WO2010060376A1

WO2010060376A1 - Video frame type identification method and device

Info

Publication number: WO2010060376A1
Application number: PCT/CN2009/075123
Authority: WO
Inventors: 陈玉芳; 李斌; 谷安文; 周素萍
Original assignee: 华为技术有限公司
Priority date: 2008-11-25
Filing date: 2009-11-25
Publication date: 2010-06-03
Also published as: CN101426137A; CN101426137B

Abstract

A video frame type identification method and device are provided, wherein, the method includes that: receiving the video frames to be identified; obtaining the size of the received video frames to be identified; and determining the types of the video frames to be identified based on the size of the video frames to be identified. The device includes: a receiving unit, for receiving the video frames to be identified; a first statistical unit, for obtaining the size of the received video frames to be identified; and a determining unit, for determining the types of the video frames to be identified based on the size of the video frames to be identified stated by the first statistical unit. The scheme uses the length of the video frames to identify the types of the video frames, without analyzing the head information of each video frame, thereby saving the network overhead.

Description

The invention relates to a Chinese patent application filed on November 25, 2008, the Chinese Patent Application No. 200810203702.4, entitled "A Method and Apparatus for Identifying Video Frame Types". Priority is hereby incorporated by reference in its entirety.

Technical field

The present invention relates to the field of streaming media technologies, and in particular, to a video frame type identification method and apparatus.

Background technique

At present, with the development of wireless communication technology and streaming media technology, streaming video services have become an important application of 3G. Commercial 3G networks can provide rich streaming video on demand and video live broadcast services, and streaming video services. Has been favored by more and more users.

In the streaming video service, the video frame is the smallest unit of streaming media playback, and can be classified into an I frame, a P frame, and a B frame according to the type. A set of consecutive video frames constitutes a GOP (Group of Pictures). A GOP consists of a series of I frames, P frames and B frames with the I frame as the starting frame. Among them, the I frame adopts the intra coding mode, and the P frame and the B frame adopt the interframe coding mode. That is, the B frame refers to the I frame or P frame that is the most recent in the GOP and the nearest I frame or P frame in the GOP when encoding; the P frame refers to the most recent in the GOP before the coding. I frame or P frame. When decoding is performed at the receiving end, I frames that are not referenced to B frames or P frames can be decoded independently, while B frames and P frames must be decoded after the reference frames referenced by them are decoded.

Different types of video frames have different importance, and thus have different effects on video quality. When a video defect occurs in a reference frame, a video frame that references the reference frame at the time of encoding will correspondingly cause a video defect. Only when a video defect occurs in a B frame does not cause a video defect of other video frames.

If the wireless network can identify the type of each video frame, different types of video frames can be separately analyzed and processed according to their types, thereby ensuring that the wireless network better carries the streaming video service.

Although the header information of each video frame includes a type identifier indicating the type of the video frame, the type of the video frame can be accurately read by these types of identifiers, but for each video frame on the network side Parsing header information can take up very large network overhead.

Summary of the invention

Embodiments of the present invention provide a method and an apparatus for identifying a video frame type, which can save network overhead for identifying a type of a video frame.

The embodiment of the invention discloses a method for identifying a video frame type, including:

Receiving a video frame to be identified;

Obtaining a size of the received video frame to be identified;

Determining a type of the to-be-identified video frame according to a size of the to-be-identified video frame.

An embodiment of the present invention further discloses an apparatus for identifying a type of a video frame, including:

a receiving unit, configured to receive a video frame to be identified;

a first statistic unit, configured to obtain a size of the received video frame to be identified;

And a determining unit, configured to determine, according to a size of the video frame to be identified that is counted by the first statistical unit, a type of the video frame to be identified.

As can be seen from the above embodiments of the present invention, since the lengths of different types of video frames are different, the network side uses the length of the video frame to identify the type of the video frame, and the type of the video frame is identified by parsing the header information. More network overhead.

DRAWINGS

FIG. 1 is a flowchart of a method for identifying a video frame type according to a first embodiment of the present invention; FIG. 2 is a flowchart of a method for identifying a video frame type according to Embodiment 2 of the present invention; Schematic diagram of packetization and bearer of a medium video frame;

4 is a flowchart of a method for determining a current statistical period in the present invention;

FIG. 5 is a structural diagram of a video frame type identification apparatus according to Embodiment 3 of the present invention. detailed description

The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

Embodiment 1

Referring to FIG. 1 , FIG. 1 is a flowchart of a method for identifying a video frame type according to an embodiment of the present invention, including the following steps:

Step 101: Receive a video frame to be identified and collect a length value of the to-be-identified video frame. The video frame is decomposed on the network side into a service data unit SDU for transmission in the network, so its length value can be obtained by counting the length values of all SDUs of the video frame.

For example, the service data unit SDU is received; the SDUs belonging to the same video frame are identified; and the length values of the SDUs belonging to the same video frame are added and summed.

Step 102: Identify a type of the to-be-identified video frame according to the length value of the to-be-identified video frame.

The embodiment of the present invention further provides a method for identifying a video frame type, which may include the following steps: (1) receiving a video frame to be identified;

(2) obtaining a length value or a size of the received video frame to be identified;

(3) determining a type of the to-be-identified video frame according to a length value or a size of the to-be-identified video frame.

The inventors have found that there are certain differences in the length values or sizes of different types of video frames, and the difference can be used to identify the type of video frames. For example, in the case of the same frame rate, the maximum value of the B frame is generally smaller than the minimum value of the P frame, and the maximum value of the P frame is generally smaller than the minimum value of the I frame. Therefore, the size of this video frame can be used to identify the type of video frame.

Specific embodiments thereof, for example:

Determining a first threshold and a second threshold, where the first threshold is a value between a B frame length value and a P frame length value, and the second threshold is between a P frame length value and an I frame length value The value of the video frame is determined when the length value of the video frame is less than or equal to the first threshold, or when the length value of the video frame is less than or equal to the first threshold and less than the maximum value of the B frame. B frame; when the length value of the video frame is greater than or equal to the second threshold, or when the length value of the video frame is greater than or equal to the second threshold and greater than the minimum value of the I frame, determining that the video frame is I a frame; when the length value of the video frame is greater than the first threshold and less than the second threshold, determining that the video frame is a P frame.

The first threshold value remains unchanged when the jitter of the current code rate is relatively small or does not change, and is a value between the maximum value of the B frame and the minimum value of the P frame within the current period.

In order to obtain the first threshold, identifying the type of the B frame may be implemented by: determining a current period; counting a first average value of the length of the received video frame in the current period; comparing the first coefficient The value obtained by multiplying the first average value and the maximum value of the B frame are set to the first threshold value. When the length value of the video frame is smaller than the first threshold, determining that the video frame is a B frame, where 0<first coefficient <0.6, and in the encoding mode of H.263, MPEG4, H.264 Decrease in turn.

The second threshold is also constant when the jitter of the code rate is relatively small or does not change, and is a value between the minimum value of the I frame and the maximum value of the P frame within the current period.

In order to obtain the second threshold, identifying the type of the I frame may be implemented by: determining a current period; counting a first average of the length of the received video frame in the current period; comparing the second coefficient by the first average The value obtained by the value and the preset minimum threshold of the I frame, and the maximum value of the two is set as the second threshold. When the length value of the video frame is greater than the second threshold, determining the The video frame is an I frame, where 2 <the second coefficient <4, and is sequentially incremented in the encoding modes of H.263, MPEG4, and H.264.

The current period is usually related to the change of the code rate. When the jitter of the code rate is serious or when the code rate changes, the current period needs to be changed accordingly, so that the video frame type can be more accurately identified. The start time of the current cycle can be set to the end time of the previous cycle, and the end time of the current cycle can be set to a time before the code rate is changed, or shorter.

The determining process of the current period includes, for example: setting an initial period value for the current period; estimating the number of consecutively received B frames and the number of consecutively received I frames in the initial period; if the tth third threshold

The number of continuously received B frames is greater than a third threshold, and the current period is updated to a frame rate of the video frame, where the t1 is from the initial period to the discovery that the number of consecutively received B frames is greater than a third threshold. The duration of the time period, the third threshold is the maximum number of consecutive B frames when the code rate is unchanged; or, if the number of consecutively received I frames is greater than the fourth threshold, the current period is The update is _n , where the t ₂ is a time duration from when the initial period is counted until the number of frame rate frames of the I video frame that are continuously received is greater than the fourth threshold, and the fourth threshold is the same as the code rate. Continuous situation The maximum number of I frames appears. The estimating the number of consecutively received B frames and the number of consecutively received I frames includes: calculating a second average value of video frame lengths received in the initial period; and counting in the initial period And continuously counting the number of video frames whose length is smaller than the product of the first coefficient and the second average is used as the number of B frames; in the initial period, the length of the continuous reception is greater than the second coefficient and The number of video frames of the product of the second average is taken as the number of I frames.

For the first coefficient and the second coefficient, an appropriate value can be determined empirically. The size of the first coefficient and the second coefficient are related to the encoding mode. For different video files using the same encoding format, their first coefficients are generally equal, and the second coefficients are generally equal, therefore The first coefficient and the second coefficient corresponding to different coding modes may also be determined by parsing the header information of the video frame of the video file of different coding modes in advance. The minimum threshold of the I frame is the minimum value of the length of the I frame in a video file, and the maximum threshold of the B frame refers to the maximum value of the length of the B frame in a video file, which corresponds to the same coding mode. For a video file, the I frame minimum threshold and the B frame maximum threshold should be the same. These two values can be obtained by parsing files of different encoding methods in advance.

The first threshold and the second threshold may also be obtained by other embodiments in the embodiment of the present invention. For example, you can first obtain the maximum and minimum values of the P frame, the minimum value of the I frame, the maximum value of the B frame, and then select between the minimum value of the I frame and the maximum value of the P frame by parsing the header information. A value is used as the first threshold, and a value between the maximum value of the B frame and the minimum value of the P frame is taken as the second threshold, and then the first coefficient and the second coefficient are calculated.

It can be seen from the foregoing embodiment that, since the lengths of different types of video frames are different, the network side uses the length of the video frame to identify the type of the video frame, and does not need to parse the header information of each video frame, thereby saving The network overhead. Embodiment 2

Referring to FIG. 2, FIG. 2 is a flowchart of a method for identifying a video frame type according to Embodiment 2 of the present invention, where the method includes the following steps:

Step 201: Determine the current period. The current period is usually related to the change of the code rate. When the code rate changes, the current period needs to be changed accordingly, so that the video frame type can be identified more accurately.

Please refer to FIG. 4, which is a method for determining the current period, and specifically includes the following steps.

Step 401: Set an initial period value for the current period;

The initial period value may be a previous period value, or a preset system default value. The system default value is determined by combining a reference video file GOP length and a frame rate, and generally, Is is selected, but the present invention Embodiments are not limited to this value.

Step 402: Set parameter i and parameter j to 0 respectively, where parameter i is used to count the number of consecutive B frames, and parameter j is used to count the number of consecutive I frames;

Step 403: The average value of the received video frame length values in the initial period value is calculated. The calculation formula of the average value of the received video frame length values in the initial period value is:

N _me = total length of the received SDU within the initial period value / total number of video frames received during the initial period.

Step 404: The length N _/ram of the received video frame in the initial period is sequentially counted; this step is a loop process. Starting from the initial period, the length values of the received video frames are sequentially calculated. After processing one video frame, the next video frame continues to be received until all video frames in the initial period are processed. If the parameter i is greater than the third threshold or the parameter j is greater than the fourth threshold during the loop, indicating that the code rate changes, the value of the current period is updated, and the loop of this step is suspended.

The length value of the video frame can be found by adding the SDUs belonging to the same video frame. In addition to using the video frame boundary recognition method to identify the SDUs belonging to each video frame received in the current statistical period and combining the video frames, the same RTP packet divided into the same video frame may be inserted into the same at the transmitting end. The frame identifier of the timestamp, and the network side performs video frame combination according to the frame identifier.

Step 405: Determine whether the length of the received video frame in the initial period is less than the product of the first coefficient and the average of the length of the received video frame in the initial period, and if so, proceeds to step 406, and if no, proceeds to step 408; Step 406: Add parameter i to 1, and j to 0;

Step 407: Determine whether the parameter i is greater than the third threshold. If yes, go to step 411. If no, return to step 404, calculate the length value of the next video frame, and continue to judge the next video frame until all the initial periods. The video frames are all judged;

The third threshold is determined by the file type and the coding quantization factor. The preferred value of this embodiment is 6, but the present invention is not limited to this value.

Step 408: Determine whether the video frame is greater than a product of the second coefficient and an average value of the received video frame length values in the initial period. If yes, go to step 409; if no, return to step 404 to continue to determine the next video frame until All the video frames in the initial period are judged;

Step 409: Add parameter j to 1 and set i to 0;

Step 410: Determine whether the parameter j is greater than the fourth threshold, and if yes, proceed to step 412; if no, return to step 405 to continue to determine the next video frame until all video frames in the initial period are judged to be one-, value It is 6, but the invention is not limited to this value. Step 411: End this cycle;

The frame rate of the video frame is the playback rate of the video frame, and the number of video frames played in a unit time is used to indicate that when the parameter i is greater than the third threshold, the code rate is considered to be jittered, and the length of the video frame after the code rate is changed. The value will change significantly. In this case, the value of the current period needs to be updated accordingly, and the initial period value cannot be used. t fourth threshold

Step 412: The current period is the frame rate of the video frame, and the current loop is ended. When the parameter j is greater than the fourth threshold, the code rate is considered to be jittered, and the current period needs to be changed accordingly, and the initial period cannot be adopted. Period value.

When the value of the current period is determined in any of the above steps 411, 412, if the steps 411, 412 are not executed, the current period value determined in step 401 is used to return to the identification view. In the flow of the frequency frame type, step 202 is continued.

Step 202: The network side receives an SDU (Service Data Unit) in the determined current period.

The transport protocols used by the streaming video service include RTP (Real-Time Transport Protocol), RTP (RTP Control Protocol), and Real-Time Streaming Protocol (RTSP). The protocol for audio and video data is RTP. The sender divides a video frame into several RTP packets. The RTP packets of these application layers are transmitted on the network side by using the SDU as a carrier. Referring to FIG. 3, FIG. 3 is a schematic diagram of packetization and bearer of a video frame in the prior art.

Step 203: Count an average value of length values of video frames received in the current period;

The total length value of the received SDU and the total number of video frames in the current period may be calculated according to the following: The average value of the received video frame length in the current period is: = the total length of the received SDU in the current period / The total number of video frames received during the current period. For example, in the current period, the network side receives 2 video frames, and the total length of the SDU is ₁₀ bytes, then W _{me = 10/2=5 bytes} . Step 204: The length values of the respective video frames in the current period are sequentially counted.

Identifying each video frame that belongs to the current period by the method of video frame boundary recognition

The SDU adds the length values of the SDUs of the video frames to obtain the length values of the video frames.

N _f the above step, in addition to the method using the video frame boundary identification belonging to the identified SDU than the respective video frame received in the current cycle and the combined video frame, may also be segmented in the same video frame at the transmitting side RTP A frame identifier having the same time stamp is inserted into the packet, and the network side performs video frame combination according to the frame identifier.

This step is a cyclic process. After the type of a video frame is recognized, the length value of the next video frame is continuously calculated, and the type is identified until all the video frame types in the current period are recognized. Step 205: determining whether the average value of the length of a video frame received in the current period is less than or equal to the first threshold, and if so, proceeding to step 206, if no, proceeding to step 207;

Step 206: Determine that the video frame is a B frame, return to step 204, and continue to determine the next video frame received in the current period, until all the video frames received in the current period are judged;

After determining the type of all the video frames received in the current period, return to step 201, continue to determine the next period value, and continue to determine the type of the received video frame in the next period until all video frames All types are judged.

Step 207: Determine whether the video frame is greater than or equal to the second threshold, and if so, proceed to step 208, if no, proceed to step 209;

Step 208: Determine that the video frame is an I frame, return to step 204, and continue to determine the next video frame received in the current period, until all the video frames received in the current period are judged;

Step 209: Determine that the video frame is a P frame, and return to step 204 to continue to determine the next video frame received in the current period until all video frames received in the current period are judged.

After the network side identifies the type of the video frame, the video frame with the higher priority can be preferentially processed according to the priority set for each type of video frame in advance, thereby ensuring the transmission of the video frame with the higher priority. quality.

It can be seen from the foregoing embodiment that the lengths of different types of video frames are different, and the network side compares the lengths ^N of the video frames obtained in the current statistical period with the first threshold and the second threshold respectively, which is smaller than the first At the threshold value, it is determined that the video frame corresponding to the W from ⁶ is a B frame, and when ^{N is} greater than the second threshold, it is determined that the corresponding video frame is an I frame, and when the above two conditions are not satisfied, the determination is The video frame corresponding to the WTM is a P frame, thereby saving the network overhead. The identification of the video frame type is implemented, and the network side can separately process the video frame according to the type of the video frame. At the same time, the network side performs analysis and processing on different types of video frames according to the type of the video frame to ensure that the wireless network better carries the streaming video service. Embodiment 3

Referring to FIG. 5, FIG. 5 is a structural diagram of a video type identification apparatus according to Embodiment 3 of the present invention, where the apparatus includes a first statistical unit 501 and an identification unit 502. The internal structure and the relationship are further described below in conjunction with the working principle of the device.

a first statistic unit 501, configured to calculate a length value of the video frame to be identified;

The first statistical unit 501 includes:

a receiving unit, configured to receive a service data unit SDU; an identifying unit, configured to identify an SDU belonging to the same video frame;

And a summing unit, configured to add and sum the length values of the SDUs belonging to the same video frame. The identifying unit 502 is configured to identify, according to the length value of the video frame to be identified that is counted by the first statistic unit 501, the type of the video frame to be identified.

The identification unit 502 further includes:

The first determining unit 5021, the second determining unit 5022, and the frame type determining unit 5023.

The first determining unit 5021 is configured to determine a first threshold, where the first threshold is a value between a B frame length value and a P frame length value;

In a specific implementation scenario, the first determining unit 5021 may further include: a current period determining unit, configured to determine a current period; and a second statistic unit, configured to collect a current current determined by the current period determining unit The average value of the length of the received video frame is used as a first average value; the first setting unit is configured to compare the value obtained by multiplying the first coefficient by the first average value with the preset maximum threshold value of the B frame, and The minimum of the two is set to the first threshold.

a second determining unit 5022, configured to determine a second threshold, where the second threshold is a value between a P frame length value and an I frame length value; The second determining unit 5022 may further include: a current period determining unit, configured to determine a current period; and a second statistic unit, configured to collect a current current determined by the current period determining unit, in a specific implementation scenario. The average value of the length of the received video frame in the period is used as a first average; the second setting unit is configured to compare the value obtained by multiplying the second coefficient by the first average value with a preset minimum threshold value of the I frame, The maximum of the two is set to the second threshold.

The current period determining unit includes: a preset unit, configured to set an initial period value for the current period; and an estimating unit, configured to estimate a number of consecutively received B frames in the initial period and consecutively received I frames a first update unit, configured to determine a continuously received t third threshold if the estimating unit estimates

The number of B frames is greater than a third threshold, and the current period is updated to a frame rate of the video frame, where the t1 is continued from the initial period counting until the number of consecutively received B frames is greater than a third threshold. The third threshold is the maximum number of consecutive B frames in the case where the third threshold is the same as the code rate; the second updating unit is configured to: if the estimating unit estimates the number of continuously received I frames is greater than the fourth threshold , will t the fourth threshold

The current period is updated to a frame rate of a video frame, where the t2 is a time period from when the initial period is counted to when the number of consecutively received I frames is greater than a fourth threshold, and the fourth threshold is a code rate. The maximum number of consecutive I frames occurs without change.

In a specific implementation scenario, the foregoing estimating unit may further include: a third statistic unit, configured to collect an average value of video frame lengths received in the initial period as a second average value; and a fourth statistic unit, configured to: Counting, in the initial period, the number of consecutively received video frames whose length is smaller than the product of the first coefficient and the second average is the number of B frames; the fifth statistical unit is used for statistics in the In the initial period, the number of consecutively received video frames whose length is greater than the product of the second coefficient and the second average value is taken as the number of I frames.

The frame type determining unit 5023 is configured to: when the video frame to be identified is less than or equal to the first threshold, determine that the to-be-identified video frame is a B frame; and when the to-be-identified video frame is greater than or equal to the second threshold Determining that the video frame to be identified is an I frame; when the video frame to be identified is larger than the When the first threshold is smaller than the second threshold, it is determined that the to-be-identified video frame is a P frame. The embodiment of the invention further provides a video type identification device, which may include:

a receiving unit, configured to receive a video frame to be identified;

a first statistic unit, configured to obtain a length value or a size of the received video frame to be identified, and a determining unit, configured to determine, according to a length value or a size of the video frame to be identified that is counted by the first statistic unit Identify the type of video frame.

The determining unit includes:

a first determining unit, configured to determine a first threshold, where the first threshold is a value between a B frame maximum value and a P frame minimum value;

a second determining unit, configured to determine a second threshold, where the second threshold is a value between a maximum value of the P frame and a minimum value of the I frame;

a frame type determining unit, configured to: when a length value or a size of the to-be-identified video frame is not greater than the first threshold, or when a length value or a size of the video frame is not greater than a first threshold and less than a B frame Determining, when the maximum value, the video frame to be identified is a B frame; when the length value or the size of the to-be-identified video frame is not less than the second threshold, or when the length or size of the video frame is not less than When the second threshold is greater than the minimum value of the I frame, the video frame to be identified is determined to be an I frame; when the length value or the size of the to-be-identified video frame is greater than the first threshold and less than the second threshold, Determining that the to-be-identified video frame is a P-frame.

The first determining unit further includes:

a current period determining unit, configured to determine a current period;

a second statistic unit, configured to collect an average value of a video frame length value or a size received in the current period determined by the current period determining unit, and use the average value as a first average value;

And a first setting unit, configured to compare a value obtained by multiplying the first coefficient by the first average value and a maximum value of the B frame, and setting a minimum value of the two as the first threshold.

The second determining unit further includes:

a current period determining unit, configured to determine a current period;

a second statistic unit, configured to collect an average value of lengths of video frames received in a current period determined by the current period determining unit, and use the average value as a first average value; And a second setting unit, configured to compare a value obtained by multiplying the second coefficient by the first average value with a minimum value of a preset I frame minimum value, and setting a maximum value of the two as the second threshold value.

As can be seen from the foregoing embodiment, after the first statistic unit counts the length of the video frame, the identification unit determines the type of the video frame by using the length of the video frame counted by the first statistic unit, when the video frame is less than or equal to Determining, by the first threshold, that the video frame is a B frame; when the video frame is greater than or equal to a second threshold, determining that the video frame is an I frame; when the video frame is greater than the first threshold and less than When the second threshold is described, it is determined that the video frame is a P frame. It avoids the wasted network overhead caused by the network parsing the header information of each video frame. At the same time, the network side performs analysis and processing on different types of video frames according to the type of the video frame to ensure that the wireless network better carries the streaming video service.

The embodiments of the present invention can also be applied to the fields of video quality assessment, video quality assurance, and the like.

One of ordinary skill in the art can understand that all or part of the steps of implementing the above embodiments may be completed by a program instructing related hardware, and the program may be stored in a computer readable storage medium, and the program includes the following Step: receiving a video frame to be identified and collecting a size of the to-be-identified video frame; and identifying a type of the to-be-identified video frame according to the size of the to-be-identified video frame. The storage medium is, for example, a ROM/RAM, a magnetic disk, an optical disk, or the like.

The above is only a preferred embodiment of the present invention, and it should be noted that those skilled in the art can also make several improvements and retouchings without departing from the principles of the present invention. It should be considered as the scope of protection of the present invention.

Claims

OP090604 WO 2010/060376 PCT/CN2009/075123 - 14- Claims

A method for identifying a video frame type, the method comprising:

Receiving a video frame to be identified;

Obtaining a size of the received video frame to be identified;

The method according to claim 1, wherein the determining the type of the to-be-identified video frame according to the size of the to-be-identified video frame further includes:

Determining a first threshold, the first threshold being a value between a B frame maximum value and a P frame minimum value;

Determining a second threshold, the second threshold being a value between a P frame maximum value and an I frame minimum value; when the size of the to-be-identified video frame is not greater than the first threshold, or when When the size of the video frame is not greater than the first threshold and less than the maximum value of the B frame, determining that the to-be-identified video frame is a B frame;

When the size of the to-be-identified video frame is not less than the second threshold, or when the size of the video frame is not less than a second threshold and greater than a minimum value of the I frame, determining that the to-be-identified video frame is I And determining, when the size of the to-be-identified video frame is greater than the first threshold and less than the second threshold, determining that the to-be-identified video frame is a P-frame.

3. The method according to claim 2, wherein the determining the first threshold further comprises:

Determine the current cycle;

Counting an average of the lengths of the received video frames in the current period, and using the average value as the first average value;

The value obtained by multiplying the first coefficient by the first average value and the maximum value of the B frame are compared, and the minimum of the two is taken as the first threshold.

4. The method according to claim 2, wherein the determining the second threshold further comprises:

Determine the current cycle;

Counting the average of the lengths of the received video frames in the current period, and using the average as the first OP090604

WO 2010/060376 PCT/CN2009/075123

-15- an average value;

Comparing the value obtained by multiplying the second coefficient by the first average value with a preset minimum threshold of the I frame, and using the maximum value of the two as the second threshold.

The method according to any one of claims 3 or 4, wherein the determining the current period comprises:

Setting an initial period value for the current period;

Obtaining a number of B frames continuously received in the initial period and a number of consecutively received I frames; if the number of consecutively received B frames is greater than a third threshold, updating the current period to a third threshold of _t

^Κ fps video frames, the tl is started from the initial period to find the number of B-frame timing DRX threshold is greater than a third duration, said third threshold is a case where the rate constant The maximum number of B frames consecutively appears; or, if the number of consecutively received I frames is greater than a fourth threshold, updating the current period to a fourth threshold of _t

^—Video frame ^] frame rate, wherein the _t2 is a time period from when the initial period is counted until the number of consecutively received I frames is greater than a fourth threshold, and the fourth threshold is a code rate. In the case of a change, the maximum number of I frames appears continuously.

6. The method of claim 5, wherein

Counting an average of video frame sizes received during the initial period and using the average as a second average;

And obtaining the number of B frames continuously received in the initial period, further comprising: counting, in the initial period, continuously receiving a video frame whose size is smaller than a product of the first coefficient and the second average value The number is the number of B frames;

The obtaining the number of I frames continuously received in the initial period further includes: counting, in the initial period, a video frame that continuously receives a product whose size is greater than a product of the second coefficient and the second average value The number is used as the number of I frames.

7. A device for identifying a type of a video frame, comprising:

a receiving unit, configured to receive a video frame to be identified;

The device according to claim 7, wherein the determining unit comprises: a first determining unit, configured to determine a first threshold, where the first threshold is between a B frame maximum value and a P frame minimum value Value between

a frame type determining unit, configured to: when the size of the to-be-identified video frame is not greater than the first threshold, or when the size of the video frame is not greater than a first threshold and less than a maximum value of a B frame, Determining that the identified video frame is a B frame; or when the size of the video frame to be identified is not less than the second threshold, or when the size of the video frame is not less than a second threshold and greater than a minimum value of the I frame, Determining that the to-be-identified video frame is an I-frame; and determining that the to-be-identified video frame is a P-frame when the size of the to-be-identified video frame is greater than the first threshold and less than the second threshold.

The device according to claim 8, wherein the first determining unit further comprises:

a current period determining unit, configured to determine a current period;

a second statistic unit, configured to calculate an average value of video frame sizes received in the current period determined by the current period determining unit, and use the average value as a first average value;

The device according to claim 8, wherein the second determining unit further comprises:

a current period determining unit, configured to determine a current period;

The device according to any one of claims 9 or 10, wherein the current period determining unit further comprises:

a preset unit, configured to set an initial period value for the current period;

An estimating unit, configured to estimate a number of consecutively received B frames and an consecutively received I frames in an initial period set by the preset unit;

a first updating unit, configured to: if the estimating unit estimates, the number of consecutively received B frames is greater than

t third threshold

And a third threshold, the current period is updated to a frame rate of the video frame, where the t1 is a time from when the initial period is counted to when the number of continuously received B frames is greater than a third threshold. The third threshold is the maximum number of consecutive B frames in the case where the code rate is constant; the second update unit is configured to: if the estimation unit estimates the number of consecutively received I frames is greater than the first

t fourth threshold

a fourth threshold, the current period is updated to a frame rate of the video frame, where the t2 is a time from when the initial period is counted to when the number of consecutively received I frames is greater than a fourth threshold, the fourth time The threshold is the maximum number of consecutive I frames in the case where the code rate is constant.