JP2603931B2 - Image transmission device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Field of Industrial Application) The present invention relates to an image transmission apparatus for transmitting images in real time.

(Prior Art) When image information is transmitted via a digital line, encoding is generally performed. In addition, from the economical point of view, compression is a requirement for image information.

In response to this, there have been various types of encoding means of the conventional image information transmitting apparatus, but the transmission speed of the target line is limited to a specific value or limited to a certain range. There was only a device.

This is because there is a requirement for the encoding device to occupy an area according to the transmission speed of the line. That is, compression efficiency is emphasized in an encoding device with a low transmission speed, while processing speed is emphasized in an encoding device with a high transmission speed. For example, in an encoding device with a low transmission rate, although the processing speed may be low,
The coding efficiency must be high. Conversely, in an encoding device at a high transmission rate, the processing speed must be high even though the encoding efficiency may be low.

For this reason, the conventional image information transmission apparatus limits the target transmission speed as described above. According to the state of the art, the ratio between the maximum and minimum transmission speeds is 1-6. next,
In accordance with the transmission speed, the encoding device is designed in consideration of the above items.

That is, an encoding device at a low transmission rate places importance on encoding efficiency, and an encoding device at a high transmission rate places importance on processing speed.

In such a situation, an attempt to cope with a high transmission rate, that is, to realize a high bit rate operation by an encoding device at a low transmission rate has been impossible because the processing operation cannot follow. Conversely, if the encoding device at a high transmission rate supports a low transmission rate, that is, performs a low bit rate operation, the compression efficiency is low, resulting in extreme deterioration in image quality. In particular, when dealing with moving images, the function as an encoding device could not be performed at all.

Therefore, for example, in order to perform encoding at a transmission speed in the range from 64 Kbps to 1.5 Mbps, it is necessary to use at least two different encoding devices as shown in FIG. Burden was heavy.

Also, when configuring an image communication network, since different types of encoders are scattered on the network as shown in FIG. 8 (b), the connection procedure becomes complicated, and prior to image communication, I had to check the type of device on the other side.

(Problems to be Solved by the Invention) As described above, in the related art, the requirements according to the transmission rate cannot be satisfied at the same time, and an image transmission apparatus that can support a wide range of transmission rates cannot be obtained. . For this reason, there are various drawbacks in terms of handling and equipment burden.

SUMMARY OF THE INVENTION It is an object of the present invention to eliminate the above-mentioned problems of the prior art and to provide an image transmission apparatus capable of supporting a wide range of transmission speeds.

[Configuration of the Invention] (Means for Solving the Problems) The present invention provides an encoding unit that performs orthogonal transformation of image information and then quantizes the image information, and converts the image information encoded by the encoding unit into a first or a first image. A transmission means for transmitting to a line at a transmission rate different from each other,
When transmitting coded image information at a second transmission rate lower than the first transmission rate, the maximum capacity of a transmission buffer required to temporarily store the coded image information is reduced. And a control signal converting means for outputting a control signal for the control.

(Operation) Since the image transmission apparatus of the present invention handles from a low transmission rate to a high transmission rate, the amount of information transmitted to the line changes depending on the selected rate. For example, if the transmission speed drops,
As the amount of information decreases and the transmission speed increases, the amount of information also increases.

In the present invention, among the elements defining an image, the resolution determined by the number of images and the accuracy of quantization and the minimum value of the number of screens within a predetermined time, the so-called frame drop amount, are focused on, and the transmission speed is Instead of proportionally changing the amount of dropped frames, the resolution decreases even at a low transmission speed, and the minimum value of the number of screens is reduced at a fixed rate, that is, the amount of dropped frames is increased. .

As a result, the movement of the image is slightly awkward due to the increase in the number of dropped frames, but the image quality is maintained and the image quality does not deteriorate.

Next, an embodiment of the present invention will be described with reference to the drawings.

The image transmission device in this embodiment is from 64 Kbps to 1.5 Mbps.
Up to 24 levels of transmission speed are realized with a single hardware configuration.

The image transmission apparatus in this embodiment handles image information as digital signals. That is, the image information input to the input terminal (11) is an analog signal, and the A / D converter (1
It is converted to a digital signal in 3). Of course, when the image information is supplied as a digital signal, the A / D converter (1
3) is unnecessary.

The output of the A / D converter (13) is
Supplied to 5). The output of the sub-sampling means (15) is supplied to a frame memory (17). The data read from the frame memory (17) is supplied to a two-dimensional cosine transform operation circuit (19). The output of the circuit (19) is sent to the block selector (23) together with the quantizer (21). The output of the quantizer (21) is supplied to a buffer (25). The output of the buffer (25) is supplied to the transmission means (27). Also, the transmission means depends on the transmission speed of the line.
That ability is selected. Here, the transmission speed is selected by the transmission speed selection means (29) in 24 steps (64 Kbps to 1.5 Mbp).
s). At the same time, information on which transmission rate is selected is supplied to the control signal conversion means (31).
The control signal conversion means (31) generates various control signals according to the transmission speed and supplies them to the block selection means (23), the quantizer (21), the frame memory (17), and the sub-sampling means (15). I do.

Next, the operation will be described. The image information supplied to the input terminal (11) is obtained according to a predetermined resolution. This information is digitized and subjected to sub-sampling in sub-sampling means (15). The sub-sampling here is performed according to the resolution of the image transmitted to the destination. Moreover, in this embodiment, sub-sampling is performed at a non-linear rate with respect to the transmission speed.

For example, the sub-sampling interval is set as shown in FIG. That is, when the transmission speed is 384 Kbps or more, the pixel is output every ΔT in synchronization with the input,
If the transmission speed is less than 384 Kbps, the signal is output every 2ΔT. That is, at 384 Kbps or more, the input image is output as it is, and at less than 384 Kbps, every other image is output. This results in thinning one by one.

As a specific example, if one frame of the input image is 500 × 500 pixels, if it is less than 384 Kbps, it will be 250 × 250 pixels.

Here, attention is paid to the relationship between the subsampling interval and the transmission speed. As described above, when dealing with an area where the transmission speed is high to low, what matters is the coding efficiency and the processing speed. When the transmission speed is 54 Kbps and 1.5 Mbps, the ratio is about 24. Therefore, the amount of information at 1.5Mbps is 6
At 4Kbps, the amount of information must also be reduced by a factor of 24.

Therefore, for sampling (thinning), 64Kbps
Then, it is normal to make it 1/24. However, in this embodiment, the thinning amount is saturated non-linearly with respect to the transmission speed, particularly at a low transmission speed, instead of proportionally decreasing the information amount. This is because if the ratio is reduced proportionally, the image quality is significantly deteriorated. However, there is a problem in the amount of information if it is left as it is, but this is dealt with by dropping frames as described below.

The signal thinned out by the sub-sampling means (15) is dropped in the frame memory (17) as shown in FIG. That is, signals are written and read to and from the frame memory (17) in frame units. At 384 Kbps or more, all input frames are written to and read from the frame memory (17). On the other hand, when the transmission speed is 64 Kbps or less, only one out of three frames is written in the frame memory (17). If it is greater than 64Kbps and less than 384Kbps,
For every two frames, only one frame is written in the frame memory (17). This dropped amount is also changed nonlinearly with respect to the transmission speed.

Next, the signal in which the frame is adaptively dropped is cosine-transformed by a two-dimensional cosine transform operation circuit (19). As this circuit (19), for example, the circuit shown in the 1986 National Convention of the TV Society, 6-14, "Cosine Transform Circuit" may be used. Here, for example, input pixels are divided into blocks of 8 × 8 pixels, and cosine transform is performed.

The signal that has undergone the cosine transform is quantized by a quantizer (21).
Quantization is performed and the data is transmitted. However, in this embodiment, not all the signals subjected to the cosine conversion are transmitted for transmission, but only those whose information amount after the cosine conversion is equal to or more than a predetermined threshold are transmitted. That is, the presence or absence of information transmission is changed in the space subjected to the cosine transform. Although the amount of information is compressed by cosine transform,
As described above, control in the transform space further improves compression.

In this embodiment, the threshold value used as a criterion for block selection here is controlled by the transmission rate. That is, if the transmission speed is high, the threshold is lowered, and if the transmission speed is low, the threshold is raised. For example, a block is selected as shown in FIG. Here, the blocks are divided as shown in FIG. 4A in the space after the cosine transform. This is because the low-frequency components of the image are distributed in the upper left, and as the frequency increases, the components are distributed in a band form from the upper left to the lower right as shown in FIG. It is divided as shown in the figure.

If the threshold for block selection is reduced when the blocks are divided in this way, an area as shown in FIG. On the other hand, when the threshold value is increased, an area as shown in FIG. Specifically, for example, 1.5M
In bps transmission, a necessary part is selected from the area from FIG. 4 (a).
In s, the part from is selected.

On the other hand, the signal of the block permitted to be transmitted is quantized by the quantizer (21). Here, how many bits of the input signal are realized is controlled. That is, if the transmission speed is high, the number of bits is increased, and if the transmission speed is low, the number of bits is reduced. For example, in the same area in FIG. 4 (a), 30 bits are used for 1.5 Mbps transmission, but 20 bits for 384 Kbps and 15 bits for 64 Kbps.

Here, the change of the quantization and the change of the criterion of block selection directly affect the resolution. Therefore, in this embodiment, these various amounts and the amount of sub-sampling are controlled so as to be kept at a certain value or more even at a low transmission speed in consideration of affecting the resolution (image quality). FIG. 5 shows this state.

The signal quantized by the quantizer (21) is buffered (2
Supplied to 5). The buffer (25) includes a RAM (33) and an address control unit (35). Where the quantizer (2
The signal instructing the output from 1) is sent to the address control unit (3
Supplied to 5). In response, the address control unit (35)
Now, generate the address. The RAM (33) sequentially stores the output of the quantizer (21) according to the address.

However, in this embodiment, the maximum capacity of the RAM (33) is
It is changed according to the transmission speed. This is shown in Table 1.

As described above, when the maximum capacity of the buffer (25) (RAM (33)) is changed, the following effects are obtained.

The amount of information (code) to be transmitted changes according to the transmission speed. For example, if the transmission speed is high, the amount of information increases,
If the transmission speed is low, the amount of information is small. At this time, if you want to set the maximum capacity of the buffer (25) constant,
Of course, it is necessary to determine the maximum capacity on the assumption that the amount of information is maximum. Therefore, the maximum capacity of the buffer (25) is
Optimally set for high transmission rates.

If the transmission is performed at a low transmission speed when the capacity is set, it takes a considerable time before the buffer (25) becomes full. That is, the control delay due to the buffer accumulation amount occurs, and a still image is continued for several seconds even though the moving image is transmitted.

On the other hand, in the present embodiment, the maximum capacity of the buffer (25) is reduced in accordance with the transmission speed, so that there is no time delay in transmission, and furthermore, the threshold of the significant block selection or the like based on the buffer accumulation amount. Control is also performed smoothly.

Although the maximum capacity of the buffer (25) is changed according to the transmission speed, as shown in Table 1, the maximum capacity does not have to be proportional to the transmission speed.

This is because the lower the bit rate, the greater the number of dropped frames, resulting in variations in the amount of generated codes. At a high bit rate, the buffer memory ratio can be set to be slightly smaller. That's why.

In the above description, the buffer (25) is a RAM
It is not necessary to configure by (33).

When the signal stored in the RAM (33) reaches its variable maximum capacity, a signal instructing output is supplied from the address control unit (35) to the transmitting means (27). If the transmission means (27) can transmit, the address control unit (35) reads out the signals stored in the RAM (33) in the stored order. The read signals are sequentially supplied to the transmission means (27) and transmitted to a digital line.

Next, a specific configuration for variably setting the variable maximum capacity of the buffer (25) will be described.

As described above, in the device of this embodiment, 64 Kbps to 1.5 M
Transmission speeds up to bps are selected in 24 steps. This selection is made, for example, by a transmission rate selection means (2
This is done in 9). From this means (29), the selected transmission rate is supplied to the control signal conversion means (31) as, for example, 5-bit information. The control signal conversion means (31) is a conversion table. That is, transmission speed selection means (29)
Then, a value indicating the maximum capacity of the buffer (25) is output from the signal indicating the specific transmission speed. For example, if 64 Kbps is selected as the transmission speed, the control signal converter (31) outputs a maximum capacity value of 8 Kbits to the address controller (35). Similarly, when the transmission speed is greater than 64 Kbps and less than 384 Kbps, a maximum capacity value of 48 Kbits is output. Greater than 384Kbps, 1.5
At Mbps, the maximum capacity value of 128K bits is output. Here, the maximum capacity value of the buffer (25) may be linearly changed between 64 Kbps, 384 Kbps, and 1.5 Mbps shown in Table 1.

In response to such a maximum capacity value, the address control unit (35)
Then, the address is controlled. That is, upon receiving the designated maximum capacity value, the RAM (33) is compared with the write address. When the writing area reaches the maximum capacity value, the writing is stopped. This buffer is always read out to the transmission means (27).

On the other hand, in the control signal conversion means (31), not only the address control unit (35) but also control signals to the sub-sampling means (15), the frame memory (17), the quantizer (21), and the block selection means (23) create. For example, the subsampling unit (15) is instructed on the amount of thinning for the transmission speed. In the above description, the control of the sub-sampling interval has been described as shown in FIG. To the frame memory (17), the amount of dropped frames is supplied as a control signal as shown in FIG. Quantizer (21)
Indicates the number of quantization bits. In addition, information on a threshold value used as a criterion for block selection is transmitted to the block selection means (23) with respect to the transmission speed.

As described above, in the image transmission device of this embodiment, the operation and characteristics of each component of the component are controlled in accordance with the transmission speed. Moreover, the control is performed nonlinearly with respect to the transmission speed for each component. That is, for the functional components that control the resolution, the capability is not reduced in proportion to the transmission speed, but is kept at a certain value or more even at a low transmission speed, while the information in the time axis direction is kept. The amount reduces the amount of information beyond a value proportional to the transmission rate.

Thus, at a low transmission rate, the motion is slightly awkward, but the resolution is kept good.

Here, reference will be made to the case where a network is formed by the image transmission device in this embodiment. When a device at a certain point is used at a certain speed, the transmission speed of the line is selected by the transmission speed selecting means (29). Prior to transmitting / receiving a significant image signal, this apparatus sends the transmission speed selected by itself to the other image receiving apparatus. After the transmission speed is confirmed between the transmission and reception devices, the image signal is transmitted and received. Usually, an image transmitting apparatus and an image receiving apparatus are integrally provided and provided at each point.

 Next, another embodiment of the present invention will be described.

In this embodiment, a signal subjected to cosine transform (hereinafter, referred to as a converted signal) is not subjected to quantization and transmitted as it is, but rather to a converted signal at a certain time and one time before (one sampling time). Since the difference from the converted signal in the previous step is obtained and quantized, the coding efficiency is increased.

When taking the difference, an inverse quantizer (61), an adder (62), a frame memory (63), and a subtractor (64) are used. That is, the output of the quantizer (21) is dequantized by the inverse quantizer (61), and the adder (62) takes the sum of the cosine transform signal two times before. As a result, a converted signal one time earlier is obtained in the frame memory (64). Here, “time” is based on the time at which the subtractor (63) performs the subtraction.

 Next, another embodiment of the present invention will be described.

In this embodiment, differential encoding is performed more efficiently by using motion prediction data for information compression.

That is, using the output of the frame memory (77), the motion compensation circuit (79) detects a motion vector (MV) from the image signal one time earlier. Furthermore, this circuit (79)
In, motion prediction data is obtained from these data. The subtracter (81) calculates the difference between the output of the frame memory (17) and the motion prediction data. This difference is encoded by a cosine transform circuit (19) and a quantizer (21).

On the other hand, the output of the quantizer (21) is the inverse quantizer (71),
The inverse cosine transform circuit (73) converts the signal into a differential signal before being encoded. This signal is sent to the frame memory (7
The signal stored in 7) is added to become the image signal one time before.

[Effects of the Invention] According to the present invention, it is possible to transmit an image whose image quality is maintained even at a low transmission speed in an image transmission device that handles from a low transmission speed to a high transmission speed.

[Brief description of the drawings]

FIG. 1 is a diagram showing an outline of an image transmission apparatus according to an embodiment of the present invention. FIGS. 2 to 5 are diagrams for explaining the operation of each section of the apparatus shown in FIG. FIG. 7 and FIG. 7 show another embodiment of the present invention, and FIG. 8 shows a prior art. 15 Sub-sampling means 17 Frame memory 19 Two-dimensional cosine transform operation circuit 21 Quantizer 23 Block selecting means 25 Buffer 29 Transmission speed selecting means

 ──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-59-97286 (JP, A) JP-A-62-108687 (JP, A) JP-A-63-9391 (JP, A)

Claims (2)

(57) [Claims] An encoding means for orthogonally transforming image information and then quantizing the image information, and sending the image information encoded by the encoding means to a line at first or second different transmission rates. Means for temporarily storing the encoded image information when transmitting the encoded image information at a second transmission rate lower than the first transmission rate. An image transmission apparatus comprising: a control signal conversion unit that outputs a control signal for reducing a required maximum capacity of a transmission buffer. 2. The control signal conversion means outputs a control signal for controlling a quantization size according to the amount of image information stored in the transmission buffer, and the encoding means is designated according to the control signal. 2. The image transmission device according to claim 1, wherein the image information is quantized based on a quantization size.