 120 comprises the corner pixels 103 respectively. Further, the pixels 103 of the two binning subgroups 119, 120 may collect radiation with a different exposure time. For example, the pixels 103 of all first binning subgroups 119 may collect radiation with a first short exposure time, and the pixels 103 of all second binning subgroups 120 may collect radiation with a second long exposure time. Additionally, the pixels 103 of all first binning groups 110 may collect radiation with a different third exposure time.

FIG. 7b shows a corresponding binning frame 118 according to the exemplary pixel block 108 shown in FIG. 7a. As there are in total three binning sub-/groups in each subblock 109, i.e. one first binning group 110, one first binning subgroup 119 and one second binning subgroup 120, the pixels 103 may be binned into three binning frames 118a, 118b, 118c, wherein two binning frames 118a, 118b include color information. As the pixels 103 of the two binning subgroups 119, 120 may collect radiation with a different exposure time, the two binning frames 118a, 118b including color information may represent different exposure times. Therefore, there can be short and long exposure times in the same sensor for the same color channels, which enables easier high-dynamic-range capture.

FIG. 8a shows an exemplary block of pixels 108 comprising two binning subgroups 119, 120 in each subblock 109 according to an embodiment of this disclosure. The block 108 is identical to the block 108 shown in FIG. 7a except of the CFA 102 patterns of the two subblocks 109c, 109d comprising color filter 105 of the green color filter type 106c, i.e. the top left 109d and bottom right 109c subblock. The center and corner pixels 103 are superposed by color filters 105 of the white color filter type 106d and the residual pixels 105 are superposed by color filters 105 of the green color filter type 106c of the respective subblock 109. Accordingly, these two subblocks 109c, 109d comprise four pixels 105 that are superposed by colored color filters 106a, 106b, 106c, wherein the other subblocks 109a, 109b in the block 108 each comprise five pixels 105 superposed with colored 106a, 106b, 106c color filters 105.

For optimal channel balance, the color channel sensitivities (assuming a typical illuminant) would be the same after binning. The channel sensitivity of the block 108 shown in FIG. 8a may or may not be balanced. In particular, it may be balanced, as color filters 105 of the green color filter type 106c typically attenuate radiation of a corresponding wavelength range less than color filters 105 of other colored color filter types 106a, 106b attenuate radiation of a respective corresponding wavelength range. In each of the two subblocks 109c, 109d there are fewer pixels 103 superposed with color filters 105 of the green color filter type 106c. Thus, the active surface area of pixels 103 collecting the radiation is decreased, and the total sensitivity to radiation of the corresponding wavelength range of each of the two subblocks 109c, 109d is decreased. Therefore, the total sensitivity to radiation of a respective corresponding wavelength range of the two subblocks 109c, 109d may be more equal to the total sensitivity to radiation of a respective corresponding wavelength range of the other subblocks 109a, 109b, i.e. the color channel sensitivities may be balanced.

FIG. 8b shows a corresponding binning frame 118 according to the exemplary pixel block 108 shown in FIG. 8a.

However, it may be calculated that the channel sensitivities would be balanced or very similar if the number of full pixels 103 superposed with colored 106a, 106b, 106c color filters 105 in the respective 3x3 pixel 103 subblocks 109 would be as follows: red = 5, green = 3, blue = 5 and jade = 4. Two CFA 102 patterns supporting this are shown in FIG. 2a and FIG. 9a.

FIG. 9a shows an exemplary block of pixels 108 comprising one binning subgroup 119 according to an embodiment of this disclosure. The block 108 is identical to the block 108 shown in FIG. 8a except of the CFA pattern in the top left subblock 109d, and that each subblock 109 comprises only one binning subgroup 119. In the top left subblock 109d, the center right, center, and center left pixels 103 are superposed with color filters 105 of the green color filter type 106c, and the residual pixels 105 are superposed by color filters 105 of the white color filter type 106d. Accordingly, the CFA 102 patterns and binning sub-/group patterns 112, 115 are point- symmetric while simultaneously the channel sensitivities would be balanced according to the above mentioned requirement, as in total only three total pixels 103 are superposed with green 107d color filters 105.

The pixel block 108 shown in FIG. 2a, FIG. 4, and FIG. 5 comprises four half pixels 104 in the top left subblock 109d. The block 108 is identical to the block 108 shown in FIG. 8a except that in the top left subblock 109d the center left and center right pixels 103 are replaced by two half pixels 104 respectively. The right half 104 of the center left pixel 103 and left half 104 of the center right pixel 103 is superposed with color filters 105 of the green color filter type 106c, and the other two halves 104 are superposed with color filters 105 of the white color filter type 106d. Accordingly, the CFA 102 patterns and binning sub-/group patterns 112, 115 are point- symmetric while simultaneously the channel sensitivities would be balanced according to the above mentioned requirement, as in total only three total pixels 103 are superposed with green 107d color filters 105. FIG. 10 and FIG. 11 show multiple exemplary blocks 108 of pixels 103 with respective exemplary CFA 102 patterns and binning patterns 123 according to an embodiment of this disclosure. The exemplary blocks 108 of pixels 103 illustrate some of the CFA 102 patterns, binning group patterns 112, binning subgroup patterns 115, pixel parts 104, pixel subblocks 109, number of pixels 103 and number of pixel subblocks 109 that may be used in a block 109 of a pixel array 101 of a device 100. Combinations of some of the features of different exemplary blocks 108 may be used. For example, the colors 107 of one exemplary block 108 may be combined with the CFA 102 pattern, binning group 112 and/or binning subgroup pattern 115 of another exemplary block 108 to represent another pixel block 108 according to an embodiment of this disclosure. Further, variations of the exemplary pixel blocks 108 according to the embodiments of this disclosure may be used.

The device 100 may comprise a processor 200 as shown in FIG. 12 or an analog binning module 201 as shown in FIG. 13 for binning the pixel parts 104 of the same binning group 110, 111 or binning subgroup 114 together. Pixel parts 104 may be binned together digitally or analog. In particular, analog binning may comprise adding the collected charges of pixel parts 103 in the same binning sub-/group 110, 111, 114.

The device 100 may further comprise a processor 300 as shown in FIG. 12 and FIG. 13 for jointly processing each binning frame 118. A color image representing incoming radiation may be created, by processing the binning frames 118 of one or more blocks 108, as the information contained in the binning frames 118 is dependent on the wavelength and intensity of the incoming radiation.

Generally, the processors 200 and/or 300 may be configured to perform, conduct or initiate the various operations of the device 100 described herein. The processors 200, 300 may comprise hardware and/or may be controlled by software. The hardware may comprise analog circuitry or digital circuitry, or both analog and digital circuitry. The digital circuitry may comprise components such as application-specific integrated circuits (ASICs), field-programmable arrays (FPGAs), digital signal processors (DSPs), or multi-purpose processors. The device 100 may further comprise memory circuitry, which stores one or more instruction(s) that can be executed by the processor 200, 300, in particular under control of the software. For instance, the memory circuitry may comprise a non-transitory storage medium storing executable software code which, when executed by the processor 200, 300, causes the various operations of the device 100 to be performed. In one embodiment, the device 100 may comprises one or more processors 200, 300 and a non-transitory memory connected to the one or more processors 200, 300. The non-transitory memory may carry executable program code which, when executed by the one or more processors 200, 300, causes the device 100 to perform, conduct or initiate the operations or methods described herein

FIG. 14 shows a method 1000 according to an embodiment of this disclosure. The method 1000 may be performed by the device 100. The method 1000 comprises a step 1001 of associating for each subblock 109 all pixel parts 104 superposed with color filters 105 of the white color filter type 106d with a first binning group 110. Further, the method 1000 comprises a step 1002 of associating for each subblock 109 all pixel parts 104 superposed with color filters 105 of the red 106a, blue 106b, or green 106c color filter type with a second binning group 111. The disclosure has been described in conjunction with various embodiments as examples as well as implementations. However, other variations can be understood and effected by those persons skilled in the art and practicing the claimed matter, from the studies of the drawings, this disclosure and the independent claims. In the claims as well as in the description the word “comprising” does not exclude other elements or steps and the indefinite article “a” or “an” does not exclude a plurality. A single element or other unit may fulfill the functions of several entities or items recited in the claims. The mere fact that certain measures are recited in the mutual different dependent claims does not indicate that a combination of these measures cannot be used in an advantageous implementation.

Claims

1. A radiation detection device (100) comprising a pixel array (101) superposed with a color filter array (102), wherein each pixel (103) of the pixel array (101) comprises one or more pixel parts (104), wherein each pixel part (104) is superposed by one color filter (105) of the color filter array (102), and wherein each color filter (105) is of a red (106a), blue (106b), green (106c), or white (106d) color filter type (106) of a respective color (107); wherein the pixel array comprises one or more blocks (108) of pixels, each block (108) of pixels (103) being superposed with an identical pattern of color filters (105) of the color filter array (102); wherein each block (108) comprises at least four subblocks (109) of pixels (103), each subblock (109a, 109b, 109c, 109d) of pixels (103) being superposed by the color filters (105) of the white color filter type (109d) and the color filters of either the red (106a), blue (106b), or green (106c) color filter type (106) of a same respective color (107); wherein for each subblock (109a, 109b, 109c, 109d), all pixel parts (104) superposed with the color filters (105) of the white color filter type (106d) are associated with a first binning group (110); wherein for each subblock (109a, 109b, 109c, 109d), all pixel parts (104) superposed with color filters of the red, blue, or green color filter type (106a, 106b, 106c) are associated with a second binning group (111); wherein in each subblock (109a, 109b, 109c, 109d) the positions of the pixel parts (104) associated with the second binning group (111) are arranged in a binning group pattern (112) that is point-symmetric with respect to a geometrical center (121) of the respective subblock (109a, 109b, 109c, 109d); wherein the binning group patterns (112) are the same in each block (108); and wherein no color filter or a color filter (105) that is without wavelength- selectivity is a color filter (105) of the white color filter type (106d).

2. The radiation detection device (100) according to claim 1, wherein each block (108) comprises four subblocks (109) of pixels (103), wherein two of the subblocks (109) in each pixel block (108) comprise color filters (105) of the green color filter type (106c), one of the subblocks (109) in each pixel block (108) comprises color filters (105) of the red color filter type (106a), and one of the subblocks in each pixel block (108) comprises color filters (105) of the blue color filter type (106b).

3. The radiation detection device (100) according to claim 2, wherein the subblocks (109) in each block (108) are arranged in a grid of 2x2 subblocks, and the two subblocks comprising color filters (105) of the green color filter type (106c) are diagonally arranged in the grid.

4. The radiation detection device (100) according to claim 3, wherein the top left subblock (109d) comprises color filters (105) of a green color filter type (106c) of a first color (107d), and the bottom right subblock (109c) comprises color filters (105) of a green color filter type (106c) of a second color (107c), and the top right subblock (109a) comprises color filters (105) of the red color filter type of a color (107a), and the bottom left subblock (109b) comprises color filters (105) of the blue color filter type of a color (107b).

5. The radiation detection device (100) according to any of the preceding claims, wherein each second binning group (111) comprises one binning subgroup (114) or each second binning group (111) comprises two binning subgroups (114), wherein in each subblock the positions of the pixel parts associated with each binning subgroup (114) are arranged in a binning subgroup pattern (115) that is point-symmetric with respect to the geometrical center 121 of the respective subblock (109), wherein the binning subgroup patterns (115) are the same in each block.

6. The radiation detection device (100) according to claim 5, wherein each binning subgroup (114) in each subblock (109) in a block (108) corresponds mutually to one binning subgroup (114) of each other subblock (109) in the same block (108), and each first binning group (110) in a subblock (109) corresponds mutually to the first binning group (110) in each other subblock (109) in the same block (108), wherein all mutually corresponding first binning groups (110) and binning subgroups (114) are grouped into respective binning frames (118) for processing.

7. The radiation detection device (100) according to claim 6, comprising a processor (300) for jointly processing each binning frame (118).

8. The radiation detection device (100) according to claim 6 or 7, wherein each binning subgroup (114) corresponds to radiation of a wavelength range depending on the bandpass of the color filters (105) of a color filter type (106) of a color (107) superposing the pixel parts (104) of the respective binning subgroup (114).

9. The radiation detection device (100) according to claim 8, wherein in each subblock (109) of the block (108), a total sensitivity to radiation of a first respective corresponding wavelength range of a binning subgroup (114), is approximately equal to a total sensitivity to radiation of a second respective corresponding wavelength range of each respective corresponding binning subgroup (114) in the other subblocks (109) of the block (108).

10. The radiation detection device (100) according to claim 9, wherein a total sensitivity of a binning subgroup (114) superposed with color filters (105) of a color filter type (106) of a color (107) to radiation of a corresponding wavelength range is the product of a sensitivity to radiation of the corresponding wavelength range of the color filters (105) of the color filter type (106) of the color (107) superposing the pixel parts (104) of the binning subgroup (114) and a combined active surface area of the pixel parts (104) of the binning subgroup (114).

11. The radiation detection device (100) according to one of the claims 6 to 10, wherein the pixel parts (104) of each binning subgroup (114) and each first binning group (110) in each subblock (109) of the block (108) are configured to collect radiation with a different first exposure time compared to other binning subgroups (114) and/or another first binning group (110) in the respective subblock (109), wherein the pixel parts (104) of each respective corresponding binning subgroup (114) and each respective corresponding first binning group (110) in the other subblocks (109) of the block (108) are configured to collect radiation with the same first exposure time.

12. The radiation detection device (100) according to any of the preceding claims, wherein in each block (108) at least one binning group (110, 111) forms a disjoint cross pattern.

13. The radiation detection device (100) according to any of the preceding claims, wherein in each subblock (109) a pixel array comprising all of the pixel parts (104) in the respective subblock (109) has a rectangular shape.

14. The radiation detection device (100) according to any of the preceding claims, wherein at least one subblock (109) in a block (108) comprises less surface area of pixel parts (104) being superposed with white color filters (106d) compared to another subblock (109) in the block (108).

15. The radiation detection device (100) according to any of the preceding claims, wherein each pixel part (104) is half the size or the same size of a pixel (103) in the pixel array (101).

16. The radiation detection device (100) according to any of the preceding claims, wherein each block (108) comprises a pixel array of 9x9 pixels (103) comprising four subblocks (109a, 109b, 109c, 109d), each subblock (109a, 109b, 109c, 109d) comprising a pixel array of 3x3 pixels (103).

17. The radiation detection device (100) according to claim 16, wherein the top left subblock (109d) comprises a first binning subgroup (119d) comprising the top center and the bottom center pixel, and a second binning subgroup (120d) comprising the right half of the center left pixel and the left half of the center right pixel, and the bottom right subblock (109c) comprises a first binning subgroup (119c) comprising the top center and the bottom center pixel, and a second binning subgroup (120c) comprising the center left pixel and the center right pixel, and the bottom left subblock (109b) comprises a first binning subgroup (119b) comprising the comer pixels, and a second binning subgroup (120b) comprising the center pixel, and the top right subblock (109a) comprises a first binning subgroup (119a) comprising the comer pixels, and a second binning subgroup (120a) comprising the center pixel, wherein the residual pixel parts in each subblock (109) are grouped into the first binning group (110a, 110b, 110c, 1 lOd) respectively.

18. The radiation detection device (100) according to claim 17, wherein in each block all of the first binning subgroups (119a, 119b, 119c, 119d) correspond to each other, and all of the second binning subgroups (120a, 120b, 120c, 120d) correspond to each other, and all of the first binning groups (110a, 110b, 110c, l lOd) correspond to each other.

19. The radiation detection device (100) according to any of the preceding claims, comprising a processor (200) or an analog binning module (201) for binning the pixel parts (104) of the same binning group (110, 111) and/or binning subgroup (114) together.

20. The radiation detection device (100) according to any of the preceding claims, wherein each of the red (106a), green (106c) and blue (106b) color filter types is a color filter type of a respective color (107) comprising a respective red (113a), green (113c) or blue (113b) color component (113) with a largest magnitude compared to the other color components; and wherein each color filter type (106) of a respective color (107) comprising two components (113) of the same largest magnitude is a color filter type (106) of one of the two components according to the following hierarchy in descending order: red (113a), blue (113b), green (113c), red (113a).

21. A method (1000) of operating a radiation detection device (100), the device comprising a pixel array (101) superposed with a color filter array (102), wherein each pixel (103) of the pixel array (101) comprises one or more pixel parts (104), wherein each pixel part (104) is superposed by one color filter (105) of the color filter array (102), and wherein each color filter (105) is of a red (106a), blue (106b), green (106c), or white (106d) color filter type of a respective color (107); wherein the pixel array (101) comprises one or more blocks (108) of pixels (103), each block (108) of pixels (103) being superposed with an identical pattern of color filters (105) of the color filter array (102); and wherein each block (108) comprises four subblocks (109) of pixels (103), each subblock (109) of pixels (103) being superposed by color filters (105) of the white color filter type (106d) and color filters (105) of either the red (106a), blue (106b), or green (106c) color filter type of a same respective color (107); the method comprising

- associating (1001) for each subblock (109) all pixel parts (104) superposed with color filters (105) of the white color filter type (106d) with a first binning group (110); and

- associating (1002) for each subblock all pixel parts superposed with color filters of the red, blue, or green color filter type with a second binning group (111); wherein in each subblock (109a, 109b, 109c, 109d) the positions of the pixel parts (104) associated with the second binning group (111) are arranged in a binning group pattern (112) that is point-symmetric with respect to a geometrical center (121) of the respective subblock (109a, 109b, 109c, 109d), wherein the binning group patterns (112) are the same in each block (108), wherein no color filter or a color filter (105) that is without wavelength- selectivity is a color filter (105) of the white color filter type (106d).