JPH0435136A - Picture packet receiver - Google Patents

Abstract

PURPOSE:To reduce the deterioration in the picture quality by generating a pseudo packet when an aborted packet is known and supplying the packet to a buffer memory section to be supplemented to be the packet abort location and implementing inter-frame and in-frame interpolation processing. CONSTITUTION:An information allocation processing section 7 based on an output from a packet abort detection section 5 generates a pseudo packet in the presence of an aborted packet and supplements the packet to a packet abort location as a packet information signal. The packet information signal is stored in a buffer memory section 2, a buffer operating detection section 8 detects the buffer operating rate of the memory section 2 and it is outputted to the processing section 7. When a high load state is discriminated based on the buffer operating rate, the processing section 7 gives a command to a buffer read control section 9, reads the memory section 2 quickly to decrease the buffer operating rate thereby preventing the overload state in advance. Moreover, simultaneously the interpolation processing is implemented by a frame memory section 3. Thus, the deterioration in the picture quality is reduced.

Description

Detailed Description of the Invention (Industrial Field of Application) The present invention relates to an image packet transmitter or an image packet multiplexer on the transmitting side in an image packet communication device system connected to a high-speed packet switching network. The present invention relates to an image packet receiving device that receives an image packet signal generated by a computer and transmitted through a high-speed packet switching network, and reproduces image information.

(Conventional technology) For technical trends and technical issues related to image packet technology, see Hatahiko Yasuda, ``Video Packet Transmission Technology J'', 1987 Image Coding Symposium (IEICE PC3J87).
(Sep, 1987).

Also, Hatahiko Yasuda "Video Packet Transmission" Journal of the Television Society Vol. 1.42. No. 6 (1988) p.
533 (17) ~p, 536 (20),
It describes the characteristics of packetized video transmission, the burst nature and statistical multiplexing of video signals, and issues with video packet transmission.

The high-speed packet switching described in the above document simplifies the communication protocol within the network and increases the packet transmission rate to 1.5 Mbit/s or more, which is the primary rate of the digital network, thereby converting real-time packets into It enables communication. In a high-speed packet switching network, it is possible to handle various types of information such as audio, nine images, and data, which have different transmission speeds, all in the same packet, and it is possible to flexibly handle changes in each information over time. In addition, efficient operation of the transmission path can be expected due to the statistical multiplexing effect that takes advantage of the burst nature of the information source.

Here, nine audio images are used as high-speed packets.

It is assumed that all data transmission information is handled in short fixed-length packets (for example, 32 bytes = 256 bits).

An increase or decrease in the amount of generated transmission information that differs from each information source corresponds to an increase or decrease in the number of high-speed packets generated.

Among various types of information, image information has a larger amount of information than audio information, and the burst nature of the information source is large.

The object of image communication that requires real-time performance is moving images. A television signal as moving image information is a still image signal (hereinafter referred to as a frame signal) that is continuous in the time axis direction.

). Since the amount of information in a moving image signal has particularly large temporal fluctuations due to the nature of the information source and the movement of the screen, high-efficiency encoding is generally applied to compress the band and redundancy of the image information.

Therefore, in the image packet communication device system of the high-speed packet switching network targeted here, the image packet transmitting device or image packet multiplexing device on the transmitting side applies high-efficiency encoding to the image signal and converts the image signal. After creating an encoded signal, an image packet is generated by packet assembly and sent to a transmission path. The image packets sent out to the transmission path are transmitted within the high-speed packet network and arrive at the image packet receiving device on the receiving side. The image packet receiving device decomposes the received image packet into packets and then decodes the encoded signal to reproduce the image signal.

(Problems to be Solved by the Invention) In the transmission path (link) between exchanges (nodes) of the high-speed packet switching network in the image packet communication device system of the high-speed packet switching network, it is necessary to Since link-pi-link flow control and error control are not performed, it is affected by communication traffic within the network, resulting in packet overflow due to packet multiplexing in the notebook, and packet loss due to packet loss. It is the seller's responsibility to dispose of it.

In an image packet communication device system connected to a high-speed packet switching network, when an image packet output from the image packet transmitting device or image packet multiplexing device on the sending side is discarded within the high-speed packet switching network. In the image packet receiving device on the receiving side, which is the arrival point of the image packet, there is a problem that image information is lost, resulting in significant deterioration of image quality and making image communication difficult.

SUMMARY OF THE INVENTION In view of the above-mentioned problems, an object of the present invention is to reduce the deterioration of the image quality of the image signal to be reproduced when packets are discarded in a high-speed packet switching network, and to prevent image communication from becoming difficult. An object of the present invention is to provide an image packet receiving device that prevents the above.

(Means for Solving the Problems) The present invention provides a packet disassembly section that receives an image packet input signal, separates it into a packet information signal and a header information signal, and outputs the separated packet information signal, and a packet information signal that is the output of the packet disassembly section. a buffer memory unit that inputs the packet image signal that is the output of this buffer memory unit, a frame memory unit that inputs the packet image signal that is the output of this buffer memory unit, and a decoding unit that inputs the image frame signal that is the output of this frame memory unit, and reproduces and outputs the image signal. a packet discard detection section that detects a discarded packet based on the header information signal of the image packet from the packet disassembly section and sends out the detection output; an encoding type detection unit that detects the encoding type from the header information signal and sends out the detection output; a buffer usage detection unit that detects the usage rate of the buffer memory of the buffer memory unit and sends out the detection output; Each output of the packet discard detection section, the encoding type detection section, and the buffer use detection section is supplied, and when a packet is discarded, a pseudo packet is generated to be assigned to a packet discard position, and this pseudo packet is assigned to the packet discard position. an information allocation processing unit that supplies image packets to the buffer memory unit based on instructions from the information allocation processing unit when the usage rate of the buffer memory is in a normal state or a high load state; Based on the instructions from the buffer readout control unit that controls the readout of the packet information signal of the frame memory unit and the information allocation processing unit, the buffer readout control unit performs readout control of the packet information signal of On the other hand, based on instructions from the frame write control section, which controls writing of the packet information signal read from the buffer memory section and also performs interpolation processing if a packet is discarded, and the information allocation processing section,
The frame memory section includes a frame readout control section that performs readout control of image frame signals from the interframe encoding frame memory and the intraframe encoding frame memory of the frame memory section.

(Operation) The packet decomposition unit converts the received image packet input signal into
It separates into a packet information signal and a header information signal, outputs the packet information signal to the buffer memory section, and outputs the header information signal to the packet discard detection section and the encoding type detection section. When the information allocation processing section learns of packet discard from the output from the packet discard detection section, it generates a pseudo packet to be allocated to the packet discard position, and supplies it to the buffer memory section to be allocated to the packet discard position.

The buffer read control unit controls readout of packet information signals of image packets from the buffer memory unit based on instructions from the information allocation processing unit when the buffer memory usage rate of the buffer memory unit is in a normal state or a high load state. Do this. Based on instructions from the information allocation processing section, the frame writing control section writes data read from the buffer memory section to the corresponding interframe encoding frame memory or intraframe encoding frame memory of the frame memory section. It controls the writing of packet image signals, and also performs interframe interpolation processing and intraframe spacing processing if packets are discarded. The frame readout control section performs readout control of image frame signals from the interframe encoding frame memory and the intraframe encoding frame memory of the frame memory section. The image frame signal read from the frame memory section is decoded by the decoding section, and the image signal is reproduced.

Therefore, when an image packet is discarded due to an overflow of the buffer memory of a node that is temporarily generated due to communication traffic in the high-speed packet switching network in the image packet communication device system of the high-speed packet switching network, the reproduced image Deterioration in the image quality of the signal can be reduced, and it is possible to prevent image communication from becoming difficult.

(Example) Next, embodiments of the present invention will be described using the drawings.

FIG. 1 is a block diagram showing an embodiment of an image packet receiving device according to the present invention, and FIG. 4 is an information allocation processing section 7 of FIG.
2 is an operation flowchart showing one embodiment of the present invention. here,
The image packet receiving device is a receiving side device of the image packet communication device system. The image packet communication device system includes an image packet transmitter or image packet multiplexer on a transmitting side that is supplied with an image input signal, generates and outputs an image packet signal, and from this image packet transmitter or image packet multiplexer. A high-speed packet switching network to which image packet signals are supplied, and an image packet receiving unit of the present invention which receives the image packet signals sent from the high-speed packet switching network via a transmission line and reproduces the image signals. It consists of a device.

In FIG. 1, a packet decomposition unit 1 receives from a transmission path an image packet signal generated by an image packet transmitting device or an image packet multiplexing device on the transmitting side and transmitted within a high-speed packet switching network, and packetizes the image packets. Separates the input signal into a header information signal and a bucket 1 to information signal, outputs the packet information signal to the buffer memory section 2, and outputs the header information signal to the packet discard detection section 5 and the encoding type detection section 6. . The packet discard detection section 5 receives a header information signal and detects packet discard. The encoding type detection unit 6 inputs the header information signal and
Coding type (interframe coding or intraframe coding)
This is to detect. Here, the header information signal of the image packet input signal received by the packet disassembly unit 1 is preliminarily assigned a packet sequence number (hereinafter referred to as PS) in the packet assembling unit of the image packet transmitter or image packet multiplexer on the transmitting side. 1 byte: 8 bits PS
In this case, the numbers from 0 to 255 are repeated sequentially. ) and the encoding type (referred to as CI; here, interframe encoding and intraframe encoding) are assigned. The header information signal also includes the frame sequence number (FS) of the image.Here, in the case of a 5-bit FS, O
The numbers .about.29 are repeated sequentially. ) shall be given. Therefore, the packet format of the image packet signal is shown in FIG.

FIG. 2 is an explanatory diagram showing the packet format of the header information signal of the image packet signal and the packet information signal in the packet decomposition unit 1 of the image packet receiving device. In the figure, the header information signal (HD) is called a logical channel number (LC).

16 bits) and bucket 1 to sequence number (PS: 8
The header information length is 4 bytes (32 bits). . Here, the logical channel number (LC16 bits) is a logical address that is predetermined from the correspondence between the channel number on the transmitting side and the channel number on the receiving side at the time of call setup when starting image communication. Packet f reprint (PI: 2 bits) is used for, for example, audio,
It identifies moving images, still images, and data, and in this embodiment, it is assumed that moving images are selected. Encoding type (C
In this embodiment, I: 1 bit) is used to distinguish between interframe encoding and intraframe encoding of the image encoding method. As for the frame sequence number (FS: 5 bits), in this embodiment, the numbers from ○ to 29 are repeated in sequence, and 1 second of a moving image is played back by 30 consecutive image frames. .

Further, the packet information signal (referred to as IF) is 32 bytes (256 bits) as fixed length packet information, and here is image packet information (pv).

FIG. 3 shows the correspondence of information blocks between an image encoded signal and an image packet signal.

The image packet signal shown in FIG. 2B is the same as that shown in FIG. 2, and is composed of a header information signal (HD) and a packet information signal (IF). The image encoded signal shown in FIG. 3(a) is an intra-frame encoded signal (IA) that is alternately generated every cycle of a fixed number of frames in the image packet transmitter or image packet multiplexer on the transmitting side. It consists of an interframe coded signal (IR). Note that the intra-frame encoded signal (IA) is obtained by encoding all pixel samples within the frame. Further, the interframe encoded signal (IR) is obtained by encoding all pixel samples between adjacent frames. and,
Since the coding efficiency of the interframe coded signal (IR) is higher than that of the intraframe coded signal (IA), from the point of view of information compression, the ratio of the interframe coded signal (IR) is increased. (See Figure 3(a)). Here, the intra-frame encoded signal (IA) is generated every cycle of a fixed number of frames, so that when a part of the image encoded information is lost due to packet discard, the intra-frame encoded signal (IA) is generated from the closest frame after the time of the loss. This is to enable normal reproduction of image information using the inner encoded signal (IA). In addition, as for the image packet signal sent from the image packet transmitting device or the image packet multiplexing device, as shown in FIG.
The interframe encoded signal (IR) is allocated to a plurality of packet information signals (IF).

Next, the packet discard detection unit 5 receives the header information signal, detects packet discard by determining the discontinuous part of the packet sequence number (PS: 8 bits) as the position of packet discard, and outputs the detection output. is sent to the information allocation processing section 7. Further, the encoding type detection unit 6 inputs the header information signal, detects whether it is intra-frame encoding or inter-frame encoding from the encoding type (CI: 1 bit), and sends the detected output to the information allocation processing unit 7. Send to. The buffer usage detection unit 8 receives the packet information signal input to the buffer memory unit 2, detects how much of the buffer memory is used (buffer usage rate), and sends the detection output to the information allocation processing unit 7. do.

Based on the output of the packet discard detection section 5, the information allocation processing section 7 assumes a normal state when no packet is discarded, enters a determination process by detecting buffer usage, and sets it as a discard state when a packet is discarded, and discards the discarded packet part. pseudo packets (if a packet is discarded,
This is a temporary packet to fill in the discarded packet portion, and a special code is inserted into this temporary packet portion to indicate that it is a portion to be interpolated later. ) and inputs it as a packet information signal to the buffer memory section 2 in order to allocate it to the packet discard position (
Steps 81 to S4 in FIG. 4). The buffer memory unit 2 stores image information and performs speed conversion by first-in-first-out (FIFO) operation, and outputs a packet image signal to the frame unit 3.

The buffer usage detection unit 8 detects the buffer usage rate of the packet information signal in the buffer memory of the buffer memory unit 2, and the buffer usage detection unit 8 uses the detected buffer usage rate to the information allocation processing unit 7. Send to. The information allocation processing unit 7 performs determination processing based on buffer usage detection based on the output from the buffer usage detection unit 8 (
Step S4 in FIG. 4), when the buffer usage rate F of the buffer memory is smaller than a predetermined first buffer threshold value B1 (F<Bl), the normal state is set, and when F≧B1, the buffer usage rate is further increased. The magnitude of F and a predetermined second buffer threshold value B2 is determined, and when F<82, a high load state is established, and when F≧B2, an overload state is established (steps S5 to S7 in FIG. 4).

In the normal state, the information allocation processing unit 7 instructs the buffer readout control unit 9 to permit buffer readout to read packet image information from the buffer memory of the buffer memory unit 2 at a predetermined first cycle clock (Pl). Steps S5 and S6 in Figure 4). In addition, in a high load state, the information allocation processing unit 7 uses a predetermined second cycle clock (P2, where P2≦
It is assumed that the time relationship is P1. ) instructs the buffer readout control unit 9 to permit buffer readout so as not to read the packet image information from the buffer memory of the buffer memory unit 2 (fourth
Steps S5, 57S8 in the figure). In other words, when the load increases in the determination processing based on the buffer usage detection in the information allocation processing section 7, the buffer readout control section 9 speeds up the readout clock of the buffer memory in the buffer memory section 2 to reduce the buffer usage rate. Operate. Furthermore, in an overload state, the information allocation processing section 7 instructs the buffer readout control section 9 not to read packet image information from the buffer memory of the buffer memory section 2, and also sends a quality setting request signal to the outside. Send (step S in Figure 4)
5. S7°S9,5IO). This quality setting request signal is
It is used to request a change in the packet transmission quality settings to the exchange in the network.

Next, the packet image signal read out from the buffer memory section 2 is supplied to the frame memory section 3 under the control of the buffer readout control section 9 . Then, the frame memory section 3 outputs an image frame signal corresponding to the image encoding method to the decoding section 4. Here, the frame write control unit 10 transfers the packet image signal from the buffer memory unit 2 to the frame memory for interframe encoding and the frame memory for intraframe encoding of the frame memory unit 3 based on instructions from the information allocation processing unit 7. It controls writing to memory. Furthermore, the frame readout control unit 11 controls reading of image frame signals from the interframe encoding frame memory and the intraframe encoding frame memory of the frame memory unit 3 based on instructions from the information allocation processing unit 7. It is something.

The information allocation processing unit 7 determines whether the coding type (CI) is intraframe coding or interframe coding based on the output from the coding type detection unit 6. Then, in the above-mentioned normal state without packet discard, and in the normal state/high load state, the information allocation processing section 7 instructs the frame writing control section 10 to select the coding type (CI) or intraframe coding. At some point, a frame write instruction is given to write a frame of the corresponding packet image signal from the buffer memory section 2 into the frame memory for intra-frame encoding of the frame memory section 3 (steps SL1 to 513 in FIG. 4),
Further, when the coding type (CI) is interframe coding, a frame write instruction is given to write a frame of the corresponding packet image signal from the buffer memory section 2 into the frame memory for interframe coding of the frame memory section 3. (Step Sll, S12°515 in FIG. 4).

In addition, in the above-mentioned discard state with packet discard, and in the normal state/high load state, the information allocation processing section 7 sends information to the frame writing control section 10 regarding the encoding types CC, I, etc.
) is intra-frame encoding, a frame write instruction is given to write a frame of the corresponding packet image signal from the buffer memory section 2 into the frame memory for intra-frame encoding of the frame memory section 3, and at the same time a frame write instruction is issued to the frame memory section 3. An instruction is given to perform interframe interpolation processing (steps Sll to S14 in FIG. 4). Here, the intra-frame interpolation process replaces missing pixel samples in the discarded part of the packet (this part can be determined by a special code in the pseudo packet) with pixel samples in the same adjacent frame. .

In addition, in the above-described discard state with packet discard, and in the normal state/high load state, the information allocation processing unit 7 informs the frame writing control unit 10 of the encoding type (CI
) or interframe coding, the buffer memory unit 2
The corresponding packet image signal from frame memory section 3
At the same time, a frame write instruction is given to write a frame to the interframe coding memory of frame memory section 3.
Instructs to perform interframe interpolation processing for
Steps Sll, SL2. S15°816). Here, the interframe interpolation process uses pixel samples in the previous frame one before the current frame to identify the location where the packet of the current frame is discarded (this location can be determined by a special code in the pseudo packet). The missing pixel sample is replaced by a pixel sample at the same location in the previous frame immediately before the current frame or an adjacent pixel sample.

The frame write control unit 10 writes the packet image signal from the buffer memory unit 2 to the frame memory for interframe encoding or the frame memory for intraframe encoding of the frame memory unit 3 based on instructions from the information allocation processing unit 7. At the same time as performing the control, in the discard state with packet discard, and in the normal state/high load state, the corresponding inter-frame interpolation processing or intra-frame interpolation processing is performed on the frame memory unit 3.

Next, the frame readout control unit 11 controls the frame memory unit 3 based on an instruction from the information allocation processing unit 7 (instruction as to whether the coding type (CI) is intraframe coding or interframe coding). The image frame signal is read from the corresponding frame memory for intra-frame encoding or frame memory for inter-frame encoding. The image frame signal read from the frame memory section 3 is input to the decoding section 4. The decoding unit 4 reproduces the image signal and sends out an image output signal.

As can be seen from the above explanation, in the image packet receiving device, the information allocation processing section 7 generates a pseudo packet when a packet is discarded based on the output from the packet discard detection section 5, and sends it to the location where the packet is discarded. (appropriate) as a packet information signal. The packet information signal is accumulated in the buffer memory unit 2, and the buffer usage detection unit 8 detects the buffer memory usage rate (buffer usage rate) in the buffer memory unit 2, and outputs this to the information allocation processing unit 7. . When the information allocation processing unit 7 determines that the load is high based on this buffer usage rate, it gives an instruction to the buffer readout control unit 9 to read data from the buffer memory unit 2 quickly, thereby reducing the buffer usage rate. Since the number of packets is lowered, an overload state can be prevented, and smooth reception processing can be performed in the image packet receiving device. Further, when a packet is discarded, the information allocation processing section 7 instructs the frame write control section 10 to perform write control to the frame memory section 3 and at the same time causes the frame memory section 3 to perform interpolation processing. Therefore, it is possible to reduce the deterioration of the image quality of the image signal reproduced by the image packet due to the temporary overflow of the buffer memory at the node due to the communication traffic in the high-speed packet switching network, which makes image communication difficult. You can prevent this from happening.

The present invention is not limited to this embodiment, and various applications and modifications can be made without departing from the gist of the present invention.

(Effects of the Invention) As described above, according to the present invention, in an image packet communication device system of a high-speed packet switching network, overflow of the buffer memory at a node temporarily caused by communication traffic within the high-speed packet switching network , it is possible to reduce the deterioration of the image quality of the reproduced image signal when the image packet is discarded, and it is possible to prevent the image communication from becoming difficult.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of an image packet receiving device according to the present invention, FIG. 2 is an explanatory diagram showing the packet format of the image packet signal in FIG. 1, and FIG. 3 is an image encoded signal and an image packet. FIG. 4 is an explanatory diagram showing the correspondence between signals and information blocks, and FIG. 4 is an operation flowchart showing an embodiment of the information allocation processing section 7 of FIG. 1. 1...Packet decomposition unit, 2...Buffer memory unit,
3...Frame memory section, 4...Decoding section, 5...
- Packet discard detection unit, 6... Encoding type detection unit, 7... Information allocation processing unit,
8... Buffer use detection section, 9... Buffer succession control section, 10... Frame write control section, 11... Frame readout control section. Patent applicant Oki Electric Industry Co., Ltd.

Claims (1)

[Scope of Claims] A packet decomposition unit that receives an image packet input signal, separates it into a packet information signal and a header information signal, and outputs the same; a buffer memory unit that receives the packet information signal output from the packet decomposition unit; , an image packet having a frame memory section that inputs the packet image signal that is the output of this buffer memory section, and a decoding section that inputs the image frame signal that is the output of this frame memory section, reproduces and outputs the image signal. The receiving device includes a packet discard detection unit that detects a discarded packet based on the header information signal of the image packet from the packet disassembly unit and sends out the detection output, and a packet discard detection unit that detects a discarded packet based on the header information signal of the image packet from the packet disassembly unit, and determines the encoding type based on the header information signal from the packet disassembly unit. a coding type detection section that detects the usage rate of the buffer memory of the buffer memory section and sends out the detection output; a buffer usage detection section that detects the usage rate of the buffer memory of the buffer memory section and sends out the detection output; the packet discard detection section; Each output of the encoding type detection section and the buffer use detection section is supplied, and when a packet is discarded, a pseudo packet is generated to be allocated to the packet discard position, and this is supplied to the buffer memory section to be allocated to the packet discard position. an information allocation processing unit that controls reading of packet information signals of image packets to the buffer memory unit based on instructions from the information allocation processing unit when the usage rate of the buffer memory is in a normal state or a high load state; a buffer read control unit that performs the above-mentioned buffer readout control unit; a frame write control unit that controls writing of the packet information signal read from the frame memory unit and also performs interpolation processing if a packet is discarded; 1. An image packet receiving device comprising: a frame memory for encoding; and a frame readout control section for controlling readout of image frame signals from the frame memory for intraframe encoding.