JP2007114923A - Image processing apparatus and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus and method capable of processing with efficiency, and high speed images obtained by a plurality of image pickup devices. <P>SOLUTION: In an image processing part 2, image parallelizing parts 9 for implementing parallelization image processes on images obtained by a main image pickup device 4 and an auxiliary image pickup device 5 are provided, each in one-to-one correspondence with the image pickup device 4 and the auxiliary image pickup device 5. Each of the image parallelizing parts 9 includes a correction table 10 which stores the relationship between the coordinate data of pixels after the parallelization image processes are implemented using image data obtained by image pickup and the coordinate data of pixels corresponding to the corrected pixels before the image processes are implemented. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an image processing device and an image processing method, and more particularly, to an image processing device and an image processing method using a plurality of imaging devices.

  In general, a method of recognizing a solid using an image is roughly classified into a method of recognizing a solid from one image and a method of recognizing a solid from a plurality of images. In the former method, since the depth data of the image is lost with only one image, it is necessary to compensate for it by some method. Further, the latter method can restore the three-dimensional data in more detail than when using a single image. As the latter representative method, there is known a method of recognizing a solid using a stereo image obtained by photographing the same object by two or more imaging devices installed in different places.

  In such a method for recognizing a solid using a stereo image, the same object is photographed by a plurality of imaging devices, and the corresponding points of the same object are determined from the respective images acquired by photographing. The three-dimensional coordinates of the point can be obtained.

  As described above, the method for determining the corresponding points of the same property is to re-project images taken by a plurality of cameras on the same plane and correct each camera image to the same image. A technique for processing is disclosed (for example, see Patent Document 1). This technique is a method for calculating a conversion parameter for associating a pixel position between an image of an imaging surface image obtained by a camera and an image to be reprojected, and forming a corrected image using the conversion parameter.

  In addition, a technique for efficiently calculating epipolar lines necessary for associating a plurality of images captured by a plurality of cameras with each other by simple calculation has been disclosed (for example, see Patent Document 2).

  Furthermore, a technique that can detect a correction condition used for aligning epipolar lines even when a specific test pattern does not exist in a plurality of images taken by a plurality of cameras has been disclosed (for example, (See Patent Document 3).

In addition, a technique for matching epipolar lines by correcting one of images captured by multiple cameras under various conditions and detecting the correspondence between the corrected image and the other image has been released. (For example, see Patent Document 4).
JP 7-294215 A JP-A-6-42941 Japanese Patent Laid-Open No. 9-153148 JP-A-11-203476

  However, Patent Document 1 proposes a means for parallelizing an image obtained by photographing. However, since the means for image processing is the same as the conventional one, there is a problem that the processing time cannot be shortened. Yes.

  Patent Document 2, Patent Document 3 and Patent Document 4 propose a means for detecting an epipolar line. However, since the means for parallelizing an image is the same as the conventional one, high-precision image processing is performed. Has a problem that can not be.

  The present invention has been made in view of the above points, and it is an object of the present invention to provide an image processing apparatus and an image processing method capable of performing image processing efficiently and at high speed on images obtained by a plurality of imaging devices. It is what.

In order to solve such a problem, the invention described in claim 1
A buffer for storing image data obtained by imaging the same object with a plurality of imaging devices;
An image processing apparatus comprising: an image processing unit that processes image data acquired by each of the imaging devices to form corrected image data;
The image processing unit starts processing to form the corrected image data by parallelizing the image data when the minimum unit of the image data capable of forming the corrected image data is stored in a buffer. To do.

  According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the image processing unit stores a correspondence relationship between the image data and the corrected image data that can be formed from the image data. The corrected image data is formed from the image data stored in the buffer using the correspondence relationship stored in the correction table.

  According to a third aspect of the present invention, in the image processing apparatus according to the second aspect, the correspondence relationship between the image data and a plurality of the corrected image data that can be formed from the image data is stored. To do.

  According to a fourth aspect of the present invention, in the image processing apparatus according to the third aspect of the present invention, the image processing apparatus includes a plurality of arithmetic units that form the corrected image data from the image data stored in the buffer.

The invention according to claim 5 is the image processing apparatus according to any one of claims 1 to 4,
The minimum unit is a line unit,
The buffer is an input line buffer capable of storing the image data with the minimum number of lines capable of forming the corrected image data.

  According to a sixth aspect of the present invention, in the image processing device according to any one of the first to fifth aspects, the image processing device stores the corrected image data of each of the imaging devices in units of lines. An output buffer is provided, and the corrected image data of the same object is simultaneously output from the output buffers in units of lines.

The invention described in claim 7
In an image processing method for processing image data acquired by imaging the same object with a plurality of imaging devices to form corrected image data,
Storing the image data in a buffer;
Determining whether the image data of the smallest unit capable of forming the corrected image data is stored in a buffer;
When it is determined that the minimum unit of image data that can form the corrected image data is stored in a buffer, the stored image data is parallelized to form the corrected image data.

  According to the first aspect of the present invention, when image data capable of forming corrected image data is stored in the buffer, formation of corrected image data obtained by performing parallel image processing from the image data is started. , The time to store the density data in the buffer and the time from the input of the image data to the image processing unit from the imaging device to the time of the interpolation calculation can be shortened. Thus, there is an effect that it is possible to shorten the image processing time until the image processing for forming the corrected image data is completed.

  According to the second aspect of the present invention, since the correspondence between the image data and the corrected image data that can be formed from the image data is stored in the correction table in advance, the correspondence between the image data and the corrected image data. Thus, it is possible to reduce the calculation time and form corrected image data, and the image processing time can be shortened.

  According to the third aspect of the present invention, a plurality of corrected image data can be formed using image data capable of forming certain corrected image data, so that the image processing time can be shortened. There is an effect.

  According to the fourth aspect of the present invention, since a plurality of arithmetic units for forming corrected image data are provided, there is an effect that a plurality of corrected image data can be formed at the same time to shorten the image processing time.

  According to the fifth aspect of the present invention, since image data can be stored in a line unit in a buffer instead of a frame buffer, the time for storing the image data is reduced, and image processing is performed after the image data is acquired. There is an effect that the time until the start can be shortened.

  According to the sixth aspect of the present invention, the corrected image data can be stored in line units in the output line buffer, so that the corrected image of the same portion of the same object is obtained from the main imaging device and the sub imaging device at the same timing. There is an effect that data can be output in line units.

  According to the seventh aspect of the present invention, when image data capable of forming corrected image data is stored in the buffer, formation of correction data obtained by performing parallel image processing from the image data is started. The time to store density data in the buffer and the time from input of image data from the imaging device to the image processing unit to the time of interpolation calculation can be shortened. There is an effect that it is possible to shorten the image processing time until the image processing for forming the corrected image data is completed.

  Embodiments of an image processing apparatus according to the present invention will be described below with reference to the drawings. However, the scope of the invention is not limited to the illustrated examples.

  As shown in FIG. 1, the image processing device 1 includes an image processing unit 2 that processes corrected image data using a plurality of imaging devices, that is, a main imaging device 4 and a sub imaging device 5 to form corrected image data. A post-stage image processing unit 3 that processes the corrected image data formed by the image processing unit 2 and forms image data for recognizing an object, for example, is provided.

  The image processing unit 2 is provided with a main imaging device 4 and a sub imaging device 5 that image the same object to form image data. The image data acquired by the main imaging device 4 is reference image data for obtaining a stereo image, and the image data captured by the sub-imaging device 5 corresponds to the reference image data acquired by the main imaging device 4. Reference image data.

  The main imaging device 4 and the sub imaging device 5 are each provided with an imaging lens 6 made of glass, plastic, or the like. The imaging lens 6 is used to image the same object. On the optical path of the imaging lens 6, a plurality of imaging elements 7 as light receiving elements are arranged corresponding to the pixels, and image data is generated according to the amount of light received by each imaging element 7 and output to the image processing unit 2. It is supposed to be. This image data is generated in units of pixels, and is composed of coordinate data indicating pixel coordinates and density data indicating pixel density.

  As the imaging element 7, a complementary metal oxide semiconductor (CMOS), a charge coupled device (CCD), or the like can be used.

  As shown in FIG. 2, the image processing unit 2 is provided with a control unit 8 that controls the entire image processing unit 2 so as to be connected to each unit constituting the image processing unit 2.

  In addition, the image processing unit 2 includes an image collimating unit 9 that performs corrected image processing on the respective image data acquired by the main imaging device 4 and the sub imaging device 5 to form corrected image data. Each of the device 4 and the sub imaging device 5 is provided in one-to-one correspondence. Here, the parallelization is an image acquired when the main imaging device 4 and the sub imaging device 5 are set up in parallel and photographed image data acquired by the main imaging device 4 and the sub imaging device 5. Sorting into data. The corrected image data is composed of coordinate data and density data of pixels that have undergone parallel image processing.

  Hereinafter, the details of the image collimating unit 9 provided corresponding to the main imaging device 4 will be described. However, the details of the image collimating unit 9 provided corresponding to the sub-imaging device 5 are not particularly described. It is the same.

  The image collimating unit 9 includes coordinate data of pixels after being subjected to parallel image processing using image data acquired by the main imaging device 4 (hereinafter referred to as corrected pixels), and after the correction. A correction table 10 is provided in which the correspondence relationship with the coordinate data of the pixel before the parallel image processing (hereinafter referred to as a pixel before correction) is stored. This correspondence is detected in advance by image data acquired by the main imaging device 4 and corrected image data obtained by parallelizing the image data, and stored in the correction table 10. Since the coordinate data of the pixel before correction and the coordinate data of the pixel after correction do not necessarily match the coordinate data of the image data acquired by imaging, the coordinate data of the image data acquired by imaging and the coordinates of the pixel before correction Correspondences between the data and the corrected pixel coordinate data are detected in advance and stored in the correction table 10. The correspondence between these coordinate data differs depending on the positional relationship between the main imaging device 4 and the sub imaging device 5 and the interpolation method used for interpolation. If there is distortion in the image data acquired by the imaging lens 6 included in the main imaging device 4, the correction table 10 includes the coordinate data of the pixel before correction and the image processing parallelized to the pixel before correction. The correspondence relationship with the corrected pixel coordinate data after correcting the distortion is calculated and stored in advance.

  Further, when the coordinate data of the pixel before correction does not match the coordinate data of the pixel acquired by imaging, the pixel before correction is located by using the density data of the pixel that is located around the pixel before correction and acquired by imaging. Interpolation-related data for calculating the density data is detected in advance and stored in the correction table 10. The density data of the pixel before correction obtained by this interpolation relationship becomes the density data of the pixel after correction.

For example, as shown in FIG. 3, the interpolating relationship here refers to the density data of the pixel a before correction and the density data of the four pixels p1, p2, p3, and p4 that are present around and acquired by imaging. Relationship. In the present embodiment, the pixels acquired by imaging of the main imaging device 4 are present at positions where the X and Y axes intersect with an integer value, such as p1, p2, p3, and p4. In the present embodiment, the density data of the pixel a is calculated by an interpolation calculation using the following calculation formula (1).
a (x, y) = p1 (n + 1, m + 1) (1-α) (1-β) + p2 (n + 2, m + 1) α (1-β) + p3 (n + 1, m + 2) (1-α) β + p4 (n + 2, m + 2) αβ (1)
here,
a (x, y) is the density data at coordinates (x, y) p1 (n + 1, m + 1) is the density data at coordinates (n + 1, m + 1) p2 (n + 2, m + 1) is the density at coordinates (n + 2, m + 1) Data p3 (n + 1, m + 2) is density data at coordinates (n + 1, m + 2) p4 (n + 2, m + 2) is density data at coordinates (n + 2, m + 2) α is the difference between x and (n + 1), x Directional weighting β is the difference between y and (m + 1), and is the weighting in the y direction.

A specific example of the relationship between the corrected image data and the correction table 10 in the present embodiment will be described below.
For example, as shown in FIG. 4, corrected pixels a1, b1, and c1 located at coordinates (n, m), coordinates (n + 1, m), and coordinates (n, m + 1) are respectively coordinated (n + 1 + α1, m + 1 + β1). ), Coordinates (n + 1 + α2, m + 1 + β2) and coordinates (n + 1 + α3, m + 1 + β3) corresponding to uncorrected pixels a2, b2, and c2. The correspondence between the coordinate data of the pixel before correction and the coordinate data of the pixel after correction, the arithmetic expression used for the interpolation calculation, and the weights α1 to α3 and β1 to β3 possessed by the individual pixels before correction are the correction table. 10 is stored.
Then, the density data of the uncorrected pixel that is interpolated using the arithmetic expression (1) becomes the density data of the corrected pixel corresponding to the pixel before correction. For example, the density data of the density data a2 (n + 1 + α1, m + 1 + β1) of the pixel a2 is calculated by interpolation using the arithmetic expression (1), and the density data becomes the density data a1 (n, m) of the pixel a1. .

  Returning to FIG. 2, the image collimating unit 9 is provided with an input line buffer 11 for temporarily storing density data among the image data output from the main imaging device 4 in units of lines. The line unit is a minimum unit capable of forming corrected image data when processing the image data to form corrected image data. The input line buffer 11 can store at least density values for the number of lines necessary to form an interpolation relationship described later. The required number of lines varies depending on the density data interpolation method to be used. In this embodiment, since the linear interpolation method that performs linear interpolation using the density data of four pixels existing around the pixel for which density data interpolation is performed is used, the input line buffer 11 has two lines. Pixel density data can be stored at a minimum.

  The input line buffer 11 is provided with an input counter 12 connected to the input line buffer 11 for counting when image data is stored in the input line buffer 11. The input counter 12 includes an input x counter 13 that counts the x position of the input line buffer 11 and an input y counter 14 that counts the y position of the input line buffer. By using the input x counter 13 and the input y counter 14, individual density data can be stored at an appropriate position in the input line buffer 11.

  Further, the input line buffer 11 is connected to an interpolation relation generation unit 15 that performs an interpolation operation on the density data of the pixels stored in the input line buffer using the interpolation relation stored in the correction table 10.

The interpolation relation generation unit 15 performs an interpolation operation using the density data stored in the input line buffer 11 and the arithmetic expression and weight stored in the correction table 10 to calculate the density data of each pixel. An interpolation calculation unit 16 is provided.
For example, the interpolation calculation unit 16 can perform distortion correction that can be calculated based on the density data of four surrounding pixels such as a2, b2, and c2 surrounded by four pixels p1, p2, p3, and p4 in FIG. It is provided corresponding to the maximum number of previous pixels, and in this embodiment, three are provided. The number of interpolation calculation units 16 differs depending on the distortion caused by the lens and the imaging system and the interpolation method used for interpolation, and is calculated in advance based on the correspondence between the coordinate data before correction and the coordinate data after correction, and the coordinate data of the pixel to be acquired. Is done.

  The interpolation relation generation unit 15 is provided with an interpolation counter 17 that manages coordinate data before correction of the pixels that are subjected to interpolation calculation by the interpolation calculation unit 16, corresponding to the interpolation relation generation unit 15. The interpolation counter 17 includes an interpolation x counter 18 that manages the x-coordinate before correction of the pixel that is interpolated by the interpolation calculation unit 16, and an interpolation that manages the y-coordinate before correction of the pixel that is also subjected to the interpolation calculation. and y counter 19.

The interpolation calculation unit 16 is connected to an output line buffer 20 that temporarily stores the density data formed by the interpolation calculation and outputs the density data stored in the subsequent image processing unit 3 in units of lines. The output line buffer 20 is provided corresponding to each of the main imaging device 4 and the sub imaging device 5, and the timing at which the reference image data captured by the main imaging device 4 is output to the subsequent image processing unit 3; The timing when the reference image data captured by the sub imaging device 5 and corresponding to the standard image data is output to the subsequent image processing unit 3 coincides with the timing.
The output line buffer 20 can store at least the density data of the same number of lines as the input line buffer 11. When the timing at which the reference image data corresponding to the standard image data is output is later than the timing at which the standard image data is output, the output line buffer 20 provided for the main imaging device 4 includes an input line buffer. In addition to the density data of the number of lines equal to 11, the density data for the number of lines whose output is delayed can be stored.

  The output line buffer 20 counts when density data formed by the interpolation calculation unit 16 is stored in the output line buffer 20 or is output from the output line buffer 20 to the subsequent image processing unit 3. Are provided corresponding to the output line buffer 20. The output counter 21 includes an output x counter 22 that counts the x position of the output line buffer 20 and an output y counter 23 that counts the y position of the output line buffer 20.

  Next, the operation of the image processing apparatus 1 according to this embodiment will be described.

  When an image is captured using the main imaging device 4 and the sub imaging device 5, the image sensor 7 on the optical path of the imaging lens 6 of the main imaging device 4 and the sub imaging device 5 converts the image data into image data according to the amount of received light. Therefore, the image data by the main imaging device 4 and the sub imaging device 5 are respectively generated.

  When the image data generated by the main imaging device 4 and the sub imaging device 5 is input to the image parallelizing unit 9 connected to the main imaging device 4 and the sub imaging device 5, the control unit 8 configures the image data. Density data among the data is temporarily stored in the input line buffer 11. Then, the control unit 8 refers to the correction table 10 and detects a pixel before correction that can perform an interpolation calculation from the density data stored in the input line buffer 11. Then, the control unit 8 causes the interpolation relation generation unit 15 to output an interpolation calculation formula, a weighting, and the like corresponding to the pixel before correction from the correction table 10. At the same time, the control unit 8 detects the correspondence relationship of the image data calculated by the interpolation relationship generation unit 15 from the correction table 10 and causes the interpolation counter 17 to manage the coordinate data of the pixel before correction for interpolation calculation.

  More specifically, when the density data p1, p2, p3, and p4 constituting the acquired image data are stored in the input line buffer 11, the control unit 8 can perform interpolation using these density data before correction. It is detected from the correspondence relationship of the correction table 10 that the pixels are a2, b2, and c2, and the interpolation calculation formula and weight used for each pixel are output from the correction table 10 to the interpolation relationship generation unit 15. Then, the interpolation calculation unit 16 performs an interpolation calculation corresponding to each pixel to form density data.

For example, the interpolation calculation of a2 is performed using the following calculation formula (2), and the corrected pixel density data is calculated.
a2 = (1-α1) (1-β1) p1 + α1 (1-β1) p2 + (1-α1) β1p3 + α1β1p4 (2)
here,
a2 is the density data p1 of the coordinate a2, p1 is the density data of the coordinate p1, p2 is the density data of the coordinate p2, p3 is the density data of the coordinate p3, p4 is the density data of the coordinate p4 α1 and β2 are weights of a2. .

  When calculating the density data of a2, the control unit 8 causes the interpolation x counter 18 to manage (n + 1 + α1) indicating the x coordinate out of the coordinate data (n + 1 + α1, m + 1 + β1) of the pixel a2, and the y coordinate. (M + 1 + β1) indicating that the interpolation y counter 19 manages.

  Similarly, each interpolation calculation unit 16 performs interpolation calculation of b2 and c2, and calculates density data of pixels after correction of b2 and c2. The calculated density data of a2, b2, and c2 become density data of pixels after correction of parallelization (corrected pixels) and density data of a1, b1, and c1, respectively.

The control unit 8 temporarily stores the corrected pixel density data calculated by the interpolation calculation at an appropriate position in the output line buffer 20 using the output counter 21. This appropriate position is a position rearranged to image data that has been subjected to parallelization processing, which is an arrangement of image data acquired when the main imaging device 4 and the sub-imaging device 5 are installed completely in parallel. When the imaging lens 6 included in the main imaging device 4 and the sub imaging device 5 has distortion, it is also the position of the image data after correcting the distortion.
The control unit 8 detects corrected pixel coordinate data corresponding to the corrected pixel density data output from the interpolation calculation unit 16 from the correction table, and outputs the x coordinate of the coordinate data to the output x counter 22. And the output y counter 23 manages the y coordinate of the coordinate data. Then, the control unit 8 controls the output counter 21 to output and store the corrected pixel density data at the position of the output line buffer 20 corresponding to the corrected pixel coordinate data.

  Further, the control unit 8 outputs the density data stored in the output line buffer 20 connected to each of the main imaging device 4 and the sub imaging device 5 to the subsequent image processing unit 3 using the output counter 21 in units of lines. To do. At this time, the control unit 8 performs control so that the timing at which the standard image data is output matches the timing at which the reference image data corresponding to the standard image data is output.

  And the control part 8 processes the correction | amendment image data in the back | latter stage image process part 3, and forms the image data which recognizes an object, for example.

  As described above, according to the present embodiment, when image data capable of forming corrected image data is stored in the input line buffer 11, the corrected image data is formed by performing parallel image processing from the image data. Therefore, it is possible to reduce the time for storing the density data in the input line buffer 11 and the calculation time for the correspondence between the image data and the corrected image data, thereby forming the corrected image data. Time can be shortened.

  In the present embodiment, the image parallelizing unit 9 is provided with the correction table 10 in which the correspondence relationship between the coordinate data of the corrected pixel and the coordinate data of the pixel before correction is provided. However, the present invention is not limited thereto, and the correction table may not be provided in the image processing unit, and the correspondence relationship between the coordinate data of the corrected pixel and the coordinate data of the pixel before correction may be detected.

In this embodiment, an example in which the line unit is the minimum unit in which the corrected image data can be formed is shown. However, the present invention is not limited to this, and a pixel unit or an image unit may be used.
Correspondingly, in this embodiment, an example in which the input line buffer 11 for storing density data in units of lines is provided, but the present invention is not limited to this, and input for storing density data in units of pixels or images is provided. A line buffer may be provided.
Furthermore, in the present embodiment, an example in which the output line buffer 20 that outputs the stored density data in units of lines is shown. However, the present invention is not limited to this. An output line buffer that stores data in units or images may be provided.

  Furthermore, in this embodiment, although the example which provided the single sub imaging device 5 was shown, it is not limited to this, The structure which provides a some sub imaging device may be sufficient.

  In the present embodiment, the example in which the interpolation calculation unit 16 is provided corresponding to the maximum number of density data before correction that can be calculated at one time has been described. However, the present invention is not limited to this, and one interpolation calculation is performed. The configuration may be such that the interpolation calculation is sequentially performed in the unit.

1 is a schematic diagram illustrating an overall configuration of an image processing apparatus according to an embodiment. It is the schematic which shows the structure of the image parallelization part in this embodiment. It is a figure which shows the relationship between the pixel which performs the interpolation calculation in this embodiment, and a surrounding pixel. It is a figure which shows the correspondence of the pixel before distortion correction in this embodiment, and the pixel after distortion correction.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image processing apparatus 2 Image processing part 3 Subsequent image processing part 4 Main imaging device 5 Sub imaging device 6 Imaging lens 7 Imaging element 8 Control part 9 Image parallelizing part 10 Correction table 11 Input line buffer 12 Input counter 15 Interpolation relation production | generation part 16 Interpolation Operation Unit 17 Interpolation Counter 20 Output Line Buffer 21 Output Counter

Claims (7)

A buffer for storing image data obtained by imaging the same object with a plurality of imaging devices;
An image processing apparatus comprising: an image processing unit that processes image data acquired by each of the imaging devices to form corrected image data;
The image processing unit starts processing to form the corrected image data by parallelizing the image data when the minimum unit of the image data capable of forming the corrected image data is stored in a buffer. An image processing apparatus.   The image processing unit includes a correction table that stores a correspondence relationship between the image data and the corrected image data that can be formed from the image data, and is stored in the buffer using the correspondence relationship stored in the correction table. The image processing apparatus according to claim 1, wherein the corrected image data is formed from the image data.   The image processing apparatus according to claim 2, wherein the correction table stores a correspondence relationship between the image data and a plurality of the corrected image data that can be formed from the image data.   The image processing apparatus according to claim 3, wherein the image processing unit includes a plurality of arithmetic units that form the corrected image data from the image data stored in the buffer. The minimum unit is a line unit,
The said buffer is an input line buffer which can store the said image data of the minimum line number which can form the said correction | amendment image data, The Claim 1 characterized by the above-mentioned. Image processing device.   The image processing apparatus includes an output buffer that stores the corrected image data of each imaging apparatus in units of lines, and outputs the corrected image data of the same object from the output buffers simultaneously in units of lines. The image processing apparatus according to any one of claims 1 to 5. In an image processing method for processing image data acquired by imaging the same object with a plurality of imaging devices to form corrected image data,
Storing the image data in a buffer;
Determining whether the image data of the smallest unit capable of forming the corrected image data is stored in a buffer;
An image processing method characterized in that the corrected image data is formed by parallelizing the stored image data when it is determined that the minimum unit of image data that can form the corrected image data is stored in a buffer. .

Images