Method and device for video image input, processing and output

FIELD: physics.

SUBSTANCE: this method comprises memorising the input raster video image as a flow of frames in the line input buffer. Said frames are splitted to micro blocs. The latter are compressed and stored in external memory. For processing, said micro blocs are retrieved from external memory, unclasped and written to internal memory. Raster macro blocs are formed and processed by appropriate processors.

EFFECT: efficient use of internal memory irrespective of processing algorithm type.

22 cl, 2 dwg

 

The invention relates to the field of data processing, and in particular to methods and devices for input, processing and output of video, and can be used in video surveillance systems and video Analytics.

Recently there has been active development of technologies that significantly improve the quality of input from various video sources. There is an increase in resolution and frame rate of the video stream that requires more equipment needed for processing and storing data, increase processing speed, bandwidth data paths between the processing modules and the external memory.

Modern level of development of technology VLSI (very large integrated circuits) does not allow to place on the crystal enough memory to store full-frame image high resolution. On the other hand, the capacity of the external memory interface is insufficient for the use of modern algorithms of processing of the video images of high resolution at high speed.

To solve the above problems apply data processing technology "in passing" using small buffers and row of macroblocks, decoding/encoding of information using special compression methods on which I exchange with external memory, block post-processing mode data memory-memory.

The known method and device for processing and storing the images in memory (application US 20070242077 A1), based on the modular principle, which calculates the size of the processed block in pixels (based on the performance of the processing devices) and the size of the block in megabytes (due to restrictions of the file system). When this is saved in the block must be an integer number of processed blocks. One of the processed block must get saved in the unit. The coordinates of the processed blocks is calculated from the center of the image, the coordinates of conserved blocks from the lower left corner of the image. Image save without compression.

The above-described method and apparatus is used within a computing system including processor, memory, network, disk, and so on, thus assume that the image is stored on disk, so the size of the block is large enough and is determined by the limitations of the file system. In modern systems RAM allows you to store multiple frames without storing on disk, while the greater problem is the need to reduce traffic, which in this approach is not solved.

There are the method and apparatus of image processing (patent RU 237638), in which image data to encode into modular units placed predefined way, when the processing area, according to the settings, can be formed from the data of one or more blocks (e.g., offset relative to the edges of the modular block to the left and up), and data related to processing, are placed in memory in order to handle the boundary region for processing.

These method and apparatus offer storage in memory only one line of modular units. However, when processing high-resolution images and large enough to implement complex processing algorithms size of the processed block of memory required increases, and the use of coded blocks of various sizes (JPEG, MPEG) does not allow for a quick search of any area of the image in memory, and stream processing method is the only one possible, that is not always the best treatment method.

In the application US 20100296745 A1 describes a method of image processing, namely compression block image with a certain number of pixels, consisting of several steps: 1) lossless compression, 2) if the block length after compression exceeds the preset value, the parameter is selected quantization and produce compression, steps 1) and 2) repeat until dormancy is the block length is equal to or less than a specified value, 3) in the memory writes the compressed block and the identifier to determine the quantization parameters.

The above method allows for data compression with the highest possible quality within a limited block size after compression, but iterational method has a negative impact on performance. The size of the compressed block is stored in memory, is not strictly defined value, therefore, to implement a quick search the desired area in memory (which is often relevant when solving video analysis tasks) require the storage of additional information about the correspondence between the coordinates of the regions in the source image and the address of their location in memory. In addition, does not address the use of this method in the system of processing video data.

Closest to the claimed invention is the image processing system described in the application US 20120033886 A1, which includes the input port data output port of the compressed data or the node store the compressed data memory instructions and the processing device scalable blocks, which receives and processes the blocks, performs coding based on wavelet transform and outputs the encoded stream into the system. The system further includes an input port of the compressed data and the output port is kodirovannykh data or the storage node of the decoded data. This system is chosen as the prototype of the claimed invention.

The system prototype is a General principle of software and/or hardware system, image processing involves dividing the image into blocks, which can be decoded for processing, processed in parallel, be encoded for transmission over the data paths and memory retention, it does not define the mechanisms of division into blocks for processing and storage guidelines for the storage and retrieval area in memory that is not regulated by the division of functions between software and hardware part of the system.

Objective of the claimed invention is a method and device input, processing and output video images, including raster video image of high resolution, which possess a small amount of internal memory on-chip and low bandwidth interface external memory, but at the same time provide a quick search of objects in the external memory and the efficient use of internal memory regardless of the type of processing algorithms, due to the transmission of video data between the device and external memory in a compressed form (at the level of microblocks), and also due to block processing of frames of the input image, retrieved from the internal memory (at the level of overlapping RA is throuyh macroblocks).

The problem is solved by creating a method of input, processing and output of the video image, in which the phase of the input raster image in the form of a stream of frames by row remain in the input buffer row, breaking the frames of the input video image input microblock, compress and store the input microblock in the external memory; the processing stage reads the input microblock from the external memory, open it and write the input microblock in the internal memory, thus forming a raster macroblocks, process raster macroblocks by processors, used to store the intermediate and final results of processing only the internal memory, smash processed raster macroblocks weekend microblock, compress and store the output microblock in the external memory in this form the frame of the output video; phase output line-by-line read microblog frame of the output image from the external memory, you open it, and writes the output microblock frame of the output image in the output string buffer from which to read line by line and pass to the recipient of the video.

In a preferred embodiment of the method frames of the input video image is divided into input microblock fixed size, the cat is who shrink with fixed odds, you get a compressed input microblock the same size, which remain in the external memory.

In a preferred embodiment of the method in the processing of raster macroblocks simultaneously load all raster macroblocks from several frames of the input video in the internal memory.

In a preferred embodiment, the method raster macroblocks overlap between them.

In a preferred embodiment of the method the loop macroblock raster repeat this a number of times, which is necessary for the processing is stored in the external memory with the input microblock frame of the input video image or its parts.

In a preferred embodiment, the method at the stage of processing performed several cycles of processing stored in the external memory with the input microblock frame of the input video image.

In a preferred embodiment, the method at the stage of processing performed several cycles of processing stored in the external memory with the input microblock frame of the input video image, in this case, applying different partitioning scheme frame of the input video image on a raster macroblocks.

The problem is solved by creating the input device, the processing and in the water video contains the input video data, the video processing modules and the output modules video made with the possibility of exchanging data with the external memory through the external memory controller and system connections, and the input video data is made in the form of the input controller, which is connected with the input string buffer, which is connected to the divisor input image, which is connected to the compression module input microblock, which is connected to the external connections, and the modules of the video processing is executed in the form of a collector of macroblocks, which is connected to the external system connections and decoder processing, which is connected with the internal memory, which is connected with the processors and divider processed macroblocks, which is connected to the compression module output microblocks, which is connected to the external connections, and the output video data is executed in the form of a collector on the video output, which is connected to the external connection and the output decoder, which is connected to the output string buffer, which is connected with the output controller, and an input controller configured to receive from the video source input raster video image in the form of a stream of frames and transmitting the input buffer of rows made with prob is the possibility progressive conservation of the video image; the divisor input image is configured to read data from the buffer rows and generate the input microblock; a compression module input microblock made with the possibility of compression of the input microblock fixed odds, with the formation of the compressed input microblock; the connection system and the external memory controller configured to transfer compressed input microblock to external memory is made with the possibility of recording and storing frames of the input video entirely in the form of a compressed input microblock; collector of macroblocks arranged to read from the external memory compressed input microblock; decoder processing performed with unclamping microblocks and write them in the internal memory, thus forming a raster of macroblocks; processors configured to process raster macroblock located in the internal memory, while using only internal memory for storing processing results and intermediate results of the processing; a divider processed macroblocks is performed by segmenting the processed raster macroblocks weekend microblock; a compression module output microblock made with the possibility of compression of the output microblock; system connection & controller external PA is ATI configured to send compressed output microblock the external memory, which is made with the possibility of recording and storing the compressed output microblocks, which form the frame of the output video; collector output video made with the possibility of reading line by line of the compressed output microblock frame of the video output from an external memory; an output decoder configured to unclamping line-by-line compressed output microblock frame of the video output and write it to the output string buffer made with the possibility progressive save the video output; an output controller configured to line-by-line reading of data from the output buffer of rows and transmitting the output raster of the video image in the form of a stream of frames to the recipient of the video.

In a preferred embodiment, the device input buffer row is made with the possibility of a line-save input video, while recording new lines in place of the oldest lines when filling the input buffer rows.

In a preferred embodiment, the device divisor of the input image is configured to read data from the buffer rows and generate the input microblock, the size of the input microblock vertically equal to or less than the number of rows in the input string buffer.

In preferably the m embodiment of the device, the compression module input microblock made with the possibility of compression of the input microblock fixed odds thus the formation of the compressed input microblock the same size, and as compression methods use lossy compression or by omitting the insignificant information, namely the dense packing of the component or decrease the capacity component of the pixel according to a logarithmic law or discrete cosine transform with rejection minor component.

In a preferred embodiment, the device module compression input microblock made with the possibility of compression of the input microblock fixed odds, with the formation of the compressed input microblock the same size, and as a method of partial compression using JPEG lossy compression: splitting microblock size of 8 by 8 pixels, discrete cosine transform, quantization of the input microbiota, as a result of receiving input microblock fixed length and reduced size in relation to the original.

In a preferred embodiment, the device decoder processing performed with unclamping microblocks and write them in the internal memory, thus forming a raster of macroblocks, with each macroblock raster is chosen so that the raster macroblock consisted of integer input microblock; fully, the key intermediate processing results, was suitable for placement in the internal memory; was suitable for processing independently from the rest of the frame of the input video image.

In a preferred embodiment, the device processor is designed as a device selected from a set that includes a scalar processor, a vector processor, hardware accelerator.

In a preferred embodiment, the device compression method used in the compression module output microblock, similar to the method of compression used in the compression module the input microblock.

In a preferred embodiment, the device processors configured to process raster macroblock, while simultaneously load a bitmap macroblocks from several frames of the input video image.

In a preferred embodiment, the device processors made with the possibility of repetition of the loop raster macroblock, the number of times that you want to process stored in the external memory frame of the input video image or its parts.

In a preferred embodiment, the device processors made with the possibility of repetition of the loop raster macroblock, the use of in each cycle of different partitioning schemes the frame of the input image on a raster macroblocks.

In a preferred embodiment, the device processors configured to process raster macroblock various processing procedures with saving intermediate results in the external memory.

In a preferred embodiment, the device bitmap macroblocks overlap between them, while the overlap area is determined by the processing algorithm raster macroblock in processing.

In a preferred embodiment, the device bitmap macroblock consists of a workspace and a boundary data area, located around the workspace, while the boundary region data of one raster macroblock is part of the working area neighboring macroblock raster.

In a preferred embodiment, the device output buffer row is made with the possibility of a line-save the video output while recording new lines in place of the oldest lines in the output buffer is full rows.

For a better understanding of the claimed invention the following is a detailed description with appropriate graphics.

Fig.1. Block diagram of the input, processing and output video made according to the invention.

Elements:

1 - input control is EP (IICtr);

2 - input buffer rows (ILBuff);

3 is a divisor of the input image (ISegm);

4 - compression module input microblock (IEnc);

5 - system connections (Intrconnect);

6 - external memory controller (MemCtr);

7 - external memory (ExtMem);

8 - collector macroblocks (PAssm);

9 - decoder processing (PDec);

10 - internal memory (VRAM);

11 - processors (Proc1 - ProcN);

12 is a divisor of processed macroblocks (PSegm);

13 is a compression module output microblocks (PEnc);

14 - collector output video (OAssm);

15 - output decoder (ODec);

16 - output string buffer (OLBuff);

17 - output controller (OLCtr).

Fig.2. The block diagram of the processing chain video, made according to the invention.

Consider a General embodiment of the inventive method and device input, processing and output video image (Fig.1-2).

The claimed device input, processing and output video image (Fig.1) is composed of modules 1 to 4 input modules 8, 9, 12, 13 and processing modules 14-17 output video data that communicate with the external memory 7 is controlled by the controller 6 of external memory via 5 connections. The claimed method of input, processing and output of the video image based on two-level splitting of the original image on compressible microblock used as storage elements and data exchange m is waiting for the external memory 7 and the internal memory 10, and overlapping raster macroblocks used as elements on which exercise independent processing of the data in the internal memory 10.

The input video image received from an external source (CMOS sensor, the camcorder, the transmission line), remain in the input buffer row 2, then split the input microblock fixed size, which is compressed with a fixed ratio so that the compressed input microblock also have the same size. Compressed input microblock save in external memory 7, designed for the storage required to process the number of frames, so that they could read and write, with no one to calculate the address of the data on the coordinate input microblock in the frame of the input image.

For processing video frames compressed input microblock according to a certain algorithm read from the external memory 7, open it and write to the internal memory 10, thus forming a raster macroblocks. The processors 11, which may be scalar, vector processors or specialized hardware accelerators, process each macroblock raster, used to store the intermediate and final results only internal memory 10. For processing raster macroblock can provide the required simultaneous loading bitmap macroblocks from several frames of the input video image, however, they all must be fully allocated in the internal memory 10. After processing the raster macroblock is divided into a weekend microblock that compress and store in the external memory 10, thus forming a frame of the output video image consisting of a compressed output microblock. The loop raster macroblock is repeated the number of times required for the processing is stored in the external memory frame of the input video image or parts of interest. Raster macroblocks can overlap, and the overlap is determined by the processing algorithm. For a full treatment of a frame of the input video may take several cycles of the processing of the video frame, wherein in each cycle can be applied to various circuit breaker frame to bitmap macroblocks.

To output the fully processed frame of the compressed video output microblock frame of the output image line by line read from the external memory 10, open it and write to the output buffer 16 rows, where each line read, and passed to the receiver or video display, the transmission line).

Let us consider in more detail the work of the best embodiments of the claimed method and device input, processing and output video image (Fig.1-2).

The input raster movie and the imagination (the flow rate) is taken from the video source (CMOS-sensor, camcorder, the transmission line) through the input controller 1 and the line remain in the input buffer row 2, suitable for the storage of small (e.g., 4-16) number of rows. When filling the input buffer row 2 new row record for most old rows.

The divisor input image 3 reads data from the input buffer lines 2 and forms the input microblock small size (for example, 4x4, 4x8 or 16x16 pixels). The size of the input microblock vertically equal to or less than the number of rows in the input buffer row 2.

Module 4 compression input microblock compresses the input microblock with a fixed ratio so that the compressed input microblock also have the same size. As methods of compression can be lossy or discarding unimportant information, such as: dense packing component; reducing the bit component pixel logarithmic law; discrete cosine transform with a drop minor component.

For compression microblocks can be partially used the JPEG lossy compression. This algorithm involves partitioning into microblock size of 8 by 8 pixels and the subsequent discrete-cosine transformation. The converted microblocks is subjected to quantization (i.e., the width of some comp the element is reduced or even zero), then get microblocks fixed length and reduced size in relation to the original. This data compression stop, and microblock Packed sequentially, one after the other.

To obtain the brightness of the pixel at known coordinates (x,y) first define the input microblocks, which contains the pixel - to do this, divide the coordinate value on the size microbiota in pixels (x/8, y/8). Knowing the size of the Packed microbiota in bits (M), and determine its position in memory and read ((round(x/8)+round(y/8)*W/8*)M, where W is the width of the frame in pixels). To read the data used dequantization (multiplied by the appropriate coefficients components microbiota). Do a reverse discrete cosine transformation. In the resulting microplate image to select the pixel.

The above-described compression method of microblocks set forth to illustrate, there may be other methods of compression microblocks, allowing to receive compressed microblock the same size.

Compressed input microblock through the system 5 connections and the controller 6 external memory save in external memory 7, designed to hold the required number of video frames as a whole. Compressed input microblock feature in the external memory 7 so that in the future they could read and write, independently computing dreadinny coordinate input microblock in the frame of the input video (e.g., series-continuous).

For processing frames of the input video image stored in the external memory 7 in the form of a compressed input microblock, collector 8 macroblocks for a given algorithm reads from the external memory 7 compressed input microblock, which then open the decoder processing 9 and recorded in the internal memory 10, thus forming a raster macroblocks.

The raster size of the macroblock is chosen so that it: consists of an integer input microblock; fully, including the intermediate results of processing, could be placed in the internal memory 10; may be handled independently from the rest of the frame of the input video image.

The CPU 11 of the processing process raster macroblock located in the internal memory 10, while using only the internal memory 10 for storing processing results and intermediate results. As processors 11 processing can be applied to scalar, vector processors or dedicated hardware accelerators.

After completing the processing of the divider 12 is processed macroblock raster raster splits the macroblock, which is the result of processing on the output microblock that are compressed using the compression module 13 output microblock. The compression method in the compression module 13 output microblock is similar to the used in module 4 compression input microblock. Compressed output microblock save in external memory 7, in this form the frame of the output video as compressed output microblock.

For processing raster macroblock may require simultaneous loading bitmap macroblocks from several frames of the input video. In this case, the size of the raster macroblocks from different frames should also choose based on the requirements similar to the requirements for raster macroblocks generated by the decoder processing 9.

Raster macroblocks may overlap each other, providing a processing edge regions of the raster macroblock. The overlap area is determined by the processing algorithm and contains an integer number of input microblock.

The above loop raster macroblock can be repeated as many times as necessary to process stored in the external memory 7 frame of the input video, in whole or in part, of interest.

For full processing of the frame may take several cycles of the video processing various processing procedures with saving intermediate results in the external memory 7, for each fragment of the algorithm can be applied to various circuit breaker frame input videosor is the position on the raster macroblocks.

In Fig.2 shows an example of a chain of video processing. The input video is processed by the first processing procedure, for this, the image is divided into a raster macroblock type MBO, which are processed by the first processing procedure. The resulting processing raster macroblock type MV1 without saving in the external memory 7 is processed by the second processing procedure. The resulting processing raster macroblock type MD2 also without saving in the external memory 7 is processed by the third processing procedure, and the resulting processing raster macroblock type MV remain in the external memory 7. The intermediate video image read from the external memory 7, is divided into a raster macroblock type MV' (different from type MW) that handle the fourth procedure, thus form a processed output video image, which is recorded in the external memory 7.

Raster macroblock consists of a workspace and a boundary data area, located around the work area. The boundary area is part of the working area neighboring macroblock raster. Below are examples select the size of the bitmap macroblock for different processing procedures.

Example # 1: To apply filters with different aperture on the input image is agenies choose the filter with the largest aperture. Based on the size of the aperture calculate the minimum size of a macroblock raster in pixels and rounded upward rounding to integer values of the raster size of the macroblock in microblock.

Example # 2: consistent application of filters with different aperture size of the boundary region is the sum of the apertures of filters.

Example No. 3: When small apertures filters the block size is chosen for reasons of reducing the overhead of data - the more the workspace with respect to the boundary region, the less repeated readings of the data.

To output the fully processed frame of video collector 14 of the video output reads line-by-line compressed output microblock frame of the video output from the external memory 7, which then open the output decoder 15 and write to the output buffer row 16 designed to hold a small (e.g., 4-16) the number of rows of the video output. When the output buffer is full row 16 new output microblock record in place of the oldest output microblock.

The output controller 17 reads the data line by line from the output buffer rows 16 and transmits the generated output raster image (the flow rate) to the recipient of the video display, the transmission line).

The claimed method and device is input, processing and displaying bitmap images using the internal memory 10, the volume of which may be less than the amount of a particular frame of video. High processing performance in terms of bandwidth limitations of the interface with the external memory 7 is achieved through: the transfer of video data between the device and the external memory 7 in a compressed form (at the level of microblocks); block-by-block processing of the video frames of the input image that is extracted from the internal memory 10 (at the level of overlapping bitmap macroblocks).

The greatest efficiency of the device is achieved when parallel execution is the following: the processing of the current macroblock raster; sending data for processing of the next raster macroblock from the external memory 7 internal memory 10; the sending of the data of the previous processed raster macroblock from the internal memory 10 to the external memory 7. At the same time, processing of the current macroblock raster must be equal to the time of sending data.

Although the above described embodiment of the invention has been set forth to illustrate the present invention, the experts it is clear that various modifications, additions and substitutions, without departing from the scope and meaning of the present invention disclosed in the accompanying claims

1. Method of input, processing and output of the video image, in which the phase of the input raster image in the form of a stream of frames by row remain in the input buffer row, breaking the frames of the input video image input microblock, compress and store the input microblock in the external memory; the processing stage reads the input microblock from the external memory, open it and write the input microblock in the internal memory, thus forming a raster macroblocks, process raster macroblocks by processors, used to store the intermediate and final results of processing only the internal memory, split processed raster macroblocks weekend microblock, compress and store the output microblock in the external memory in this form the frame of the output video; phase output line-by-line read microblog frame of the output image from the external memory, you open it, and writes the output microblock frame of the output image in the output string buffer from which to read line by line and pass to the recipient of the video.

2. The method according to p. 1, characterized in that the frames of the input video image is divided into input microblock fixed size, which is compressed with a fixed factor, while the floor is with compressed input microblock the same size, which store in the external memory.

3. The method according to p. 1, characterized in that the processing of the raster macroblocks simultaneously load all raster macroblocks from several frames of the input video in the internal memory.

4. The method according to p. 1, characterized in that the raster macroblocks overlap between them.

5. The method according to p. 1, characterized in that the loop macroblock raster repeat this a number of times, which is necessary for the processing is stored in the external memory with the input microblock frame of the input video image or its parts.

6. The method according to p. 1, characterized in that the processing stage to perform several cycles of processing stored in the external memory with the input microblock frame of the input video image.

7. The method according to p. 6, characterized in that the processing stage to perform several cycles of processing stored in the external memory with the input microblock frame of the input video image, in this case, applying different partitioning scheme frame of the input video image on a raster macroblocks.

8. The device input, processing and output of the video image containing the input video data, the video processing modules and the output modules video made with the possibility of exchanging data with the external memory through the troller external memory and system connections, moreover, the input video data is made in the form of the input controller, which is connected with the input string buffer, which is connected to the divisor input image, which is connected to the compression module input microblock, which is connected to the external connections, and the modules of the video processing is executed in the form of a collector of macroblocks, which is connected to the external system connections and decoder processing, which is connected with the internal memory, which is connected with the processors and divider processed macroblocks, which is connected to the compression module output microblocks, which is connected to the external system connections, and the output video data is executed in the form of a collector on the video output, which is connected to the external connection and the output decoder, which is connected to the output string buffer, which is connected with the output controller, and an input controller configured to receive from the video source input raster video image in the form of a stream of frames and transmitting the input buffer of rows made with the possibility of a line-save video; the divisor of the input image is configured to read data from the buffer rows and generate the input microblock; a compression module input microblock ispolnen with the possibility of compression of the input microblock fixed odds thus the formation of the compressed input microblock; the connection system and the external memory controller configured to transfer compressed input microblock to external memory is made with the possibility of recording and storing frames of the input video entirely in the form of a compressed input microblock; collector of macroblocks arranged to read from the external memory compressed input microblock; decoder processing performed with unclamping microblocks and write them in the internal memory, thus forming a raster of macroblocks; processors configured to process raster macroblock located in the internal memory when using only the internal memory for storing processing results and intermediate results of the processing; a divider processed macroblocks is performed by segmenting the processed raster macroblocks weekend microblock; a compression module output microblock made with the possibility of compression of the output microblock; the connection system and the external memory controller is configured to send compressed output microblock the external memory, which is configured to record and store the compressed output microblocks, which form the frame of the output video; the collector of the output video is made with the considerations applying a read line by line of the compressed output microblock frame of the video output from an external memory; the output of the decoder is arranged to unclamping line-by-line compressed output microblock frame of the video output and write it to the output string buffer made with the possibility progressive save the video output; an output controller configured to line-by-line reading of data from the output buffer of rows and transmitting the output raster of the video image in the form of a stream of frames to the recipient of the video.

9. The device under item 8, characterized in that the input string buffer is made with the possibility of a line-save input video, while recording new lines in place of the oldest lines when filling the input buffer rows.

10. The device under item 8, characterized in that the divider input image is configured to read data from the buffer rows and generate the input microblock, the size of the input microblock vertically equal to or less than the number of rows in the input string buffer.

11. The device under item 8, wherein the compression module input microblock made with the possibility of compression of the input microblock fixed odds, with the formation of the compressed input microblock the same size, and as compression methods use lossy compression or by omitting the insignificant information, and the military: the dense packing of the component or decrease the capacity component of the pixel according to a logarithmic law or discrete cosine transform with rejection minor component.

12. The device under item 8, wherein the compression module input microblock made with the possibility of compression of the input microblock fixed odds, with the formation of the compressed input microblock the same size, and as a method of partial compression using JPEG lossy compression: splitting microblock size of 8 by 8 pixels, discrete cosine transform, quantization of the input microbiota, as a result of receiving input microblock fixed length and reduced size in relation to the original.

13. The device under item 8, characterized in that the decoder processing performed with unclamping microblocks and write them in the internal memory, thus forming a raster of macroblocks, with each macroblock raster is chosen so that the raster macroblock consisted of integer input microblock; fully, including intermediate processing results, was suitable for placement in the internal memory; was suitable for processing independently from the rest of the frame of the input video image.

14. The device under item 8, wherein the processor is made in the form of the device selected from a set that includes a scalar processor, a vector processor, hardware accelerator.

15. The device is .8, characterized in that the compression method used in the compression module output microblock, similar to the method of compression used in the compression module the input microblock.

16. The device under item 8, wherein the processors configured to process raster macroblock, while simultaneously load a bitmap macroblocks from several frames of the input video image.

17. The device under item 8, wherein the processors configured to repeat the processing cycle of the raster macroblock, the number of times that you want to process stored in the external memory frame of the input video image or its parts.

18. The device under item 8, wherein the processors configured to repeat the processing cycle of the raster macroblock, the use of in each cycle of different partitioning schemes frame of the input image on a raster macroblocks.

19. The device under item 8, wherein the processors configured to process raster macroblock various processing procedures with saving intermediate results in the external memory.

20. The device under item 8, characterized in that the raster macroblocks overlap between them, while the overlap area is determined by the processing algorithm R is strologo macroblock in processing.

21. The device under item 8, characterized in that the raster macroblock consists of a workspace and a boundary data area, located around the workspace, while the boundary region data of one raster macroblock is part of the working area neighboring macroblock raster.

22. The device under item 8, characterized in that the output string buffer is made with the possibility of a line-save the video output while recording new lines in place of the oldest lines in the output buffer is full rows.



 

Same patents:

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to image processing, particularly to a device and a method of processing images, which enable to classify composition of an input image. The technical result is achieved through a device which includes a separation line detection module, which can be configured to receive an input image and detect a separation line which separates two sides of the input image, a classification module which can be configured to classify an input image based on the degree of symmetry and the separation line; the classification module can also be configured to generate a classification signal to provide at least one of display or storage of classification.

EFFECT: automating classification of composition of an input image into detailed composition structures based on the degree of symmetry and a separation line.

19 cl, 46 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to computer engineering and can be used in video sequence analysis and processing systems and digital television. The method of detecting defects on video signals includes analysing differential images of adjacent frames. Binarisation, expansion and joining operations are then applied to the differential images of adjacent frames. Non-zero values are analysed on the obtained arrays, for which a defectiveness decision is made on the initial frames based on dispersion of the initial values.

EFFECT: detecting the position of defects on video signals with insufficient prior information about statistical characteristics of additive noise and useful component function.

2 cl, 1 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to multispectral reading photosensitive devices for reading sub-sampled data of photosensitive pixels in large-scale array photosensitive chips. The multispectral photosensitive device and pixel sampling method include: a first merging process for merging and sampling two adjacent pixels in the same row and the different columns, or in different rows and the same column or different rows and different columns in a pixel array to obtain sampling data of a first merged pixel; a second merging process for merging and sampling the sampling data of the first merged pixel obtained in the first merging process to obtain sampling data of a second merged pixel; and a third merging process; sampling data of a third merged pixel are obtained in a digital space by colour changing and image zooming methods.

EFFECT: enabling sub-sampling with high output and efficient image processing.

18 cl, 26 dwg

FIELD: personal use articles.

SUBSTANCE: system includes movie server configured for movie show and movie time code transmission; central server configured to transmit movie reference time; movement control unit consisting of preliminary storage unit for movement code data corresponding to time code, before movie show; receiver unit; actuator movement control unit aligned with the stored movement code when a movie starts.

EFFECT: smoothing of seat actuator drive movement with time synchronisation to the movie shown.

4 cl, 3 dwg

FIELD: physics, photography.

SUBSTANCE: invention relates to frame grabbers. The result is reached by that the frame grabber includes the generation unit designed with a possibility of generation of the image data, and the resolution unit designed with a possibility, on the basis of the first image data generated by the generation unit, when the position in focus is located in the first focal position, when the object is located in the focused state, or the second focal position on the side of smaller distance of the first focal position, and the second image data generated by the generation unit, when the position in focus is located in the third focal position on the side of greater distance of the focal position, when the back ground is located in the focused state, resolutions of the first area including the object, and the second area including the background.

EFFECT: accurate resolution of the object of shooting and background, even if the image data have poor difference by depth between the object and background.

16 cl, 11 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to image capturing devices. The result is achieved due to that the image capturing device comprises an image capturing unit configured to capture an image of an object through an optical system; a display unit configured to display an image captured by the image capturing unit on a screen; a determination unit configured to simultaneously determine a plurality of touch positions on the screen where an image is displayed; and a control unit configured to smoothly adjust the focusing state in accordance with change in distance between a first determined touch position and a second determined touch position in order to change the focusing area.

EFFECT: broader technical capabilities of the image capturing device.

13 cl, 27 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to computer engineering. An image processing device for detecting, from image data generated by an image sensor formed by a plurality of pixels, a signal from a defective pixel of the image sensor comprises a first determination unit for obtaining a first determination value indicating the value of the difference in signal strength between a pixel of interest and a plurality of pixels located near the pixel of interest; a second determination unit for obtaining a second determination value indicating the distribution width of the difference in signal strength between the pixel of interest and the plurality of pixels located near the pixel of interest; and a detection unit for detecting if the signal from the pixel of interest is a signal from a detective pixel using the first determination value and the second determination value, wherein the first determination unit obtains the first determination value by obtaining the difference in signal strength between the pixel of interest and each of the plurality of pixels located near the pixel of interest, obtaining from each difference value indicating the probability that the signal from the pixel of interest is a signal from a defective pixel, and multiplying the obtained values.

EFFECT: high accuracy of detecting a defective pixel.

11 cl, 22 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to an image forming apparatus. The result is achieved due to that the image forming apparatus includes a control unit and a detector which includes a plurality of pixels and which performs an image capturing operation for outputting image data corresponding to emitted radiation or light. The image capturing operation includes a first image capturing operation in which the detector is scanned in a first scanning region which corresponds to part of the plurality of pixels to output image data in the first scanning region, and a second image capturing operation in which the detector is canned in a second scanning region larger than the first scanning region to output image data in the second scanning region. The control unit prompts the detector to perform an initialisation operation for initialising a conversion element during a period between the first image capturing operation and the second image capturing operation in accordance with the switch from the first scanning region to the second scanning region.

EFFECT: design of a device capable of reducing the difference in level which might arise in a captured image and which depends on the scanning region to prevent considerable deterioration of image quality.

9 cl, 8 dwg

FIELD: physics, computer engineering.

SUBSTANCE: group of inventions relates to image processing technologies. An image processing device for reconstruction processing for correcting image quality deterioration due to aberration in an optical image-forming system. The image processing device comprises a dividing means for dividing image data of colours of colour filters into image data of corresponding colours of colour filters. The device also includes a plurality of image processing means, each designed to perform reconstruction processing by processing using an image data filter of one of the corresponding colours divided by said dividing means.

EFFECT: fewer false colours through image reconstruction processing in a RAW image, as well as reduced load on image reconstruction processing.

10 cl, 33 dwg

FIELD: physics.

SUBSTANCE: apparatus for adjusting a magnetooptical system for forming a beam of protons consists of a pulsed electromagnet which is formed by a pair or a system of pairs of thin conductors directed along the axis of a proton graphic channel spread in a transverse plane. A scaling array of metal plates mounted in a frame is placed at the output of the electromagnet. The method of adjusting a magnetic system for forming a beam of protons and a method of matching magnetic induction of an imaging system involve generating a magnetic field, through which the beam of protons is passed, the direction of said beam through the imaging system to a recording system by which the image of the scaling array is formed. Upon obtaining a distorted image, the magnetic beam forming system is adjusted and magnetic induction of the magnetooptical imaging system is adjusted by varying current of lenses of said systems and retransmitting the beam of protons until the required images are formed.

EFFECT: high quality of adjustment.

4 cl, 14 dwg

FIELD: physics, photography.

SUBSTANCE: invention relates to an image processing device and method, which can improve encoding efficiency, thereby preventing increase in load. The technical result is achieved due to that a selection scheme 71 from a prediction scheme 64 by filtering selects a motion compensation image for generating a prediction image at a high-resolution extension level from key frames at a low-resolution base level. The filter scheme 72 of the prediction scheme 64 by filtering performs filtration, which includes high-frequency conversion and which uses analysis in the time direction of a plurality of motion compensation images at the base level, selected by the selection scheme 71, in order to generate a prediction image at the extension level.

EFFECT: reducing load in terms of the amount of processing owing to spatial increase in sampling frequency at the base level for encoding the current frame.

19 cl, 26 dwg

FIELD: information technology.

SUBSTANCE: method of compression of graphic file by fractal method using ring classification of segments, in which the graphic file is split into rank regions and domains, and for each rank region the domain and the corresponding affine transformation is found, that best approximates it to the appropriate rank region, and using the obtained values of the domain parameters, comprising their coordinates, the coefficients of the affine transformations, the values of brightness and contrast, the archive is formed, and classification of domains and rank regions are introduced, based on the allocation in them of the "rings" and the calculation of the mathematical expectation of pixel intensity of these "rings", which enables to reduce the complexity of the phase of correlation of the segments and to accelerate compression.

EFFECT: reduced time of compression of the graphic file by fractal method.

3 dwg

FIELD: physics.

SUBSTANCE: method comprises making each array element in an image sensor from one "R, G, B radiation colour brightness to code" converter, which performs parallel synchronous conversion of radiation of three colours analogue video signals R, G, B into three codes. The frame image digitisation apparatus includes an objective lens, an image sensor comprising an array of elements, three switch units, three register units and a control signal generator, wherein each switch unit includes the same number of encoders as converters.

EFFECT: reduced cross dimensions of array elements in an image sensor, which enables to reduce the frame format size or increase resolution of the image sensor.

6 dwg, 1 tbl

FIELD: physics.

SUBSTANCE: disclosed is a frame image digitisation apparatus. The disclosed apparatus comprises a lens in the focal plane of which there is an image sensor having an array of elements, a control signal generator and three register units, the outputs of which are the outputs of the digitisation apparatus. Each array element consists of a converter for converting radiation of colours R, G, B into three codes. Images are input into the sensor, the number of said images being equal to the number of array elements and the number of colours R, G, B of analogue-to-digital converters (ADC).

EFFECT: high image frame resolution owing to conversion of three colours R, G, B into codes using one converter.

4 dwg, 2 tbl

FIELD: physics.

SUBSTANCE: apparatus comprises a lens, an image detector having an array situated in the focal plane of the lens, the array having elements which are converters for converting radiation to codes based on the frame resolution number 106, each having an opaque housing in the front part of which, in a partition wall, there is a microlens, on the optical axis of which and at an angle of 45 thereto semitransparent micromirrors are arranged in series and rigidly mounted based on the number of bits per code, each preceding micromirror transmitting to the next micromirror radiation flux with half the strength.

EFFECT: high speed of frame digitisation.

1 tbl, 4 dwg

FIELD: physics.

SUBSTANCE: apparatus comprises a lens, an image detector which includes an array of elements based on the frame resolution number 106, situated in the focal plane of the lens and having three groups of outputs of colour codes R, G, B, includes three register units and a control signal generator which outputs from the first output pulses with frame frequency (25 Hz), connected to the control inputs in array elements, and from the second output pulses with code sampling frequency, connected in parallel to the second control inputs of the first through third register units.

EFFECT: high frame resolution by making array element converters of brightness of radiation of colours R, G, B - three codes, which synchronously output codes of three colours R, G, B.

5 dwg, 1 tbl

FIELD: physics.

SUBSTANCE: disclosed is a method of obtaining a structural image of a biological object in optical coherence tomography. The method includes breaking down a source colour vide frame into non-overlapping spatial blocks consisting of more than one pixel. A structural image is obtained via small-angle raster scanning in the arm of an optical coherence tomography sample. The obtained image with a size of Piskh bytes is broken down into non-overlapping spatial blocks only on columns; adjacent column blocks are averaged pixel by pixel to form a new image with a size of Pstl bytes; the new image is broken down into non-overlapping spatial blocks only on rows; adjacent row blocks are averaged pixel by pixel to form a resultant image with a size of Pres bytes, and the averaging process is controlled based on an exponential relationship Pstl from the number of averaging column blocks Ustl and Pres from the number of averaging row blocks - Ustr.

EFFECT: high quality of the structural image of a biological object in optical coherence tomography.

7 dwg

FIELD: physics, video.

SUBSTANCE: invention relates to video encoding and decoding, which provides conversion between a spatial region and a conversion region. The method of encoding video includes breaking down a video image into encoding units having a maximum size. The images are encoded based encoding units according to depths obtained by hierarchically breaking down each maximum depth encoding unit, and based on the type of partition defined according to depths of the depth encoding units. The type of partition includes a data unit having the same size as the current encoding unit, and a partial data unit obtained by breaking down the height or width of the current encoding unit. Encoding units are determined according to encoding depths relative to each of the depth encoding units, and encoding units having a tree structure are therefore determined. The encoded data are output.

EFFECT: high efficiency of image compression and, as a result, high efficiency of encoding and decoding video.

15 cl, 15 dwg, 1 tbl

FIELD: physics, video.

SUBSTANCE: invention relates to encoding three-dimensional video signals, and specifically to a transport format used to transport three-dimensional content. The technical result is achieved using a device which is characterised by that it includes a means of generating a stream which is structured into multiple levels: level 0, having two independent layers: a base layer containing video data of a right-side image, and a level 0 extension layer containing video data of a left-side image, or vice versa; level 1, having two independent extension layers: a level 1 first extension layer containing a depth map relating to the image of the base layer, a level 1 second extension layer containing a depth map relating to the image of the level 0 extension layer; level 2, having a level 2 extension layer containing overlapping data relating to the image of the base layer.

EFFECT: high quality of three-dimensional images with a large number of presentations used.

6 cl, 2 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to means of encoding and decoding images. In the method, the motion vector of a reference section has the same shape as the current section and belongs to a reference image which is different from the current image and is broken down in advance as a result of encoding with subsequent decoding into a plurality of sections. When a reference section overlaps a set of reference sections from said plurality of sections of the reference image, said motion vector of the current image section is determined based on a reference motion vector function belonging to a set of reference motion vectors associated with k overlapped reference sections.

EFFECT: high accuracy of predicting the motion vector of an image section.

15 cl, 6 dwg

FIELD: systems for encoding and decoding video signals.

SUBSTANCE: method and system for statistical encoding are claimed, where parameters which represent the encoded signal are transformed to indexes of code words, so that decoder may restore the encoded signal from aforementioned indexes of code words. When the parameter space is limited in such a way that encoding becomes inefficient and code words are not positioned in ordered or continuous fashion in accordance with parameters, sorting is used to sort parameters into various groups with the goal of transformation of parameters from various groups into indexes of code words in different manner, so that assignment of code word indexes which correspond to parameters is performed in continuous and ordered fashion. Sorting may be based on absolute values of parameters relatively to selected value. In process of decoding, indexes of code words are also sorted into various groups on basis of code word index values relatively to selected value.

EFFECT: increased efficiency of compression, when encoding parameters are within limited range to ensure ordered transformation of code word indexes.

6 cl, 3 dwg

Up!