KR0162200B1 - Data compression device with overhead data insertion - Google Patents

Abstract

According to an image data compression apparatus having an overhead data insertion function of the present invention, a circuit for generating a quantization level variable according to a buffer state and a circuit for inserting an overhead data into the image data are separately configured to overwrite the image data. There is an effect that can adaptively generate the quantization level regardless of the data delay for inserting the head data. In addition, it is possible to perform parallel processing of signals by determining the output amount of data using the buffer state.

Description

Image data compressor with overhead data insertion

1 is a conventional image data compression system.

2 is a screen and data division unit diagram.

3 is a block diagram showing a parallel processor of an image data compression apparatus implementing one preferred embodiment of the present invention.

4 is a down counter connection diagram of parallel processors.

5 is a combined arrangement diagram of image data and overhead data.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data compression apparatus having an overhead data insertion function, and more particularly, to an image data compression apparatus having an overhead data insertion function for image data before channel transmission.

A fully digital HD-TV encoder, a video data compression system of a digital VCR, an encoder of a video conferencing system, and the like transmit channel overhead data to the video data so that the decoder can decode the received video signal. Overhead data is additional data related to the video signal for decoding at the decoder before channel transmission, and includes an end block (EOB), a frame start code (FSC), and a quantization level (QL). Level (MV), Motion Vector (MV), Frame / Field Code (FFC), Intra / Inter Code (IIC), and the like. This overhead data is muxed with compressed video data to enable decoding and can be changed according to system specifications.

1 is a conventional image data compression system, which is a system for compressing image data about a 1/4 screen formed by dividing an entire screen into four vertically.

The apparatus of FIG. 1 comprises means for discrete cosine transforming and quantizing a first multiplexer 1 that receives and multiplexes a Y signal (luminance signal), a U signal (RY difference signal), and a V signal (BY difference signal). And a means for inverse quantization and inverse discrete cosine conversion of input data, a means for motion estimation and compensation for storing image data in units of frames, a variable length encoding means, and a quantization level according to the amount of data stored. A buffer for generating QL), a means for multiplexing the output data of the buffer and data audio data and control data related to the remaining divided screens in a predetermined order, and means for correcting and correcting data errors.

When the Y, U, V signal of the sampled image is input to the first multiplexer 1, the first multiplexer 1 selectively outputs the input data in a predetermined order. That is, all the Y signals of the 8x8 block are transmitted and then transmitted in the order of the U signal V signal. The data output through the first adder A1 is converted into data in the frequency domain by the discrete cosine transform unit 2 and quantized by the quantization unit 3 to predetermined representative values. The inverse quantization unit 4 and the inverse discrete cosine transform unit 5 perform lossless decoding and output the result.

The motion estimation unit 6 calculates a difference value between the image data input from the first multiplexer 1 and the image data output through the second adder A2 and the frame memory 7. The motion vector (MV) generated as a result is transmitted to the second multiplexer 8 and used to compensate for image data difference between frames of the motion compensator 9, that is, motion compensation. The output data of the motion compensator 9 is used to generate difference data between frames in the first adder A1, and is added to the image data decoded in the second adder A2 to form complete 8x8 block data.

The image data input to the quantization unit 3 is quantized to predetermined representative values determined by the quantization level QL transmitted from the buffer 10 and transmitted to the variable length encoding unit 11. When the length / level coded image data is input to the buffer 10 by the variable length encoder 11, the buffer 10 having the quantization level (QL) generator is used to store the amount of image data stored. Accordingly, the quantization level QL is generated again and transmitted to the quantization unit 3. The image data output from the buffer 10 is combined by the second multiplexer 8 with the image data, the audio data motion vector (MV), the control data, etc. of the remaining divided screens, and a forward error correction unit 12. Then, the bit error is corrected, and then digitally modulated by a Quadrature Amplitude Modulation unit 13 using two carriers of the same frequency and having a phase difference of 90 degrees. The modulated video data is transmitted to the receiver.

In the data compression system as described above, channel transmission of overhead data that facilitates data restoration at the receiving end by compressing and transmitting video data is essential, but if the video data is delayed by one frame interval to insert the overhead data, the data is divided into quarters. There is a problem that it is difficult to determine the quantization level having a much smaller data interval than the image data of the screen.

An object of the present invention for solving the above problems, by separately configuring the quantization level generating means and the image data delay means, to adaptively generate the quantization level according to the amount of data input, delay the image data to overhead data The present invention provides an overhead data insertion apparatus capable of inserting the.

As described above, an object of the present invention is to provide a video data compression system that performs intra frame coding and inter frame coding in which data quantization rate is variable according to a quantization level, and receives predetermined image data and overhead data. Data amount detecting means for detecting the data length of the total data including the head data, quantization level generating means for receiving the total data length information detected by the data amount detecting means and generating a quantization level corresponding to the total data length; An overhead data insertion means for receiving the predetermined image data and the overhead data and inserting the overhead data into the image data, the total data length information of the data amount detection means, and outputting from the overhead data insertion means Receives and saves the video data inserted the data, the The stored data is achieved by the overhead insertion functions of the video data compression apparatus comprising a comprising a data output means for outputting the image data for the data length of the body data is coefficient.

In addition, an object of the present invention is to divide an entire screen into a plurality of split screens, and use intra-frame encoding and inter-frame coding using a plurality of parallel processors whose data quantization rate varies according to the quantization level. In a video data compression system that encodes and outputs encoded video data related to each divided screen in a predetermined order, the predetermined data length and overhead data of each divided screen are input, and the length of the total data obtained by adding up the video data and the overhead data is calculated. Data quantity detecting means for detecting each divided screen, quantization level generating means for receiving and adding total data length information detected for each divided screen, and generating a quantization level corresponding to the added data length, and each division Video data for each screen divided by receiving predetermined video data and overhead data for each screen Overhead data inserting means for inserting overhead data into another, and image data outputted from the overhead data inserting means and inserted with overhead data are stored for each divided screen, and stored in a predetermined split screen of the data amount detecting means. Image data corresponding to the predetermined divided screen is output from the image data stored for each divided screen for a time corresponding to the accumulated data length for the entire data length, and a time corresponding to the accumulated data length of the predetermined divided screen elapses. The image corresponds to all the divided screens by outputting the image data corresponding to the other predetermined divided screens for a time corresponding to the data length accumulated in the total data length for the other predetermined divided screens and includes the overhead data. Overhead data insertion including data output means for outputting data It is also achieved by an image data compression device having an input function.

Hereinafter, an image data compression device having an overhead data insertion function according to an exemplary embodiment of the present invention will be described in detail with reference to the accompanying drawings.

Video signals of high-definition TVs have wide bandwidths, so high-speed system clocks are required for signal processing. In the existing video data compression system, the entire screen is divided into four equal parts and processed using four way parallel processors to process the video signal of high quality TV.

FIG. 2 is a screen and data division unit diagram. FIG. 2 (a) shows four divided screens in which the entire screen is vertically divided. FIG. 2 (b) shows eight blocks of 8 × 8 pixel units. Macro block (MB) of Y, U, and V signals is shown, and FIG. 2 (c) shows a slice composed of 11 macroblocks and a group of macroblocks composed of 4 slices. ; MMB). The length of the slice or macroblock group is the same as the horizontal length of the split screen in FIG. Therefore, a plurality of macroblock groups are arranged vertically in each split screen.

FIG. 3 is a block diagram illustrating a parallel processor of an image data compression apparatus implementing one preferred embodiment of the present invention. In FIG. 1, a motion compensation block transform block, a quantization block, and a second multiplexer ( 8) The subsequent blocks are not shown and only the blocks of the circuit replaced by the variable length coding unit 11 and the buffer 10 are shown. A parallel processor for inserting overhead data into video data of one split screen using the screen and data division units of FIG. 2 is shown.

The apparatus of FIG. 3 includes a data length counter 31, which is a data amount detection means for detecting the total data length by receiving image data and overhead data, and a quantization level in slice units by receiving the full data length from four parallel processors. Variable length encoding of quantization level generating means for generating a (QL), means for inserting overhead data in units of macroblock groups (MMB) connected to the above-described data amount detecting means, and image data into which overhead data is inserted. And a variable length encoding unit, and data output means for converting and outputting data of each variable length encoded split screen into serial data.

The data length counter 31 includes input terminals of the quantized data QD, the motion vector MV, and the intra / interframe decision code (IID), predetermined overhead data, that is, the frame start code (FS) and the slice start. The terminal which receives the sign (SS), the macroblock group start code (MMS), the macroblock start code (MS), the block start code (BS), and the predetermined clock (CLK1) is included. The data length is counted by using the clock CLK1 for each slice, and the quantization level generating means includes data length counters of three parallel processors and data output from the data length counter 31 described above. An adder 33 for summing the lengths and a first ROM 34 for outputting the quantization level QL corresponding to the output data of the adder 33 are provided.

The overhead data insertion means includes a first memory 35 for storing quantized data QD input to the data length counter 31, the above-described motion vector MV, and intra / frame determination code IID. And a latch unit 36 having input terminals of the quantization level QL and input terminals of the field / frame determination code FFD and the clock signal CLK2, quantization data QD stored in the first memory 35, Frame Start (FS) code and Slice Start (Slice Start) generated by the overhead data generator (not shown) by receiving the output data (MV, IID, QL, FFD) of the latch unit 36. A multiplexer 37 combining a SS) code, a Mass of Macro block Start (MMS), a Macro block Start (MS) code, and a Block Start (BS) code; do.

The data output means is connected in parallel to the output terminal of the multiplexer 37 and is variable by combining the output data of the second ROM 38 and the third ROM 39 and the ROMs 38 and 39 which store the codeword and the codeword length. The variable length encoder 41 having the long data connector 40, the second memory 42 for storing the variable length coded data, and the data length output from the data length counter 31 described above. Down-counter 32 for counting and parallel to serial converter (hereinafter referred to as P / S converter) for converting output data of second memory 42 during counting time of down-counter 32 (43) ).

The circuits of the apparatus of FIG. 3 except for the adder 33 and the first ROM 34 are configured for each parallel processor, but only one parallel processor is shown to simplify the overall block diagram.

4 is a diagram illustrating a down counter connection of the parallel processors. The data length counters 44, 45, 46, 47, down counters 48, 49, 50, and 51 and P / S of each of the four parallel processors are shown. The connection relationship between the converters 52, 53, 54, 55 and the adder 33 and the first ROM 34 generating the quantization level by adding the output data of each data length counter is shown. Each down counter is connected to operate by a circulating preset signal PE. That is, when the counting of the first down counter 48 is completed, the preset signal PE1 output from the first down counter 48 controls to start the counting operation of the second down counter 49 and the first down counter 48. 49 are connected to each other in such a manner as to be operated by the output signal PE4 of the fourth down counter 51.

FIG. 5 is a combined arrangement diagram of image data and overhead data, and is related to the data division of FIG. 2 forming one macroblock group (MMB) with four slices separated by dotted lines. FIG. 5 (a) shows four segmented images divided into three vertical lines, a slice (Slice) which is each unit box, and a macroblock group (MMB) by four vertical boxes. FIG. Represents an entire data array of one slice. FIG. 5C shows overhead data in the A slice of FIG. 5A having frame information. FIG. 5 (d) shows overhead data at the beginning of the macroblock group. That is, the overhead data in the B slice or the C slice of the first divided image that is the beginning of the macroblock group (MMB) of the second divided image that does not require the frame start information. FIG. 5E shows the overhead data of the macroblock MB. That is, the overhead data of the macroblocks that do not require the frame start information and the start information of the macroblock group are shown. FIG. 5 (f) shows an overhead data array for each macroblock MB area, and FIG. 5 (g) shows an arrangement of luminance signals and color signals constituting the macroblock.

That is, FIG. 5 (g) shows the arrangement of eight luminance signals Y1, ..., Y8 and color signals U, V. FIG.

In the frame preset in relation to the quantized discrete cosine coefficient data QD output from the quantization unit 3 of FIG. 1, the motion vector MV of the motion estimation unit 6, and the data to be encoded. When the inter-frame decision code (IID) is input to the data length counter 31, the data length counter 31 with a run length coding unit and a counter is input quantized data (QD), that is, discrete code. After the length encoding (RLC) of the sine conversion factor, the total data length obtained by inputting the overhead data (FS, SS, MMS, MS, BS) output from the overhead data generator not shown in the figure is input. By using CLK1). That is, the overhead data combined with the slice size quantization data of FIG. 2 is counted using the input output. The data length counter 31 outputs the total data length information generated by the coefficients to the down counter 32 and the adder 33, and the above-described line length coded (RLC) quantized data (QD) of the input data. And output image data to the first memory 34. The data length information counted by the data length counter unit 31 is input to the adder 33 together with the data length information in units of slices output from three other parallel processors (not shown). The adder 33 accumulates data length information transmitted from four parallel processors. The accumulated data length is composed of a programmable ROM and is used as the address of the first ROM 34 that stores the quantization level QL corresponding to the accumulated data length. Accordingly, the first ROM 34 outputs a quantization level QL corresponding to the accumulated data length. The quantization level QL generated in the first ROM 34 is used to determine the quantization representative value of the quantization units included in the four parallel processors.

The above-mentioned line length coded (RLC) image data is stored in a first-in first-out (FIFO) type first memory 35 that stores one frame of data.

The latch unit 36 includes a motion vector (MV), an intra-frame / inter-frame determination code (IID) input to the data length counter unit 31, the quantization level QL of the first ROM 34, and a frame. The above-described input signals are outputted according to the latch signal CLK2 generated at the slice interval by receiving the / field decision code FFD.

The multiplexer 37 is provided with overhead data (FS, SS, MMS, MS, BS) applied from an overhead data generator (not shown) connected to the data length counter unit 31, and output data (MV, IID, QL, and FFD) are received and combined with the image data output from the first memory 35. That is, various overhead data are combined with the image data in the form of FIG. 5 by using the vertical blanking time and the horizontal blanking time that do not include the image data. Referring to FIG. 5, the data array of a slice, which is one of the overhead data insertion units of the multiplexer 37, is included in the frame and macroblock group overhead data, the overhead data of the first macroblock, and the Y in the first macroblock. It is arranged in the order of 10 blocks, etc. by the U and V signals, and the unused area is included in the position where the data of the 11th macroblock ends. Therefore, when the apparatus of FIG. 3 processes the first slice related data of the first divided image of FIG. 5 (a), the frame / macroblock group overhead of FIG. 5 (b) is reduced to FIG. 5 (c). Has the indicated data configuration. In the data structure of FIG. 5 (c), the frame number and macroblock group number are used to synchronize and restore undamaged data when the transmitted data is damaged. The second slice of the first divided image has the data structure of FIG. In this manner, the data structure of Fig. 5 (d) is formed in the first slice of the second divided image. In FIG. 5 (b), the areas indicated by macroblocks and numbers are areas into which overhead data of macroblocks are inserted. Intra / interframe code, horizontal and vertical components of motion vectors, field / frame codes, and unused areas It has a data structure arranged in the order of. In addition, the regions indicated by the combination of numbers and blocks are composed of eight luminance signal blocks, one U signal block and one V signal block, and an end of block (EOB) indicating the end of each block.

When the multiplexer 37 inserts and outputs the overhead data into the image data, the second ROM 38 and the third ROM 39 storing the codeword and the code length are codewords and the code length corresponding to the input data. It outputs a predetermined data indicating the, the data connector 40 is connected to the input data and outputs. Data output from the variable length coding unit 41 is stored in the second memory 42.

The down counter 32 receives and counts length information about the total data including the image data and the overhead data output from the data length counter 31. The down counter 32 accumulates data length information in units of macroblock groups MMB, and is down counted by the input clock CLK3. The down counter 32 outputs the enable signal ENL to the P / S converter 43 while the counting continues.

Therefore, the image data VD stored in the first-in-first-out (FIFO) type second memory 42 is output in the form of a bitstream by performing parallel-to-serial conversion while the P / S converter 43 is enabled.

When the first down counter 48 of FIG. 4 completes counting of the macroblock group (MMB), the first down counter 48 applies the precell signal PE1 to the second down counter 49. During the counting period of the second down counter 49, the image data of the second divided image is converted and output in the form of a bit stream by the second P / S converter 53. In this manner, when the image data (VD) of the first macroblock groups of each divided image is output from the P / S converters 52, 53, 54, and 55, the data of the second macroblock groups of each divided image is obtained. Signal processing is performed in the manner described above. The image data output from the parallel processors are combined with audio data and control data by the second multiplexer 8 of FIG. 1 and output through the forward error correcting unit 12 and the quadrature amplitude modulator 13.

According to the image data compression device having the overhead data insertion function of the present invention as described above, the image data is constructed by separately configuring a circuit for generating a quantization level that varies with the state of the buffer and a circuit for inserting the image data overhead data. There is an effect that can adaptively generate the quantization level regardless of the data delay for inserting the overhead data into the. In addition, it is possible to perform parallel processing of signals by determining the output amount of data using the buffer state.

Claims (11)

In a video data compression system that performs intra-frame encoding and inter-frame encoding in which the data quantization rate varies according to the quantization level, the data of the entire data obtained by inputting predetermined video data and overhead data and adding the video data and the overhead data together. Data amount detecting means for detecting a length; Quantization level generating means for receiving full data length information detected by the data amount detecting means and generating a quantization level corresponding to the entire data length; Overhead data insertion means for receiving the predetermined image data and the overhead data and inserting the overhead data into the image data; Outputs the total data length information of the data amount detecting means and the image data inserted by the overhead data inserting means and receives and stores the image data inserted with the overhead data, and outputs the stored image data while the data length of the total data is counted. An image data compression device having an overhead data insertion function comprising a data output means. The method of claim 1, wherein the overhead data inserting means comprises a first memory for receiving and storing the predetermined image data and a multiplexer for inserting the overhead data into the image data output from the first memory. Image data compression device with overhead data insertion. The data output unit of claim 1 or 2, wherein the data output unit is connected in parallel to an output terminal of the multiplexer and stores data for storing codewords and codeword lengths, and a data connector for variable length encoding by combining output data of the storage units. A variable counter encoding unit, a down counter unit which receives the total data length information of the data amount detecting unit and accumulates the data in units of a predetermined data size, and outputs a predetermined control signal while counting the accumulated data length; A second memory for storing the video data output from the overhead data insertion means, and a parallel / serial converter for outputting the video data stored in the second memory while converting the serial data to the serial data while the count down is continued in the down counter unit. Image data compression device having an overhead data insertion function, characterized in that provided. The entire screen is divided into a plurality of split screens, and image data related to each split screen is encoded in a frame and inter-frame using a plurality of parallel processors whose data quantization rate is variable according to the quantization level. In a video data compression system that outputs related data in a predetermined order, the amount of data that receives the predetermined video data and the overhead data of each divided screen and detects the length of the total data including the combined video data and the overhead data for each divided screen. Detection means; Quantization level generating means for receiving and adding total data length information detected for each of the divided screens and generating a quantization level corresponding to the added data length; Overhead data insertion means for receiving predetermined image data and overhead data for each divided screen and inserting overhead data into the image data for each divided screen; The image data output from the overhead data inserting means and stored with the overhead data is stored for each divided screen, and divided for a time corresponding to a data length that accumulates the total data length for a predetermined divided screen of the data amount detecting means. Data length corresponding to the predetermined divided screen is output from the image data stored for each screen, and when the time corresponding to the accumulated data length of the predetermined divided screen elapses, the data length accumulates the total data length for the other predetermined divided screen. And a data output means for outputting image data corresponding to all the divided screens in a manner of outputting image data corresponding to the other predetermined divided screen for a time corresponding to Image data compression device provided. 5. The apparatus according to claim 4, wherein the quantization level generating means includes an adder for adding the total data length information detected for each of the divided screens, and a ROM for outputting a predetermined quantization level corresponding to the output data of the adder. An image data compression device with an overhead data insertion function. The image data compression apparatus as claimed in claim 6, wherein the quantization level generating means generates the quantization level in slice units. 6. The data processing apparatus according to claim 5, wherein the overhead data inserting means comprises: a first memory portion comprising a plurality of memories for storing image data of each divided screen, and a plurality of inserting overhead data into data output from the plurality of memories; An image data compression device with an overhead data insertion function, characterized in that it comprises a multiplexer. 8. The data output means according to claim 7, wherein the data output means is cyclically connected to start the counting operation of the data length accumulated by another predetermined counter when the count of the data length accumulated in the predetermined counter is completed and corresponds to the number of each divided screen. A second memory unit comprising a plurality of counters, a plurality of memories for storing image data of the overhead data inserting means into which the overhead data is inserted, for each divided screen, and one of the cyclically connected counters is a second memory Parallel / serial conversion consisting of parallel / serial converters that are connected to all counters and one by one in a manner that is connected to a single negative memory, and the connected counters receive data from the memory and convert them into serial data while performing the counting operation. Image data compression with an overhead data insertion function, characterized in that the unit is provided Value. The image data compression apparatus as claimed in claim 8, wherein the plurality of counters accumulate data of a macroblock group size. 5. The video data compression apparatus as claimed in claim 4, further comprising a variable length encoding unit for variable length encoding the output data of the overhead data insertion unit and transmitting the variable data to the data output unit. The video data compression apparatus as claimed in claim 7 or 10, wherein the first memory unit and the second memory unit are first-in-first-out memory.