一种视频序列的编码方法及采用其的差错控制系统 技术领域 本发明涉及图像信息传输和处理领域，具体涉及在进行视频传输过程中的阻止传输差 错扩散的方法。 背景技术 通过 Internet和无线移动网等主要通信网络传输视频时， 现有的视频压缩算法通常采 用基于运动补偿的混合编码框架来提高压缩率。 对于运动图像专家组标准（MPEG )视频 流， 采用三种类型的帧： 帧内编码帧 （以下简称 I帧）、 前向预测编码帧（以下筒称 P帧） 和双向预测编码帧 （以下简称 B帧） 。 I帧是一个全帧压缩编码帧， 它将全帧图像信息进 行压缩编码及传输， 解码时仅用 I帧的数据就可重构完整图像； P帧可以由 I帧或前面的 P 帧进行预测， P帧属于前向预测的帧间编码， 压缩比大于 I帧； B帧以前面的 I或 P帧和后 面的 P帧为参考帧进行预测， 因此压缩比最高。 然而在满足带宽限制条件下， 需要克服传输差错的影响。 如图 1所示为传统的视频编码方式对传输差错的处理方法。在图 1中， 采用帧内编码 模式的图像是可以独立解码的， 其它的编码图像帧采取帧间编码模式。 图 1中所示是一种 典型的视频编码码流结构， 其中在 t时刻和 t+n时刻的图像帧是帧内编码模式， 记为 I, 其 它图像帧都采取用前一帧作为参考帧的帧间编码模式， 记为 P。 当出现传输差错时， 例如在 t+i时刻发生错误， 则 P(t+i)不能正确解码。 由于后续的 帧间编码帧都是用前一帧作为参考图像， 因此错误解码图像会向后扩散， 直到 t+n时刻的 帧内编码图像帧为止。 在这种情况下， 解码错误的时间间隔为 n-i。 传输差错引起视频图像不能正常解码， 而且错误图像信息会在时域和空域上扩散。 为 了降低视频传输差错及其差错扩散的影响， 一种常用的方法是采用帧内编码更新技术， 但 是帧内编码方式会大幅度降低编码效率。 与帧内编码更新相比， 参考帧选择 ( Reference Picture Selection, RPS )具有更高的编 码效率。 其基本思想是编码器通过反馈信道得到解码器的丢失情况， 在编码下一个帧间编 码帧时选用已被解码器正确接收的较早的帧作为参考帧。 Hideaki Kimata和 Yoshiyuki Yashima在文献 "A study of key-frame reference picture selection method for error resilient l
multiple video objects distribution"中提出了一种新的基于关键参考帧的无反馈关键参考帧选 择方法。 即以帧内编码帧（I帧）作为关键参考帧， 每个 P帧均以 I帧作为参考桢。 这种方 法虽然可以有效地阻止差错的扩散，但过多地插入关键帧（I帧）， 导致数据量过大， 编码效 率降低。 发明内容 本发明所要解决的技术问题在于提供了一种视频码流的编码 /解码方法， 通过在编码 码流中插入核心参考桢， 不但可以有效地减少或者阻止视频过程中由于传输差错引起的图 像质量下降， 而且不会明显降低视频图像的编码效率。 为了实现上述目的，本发明提供了一种视频序列的编码方法， 用于视频传输的差错控 制， 包括： 步骤一， 对视频序列第一帧图像采取帧内编码方式进行编码； 步骤二，在视频序列中第一帧图像之后数个位置插入核心参考帧， 所述核心参考帧的 参考帧是前面的 I帧或者核心参考帧； 步骤三， 对视频序列中的非核心参考帧图像按照标准的编码方法进行编码； 步骤四， 对视频序列中的核心参考帧图像采取帧间编码方式编码。 所迷步骤一中采取帧内编码方式的图像独立解码。 TECHNICAL FIELD The present invention relates to the field of image information transmission and processing, and in particular to a method for preventing transmission error diffusion during video transmission. BACKGROUND OF THE INVENTION When video is transmitted over a primary communication network such as the Internet and a wireless mobile network, existing video compression algorithms typically employ a motion-compensated hybrid coding framework to increase the compression ratio. For Moving Picture Experts Group Standard (MPEG) video streams, three types of frames are used: intra-coded frames (hereinafter referred to as I-frames), forward-predicted coded frames (hereinafter referred to as P-frames), and bidirectionally predictive coded frames (hereinafter referred to as frames). B frame). The I frame is a full frame compression coded frame, which compresses and encodes the full frame image information. When decoding, only the I frame data can be used to reconstruct the complete image; the P frame can be predicted by the I frame or the previous P frame. The P frame belongs to the inter prediction of the forward prediction, and the compression ratio is larger than the I frame; the B frame is predicted by using the previous I or P frame and the following P frame as the reference frame, so the compression ratio is the highest. However, under the condition that the bandwidth limitation is met, it is necessary to overcome the influence of transmission errors. Figure 1 shows the traditional video coding method for handling transmission errors. In Fig. 1, images using the intra coding mode are independently decodable, and other coded image frames are in an inter coding mode. Figure 1 shows a typical video encoding code stream structure, in which the image frame at time t and time t+n is the intraframe coding mode, denoted as I, and other image frames are taken with the previous frame as the reference frame. The inter-frame coding mode is denoted as P. When a transmission error occurs, for example, an error occurs at time t+i, P(t+i) cannot be decoded correctly. Since the subsequent inter-coded frames use the previous frame as the reference picture, the erroneously decoded picture will be spread backwards until the intra-coded picture frame at time t+n. In this case, the decoding error interval is ni. Transmission errors cause video images to not be decoded properly, and erroneous image information spreads over time and space. In order to reduce the impact of video transmission errors and error diffusion, a common method is to use intra-frame coding update technology, but the intra-frame coding method will greatly reduce the coding efficiency. Compared to intra-coded updates, Reference Picture Selection (RPS) has higher coding efficiency. The basic idea is that the encoder obtains the loss of the decoder through the feedback channel, and selects the earlier frame that has been correctly received by the decoder as the reference frame when encoding the next inter-frame coded frame. Hideaki Kimata and Yoshiyuki Yashima in the literature "A study of key-frame reference picture selection method for error resilient l A new non-feedback key reference frame selection method based on key reference frames is proposed in multiple video objects distribution. That is, intra-coded frames (I-frames) are used as key reference frames, and each P-frame is referenced by I-frames. Although this method can effectively prevent the spread of errors, excessive insertion of key frames (I frames) results in an excessive amount of data and a low coding efficiency. SUMMARY OF THE INVENTION The technical problem to be solved by the present invention is to provide a The encoding/decoding method of the video stream can not only effectively reduce or prevent the image quality degradation caused by the transmission error in the video process by inserting the core reference frame in the encoded code stream, but also does not significantly reduce the encoding efficiency of the video image. In order to achieve the above object, the present invention provides an encoding method for a video sequence, which is used for error control of video transmission, including: Step 1: encoding an image of a first frame of a video sequence by using an intra-frame coding manner; Step 2: Inserting a core reference frame at a number of positions after the first frame image in the sequence, the core reference frame The reference frame is the previous I frame or the core reference frame; Step 3, encoding the non-core reference frame image in the video sequence according to a standard encoding method; Step 4, adopting an interframe coding manner for the core reference frame image in the video sequence Encoding In the first step, the image in the intra coding mode is independently decoded.
.所述步骤二中插入核心参考帧包括每隔固定间隔插入核心参考帧，或根据率失真优化 模型选择核心参考帧 , 或根据反馈信道传回的确认信息自适应插入核心参考帧。 所迷步驟三中的所述图像是 I帧、 P帧或 B帧。 所述步骤三中标准的编码方法为帧间编码方式。 Inserting the core reference frame in the second step includes inserting the core reference frame at regular intervals, or selecting the core reference frame according to the rate distortion optimization model, or adaptively inserting the core reference frame according to the acknowledgement information returned by the feedback channel. The image in step 3 is an I frame, a P frame or a B frame. The standard coding method in the third step is an interframe coding mode.
'所述帧间编码方式是以当前帧的前一帧作为参考帧进行编码。 所述步骤四中对核心参考帧图像采取帧间编码方式编码是以当前帧之前正确接受的 核心参考帧作为参考帧进行编码。 本发明还提供一种视频传输的差错控制方法， 包括: 对视频序列第一帧图像采取帧内编码方式进行编码;
在视频序列中第一帧图像之后数个位置插入核心参考帧; 对视频序列中的非核心参考帧图像按照标准的编码方法进行编码； 对视频序列中的核心参考帧图像采取帧间编码方式编码； 接收编码的视频序列然后解码并反馈解码信息； 传输编码的视频序列以及反馈的解码信息。 所述在视频序列中第一帧图像之后数个位置插入核心参考帧步骤包括在视频序列中 每隔固定间隔插入核心参考帧， 或根据率失真优化模型选择核心参考帧， 或才艮据反馈信道 传回的确认信息选择核心参考帧。 所述标准的编码方法为采取帧间编码方式编码。 所述帧间编码方式采取以当前帧的前一帧作为参考帧。 所 ¾J?于视频序列中的核心参考帧图像采取帧间编码方式编码步骤中是以当前帧之前 正确接收的核心参考帧作为参考帧进行编码。 本发明还提供一种视频传输的差错控制系统 , 包括： 编码器，用于对视频序列第一帧图像釆取帧内编码方式进行编码，在视频序列中第一 帧图像之后数个位置插入核心参考帧， 对视频序列中的非核心参考帧图像按照标准的编码 方法进行编码并对视频序列中的核心参考帧图像采取帧间编码方式编码； 解码器， 用于接收编码的视频序列然后解码， 并将解码状况反馈到编码器； 信道， 用 于传输编码器编码的视频序列以及反馈解码器解码的信息。 所述信道包括： 传输信道， 用于传输编码起编码的视频序列； 反馈信道， 用于反馈解 码器解码的信息。 所述编码器、 解码器和信道存在时间延迟。 本发明提出了一种通过插入核心参考帧， 可以在不明显降低视频编码效率的条件下， 实现视频传输差错控制的方法。 与原有的在视频序列中插入 I帧或者关键参考帧的方法相 比， 具有更高的编码效率， 而且能够有效减小或者阻止传输差错在视频序列上的扩散， 从 而提高视频传输的质量。
附图说明 图 1所示为传统的视频编码方式对传输差错的处理方法； 图 2所示为核心参考帧方法对传输差错的处理方法； 图 3 所示为采取固定间隔插入核心参考帧的差错控制方法； 图 4 所示为自适应核心参考帧间隔的差错控制方法。 具体实施方式 为了使本发明的目的、技术方案及优点更加清楚明白， 以下结合附图及实施例， 对本 发明进行进一步详细说明。 应当理解， 此处所描述的具体实施例仅仅用以解释本发明， 并 不用于限定本发明。 为了能够在出现传输差错时减少或者阻止错误解码图像的扩散，本发明提出了一种在 基于核心参考帧的编码方法， 可以在不明显降低视频编码效率的情况下， 阻止错误的扩散。 图 2所示为核心参考帧方法对传输差错的处理方法。 在进行视频编码的过程中， 在某些特定位置插入核心参考帧， 记为 C, 如图 2中的双 线所示。 核心参考帧也采取帧间编码方法进行编码， 与普通的帧间编码方法比较， 其不同 之处在于其参考帧是前面的 I帧或者核心参考帧。 例如在图 2中， C ( t+j )是以 I ( t )为参 考帧， C ( t+k )是以 C ( t+j )为参考帧。 当出现传输差错时， 例如在 t+i时刻发生错误， 则 P(t+i)不能正确解码。 由于后续的 帧间编码帧 P ( t+i+1 )是以 P(t+i)作为参考图像， 因此措误解码图像会向后扩散。 在 t+j时刻， 由于核心参考帧 C ( t+j )是以 I ( t )作为参考帧， 与 P ( t+i )和 P ( t+i+1 ) 没有关系， 因此核心参考帧 C ( t+j )可以正确解码。 对于 t+j之后的图像帧， 没有发生传输 差错， 图像都能够正确解码。 在这种情况下， 解码错误的时间间隔为 j-i。 The inter-frame coding method is to encode the previous frame of the current frame as a reference frame. In the fourth step, the inter-frame coding mode encoding of the core reference frame image is performed by using the core reference frame correctly received before the current frame as a reference frame. The invention also provides an error control method for video transmission, comprising: encoding an image of a first frame of a video sequence by using an intra coding method; Inserting a core reference frame at a plurality of positions after the first frame image in the video sequence; encoding the non-core reference frame image in the video sequence according to a standard encoding method; and encoding the core reference frame image in the video sequence by using an interframe coding method Receiving the encoded video sequence and then decoding and feeding back the decoded information; transmitting the encoded video sequence and the decoded decoding information. The step of inserting the core reference frame at a plurality of positions after the first frame image in the video sequence includes inserting the core reference frame at regular intervals in the video sequence, or selecting the core reference frame according to the rate distortion optimization model, or selecting the feedback reference channel The returned confirmation message selects the core reference frame. The standard encoding method is to adopt an interframe coding method. The inter-frame coding mode adopts a previous frame of the current frame as a reference frame. The core reference frame image in the video sequence is encoded in the inter-frame coding mode in which the core reference frame correctly received before the current frame is used as the reference frame. The present invention also provides an error control system for video transmission, comprising: an encoder for encoding an intra-frame coding mode of a first frame of a video sequence, and inserting a core at a plurality of positions after the first frame image in the video sequence a reference frame, encoding a non-core reference frame image in the video sequence according to a standard coding method, and performing inter-frame coding on the core reference frame image in the video sequence; and a decoder for receiving the encoded video sequence and then decoding, The decoding status is fed back to the encoder; the channel is used to transmit the encoder encoded video sequence and the information decoded by the feedback decoder. The channel includes: a transport channel for transmitting a coded encoded video sequence; and a feedback channel for feeding back information decoded by the decoder. There is a time delay for the encoder, decoder and channel. The invention proposes a method for realizing video transmission error control under the condition that the video coding efficiency is not significantly reduced by inserting a core reference frame. Compared with the original method of inserting an I frame or a key reference frame in a video sequence, it has higher coding efficiency, and can effectively reduce or prevent the spread of transmission errors on the video sequence, thereby improving the quality of video transmission. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 shows a conventional video coding method for processing transmission errors; FIG. 2 shows a method for processing a transmission error by a core reference frame method; FIG. 3 shows an error for inserting a core reference frame at a fixed interval. Control method; Figure 4 shows the error control method for adaptive core reference frame interval. DETAILED DESCRIPTION OF THE INVENTION In order to make the objects, technical solutions and advantages of the present invention more comprehensible, the present invention will be further described in detail below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In order to be able to reduce or prevent the spread of erroneously decoded pictures in the event of transmission errors, the present invention proposes a coding method based on a core reference frame, which can prevent the spread of errors without significantly reducing the efficiency of video coding. Figure 2 shows the core reference frame method for handling transmission errors. In the process of video encoding, the core reference frame is inserted at some specific position, denoted as C, as shown by the double line in FIG. The core reference frame is also encoded by the interframe coding method, which is different from the ordinary interframe coding method in that the reference frame is the previous I frame or the core reference frame. For example, in FIG. 2, C ( t+j ) is a reference frame of I ( t ), and C ( t+k ) is a reference frame of C ( t+j ). When a transmission error occurs, for example, an error occurs at time t+i, P(t+i) cannot be decoded correctly. Since the subsequent inter-coded frame P (t+i+1) is P (t+i) as the reference picture, the error-decoded picture will be spread backward. At time t+j, since the core reference frame C ( t+j ) is I ( t ) as the reference frame, it has no relationship with P ( t+i ) and P ( t+i+1 ), so the core reference frame C ( t+j ) can be decoded correctly. For image frames after t+j, no transmission errors occur and the images are correctly decoded. In this case, the time interval for decoding errors is ji.
.图 3 所示为根据本发明一示例性实施方式采取固定间隔插入核心参考帧实现传输差 错控制的方法。 视频编码器对视频序列进行实时编码，得到编码视频码流，如图 3中的上半部分所示。 然后将编码视频码流传送到视频解码器进行解码， 如图 3中的下半部分所示。 由于编码器或者信道传输过程中存在一定的时间延迟，视频解码器接收到编码枧频码
流的时刻与真实的视频序列存在一定的时间差。 同理， 由于视频解码器和或者反馈信道传 输过程存在一定的时间延迟， 视频解码器将当前图像解码信息反馈至视频编码器也存在一 定的时间延迟。 在图 3中假设刻编码器对视频序列进行编码后， 经过信道传输， 被解码器 接收并解码， 然后将解码状况反馈至编码器所需的时间在 4个单位时间以内。 即 t时刻编 码的图像帧被解码器接收后， 在 t+4时刻之前编码器可以得到解码器的反馈信息。 解码器端的核心参考帧的解码状态通过反馈信道传送给编码器， ACK表示该图像帧 能够正确解码， NACK表示该图像帧不能正确解码。如图 3中 I(t)、 C(t+4)、 C (t+12)、 C (t+16)、 C ( t+20 )等能够正确解码的图像帧通过 ACK反馈至编码器， 而不能正确解码 的 C (t+8)通过 NACK反馈至编码器。 在 t时刻， 编码器对视频序列的第一帧用帧内编码方式编码， 记为 I (t) 。 在 t+1至 t+3时刻，编码器对当前图像采取以前一帧作为参考帧的帧间编码方式编码， 记为 P (t+1 ) 、 P (t+2) 、 P (t+3) 。 在 t + 4时刻， 编码器已经接收到解码器反馈的 I (t)正确解码 ACK信息， 对当前图 像采取核心参考帧方式进行编码， 即以 I (t)作为参考帧进行编码， 记为 C (t + 4) 。 在 t+5至 t+7时刻，编码器对当前图像采取以前一帧作为参考桢的帧间编码方式编码、 记为 P (t+5) 、 P (t+6) 、 P (t+7) 。 在 t + 8时刻， 编码器已经接收到解码器反馈的 C ( t+4 )正确解码 ACK信息， 对当 前图像采取核心参考帧方式进行编码， 即以 C( t+4)作为参考帧进行编码，记为 C(t+8)。 在 t+9至 t+11时刻， 编码器对当前图像采取以前一帧作为参考帧的帧间编码方式编 码， 记为 P (t+9) 、 P (t+10) 、 P (t+11 ) 。 在 t + 12时刻， 编码器没有接收到解码器反馈的 C ( t+8 ) 正确解码 ACK信息， 但是 按照固定间隔核心参考帧方法需要对当前图像采取核心参考帧方式进行编码。 所以即以前 面正确接收的核心参考帧 C (t+4)作为参考帧进行编码， 记为 C (t+12) 。 后面时刻的编码方法与前面类似， 在此不再重复。 图 4所示为根据本发明另一示例性实施方式的自适应插入核心参考帧实现传输差错 控制的方法。 即在丢包率较低时采取普通帧间编码方法， 在丢包率较高时采取核心参考帧 方法。 视频编码器对视频序列进行实时编码，得到编码视频码流，如图 4中的上半部分所示。
然后将编码视频码流传送到视频解码器进行解码， 如图 4中的下半部分所示。 由于编码器或者信道传输过程中存在一定的时间延迟，视频解码器接收到编码视频码 流的时刻与真实的视频序列存在一定的时间差。 同理， 由于视频解码器和或者反馈信道传 输过程存在一定的时间延迟， 视频解码器将当前图像解码信息反馈至视频编码器也存在一 定的时间延迟。 在图 4中假设刻编码器对视频序列进行编码后， 经过信道传输， 被解码器 接收并解码， 然后将解码状况反馈至编码器所需的时间在 2个单位时间以内。 即 t时刻编 码的图像帧被解码器接收后， 在 t+2时刻之前编码器可以得到解码器的反馈信息。 FIG. 3 illustrates a method of implementing transmission error control by inserting a core reference frame at fixed intervals in accordance with an exemplary embodiment of the present invention. The video encoder performs real-time encoding of the video sequence to obtain an encoded video stream, as shown in the upper part of FIG. The encoded video stream is then passed to a video decoder for decoding, as shown in the lower half of Figure 3. The video decoder receives the encoded chirp code due to a certain time delay in the encoder or channel transmission process. There is a certain time difference between the moment of the stream and the real video sequence. Similarly, due to the certain time delay of the video decoder and or the feedback channel transmission process, the video decoder also has a certain time delay to feed back the current image decoding information to the video encoder. In Fig. 3, it is assumed that the engraving encoder encodes the video sequence, after channel transmission, is received and decoded by the decoder, and then the time required to feed back the decoding status to the encoder is within 4 unit time. That is, after the image frame encoded at time t is received by the decoder, the encoder can obtain the feedback information of the decoder before the time t+4. The decoding status of the core reference frame at the decoder end is transmitted to the encoder through the feedback channel, ACK indicates that the image frame can be correctly decoded, and NACK indicates that the image frame cannot be correctly decoded. As shown in Figure 3, the image frames that can be correctly decoded, such as I(t), C(t+4), C(t+12), C(t+16), C(t+20), are fed back to the encoder through ACK. C (t+8), which cannot be correctly decoded, is fed back to the encoder through NACK. At time t, the encoder encodes the first frame of the video sequence by intra coding, denoted as I(t). At t+1 to t+3, the encoder encodes the current picture as the interframe coding mode of the previous frame as P (t+1 ), P (t+2), P (t+3). ). At time t + 4, the encoder has received the I (t) correctly decoded ACK information fed back by the decoder, and encodes the current image in a core reference frame manner, that is, I (t) is used as a reference frame for encoding, denoted as C (t + 4). At t+5 to t+7, the encoder encodes the current image using the previous frame as the reference frame, denoted as P (t+5), P (t+6), P (t+7). ). At time t + 8, the encoder has received the C ( t+4 ) feedback from the decoder to correctly decode the ACK information, and encodes the current image in the core reference frame mode, that is, encodes with C ( t+4) as the reference frame. , recorded as C (t + 8). At t+9 to t+11, the encoder encodes the current frame as the interframe coding mode of the previous frame as P (t+9), P (t+10), P (t+11). ). At time t + 12, the encoder does not receive the C ( t+8 ) feedback from the decoder to correctly decode the ACK information, but according to the fixed interval core reference frame method, the current image is encoded in the core reference frame mode. Therefore, the core reference frame C (t+4) received correctly before is encoded as a reference frame, and is recorded as C (t+12). The coding method at the latter moment is similar to the previous one and will not be repeated here. 4 illustrates a method of adaptively inserting a core reference frame to implement transmission error control, in accordance with another exemplary embodiment of the present invention. That is, the normal interframe coding method is adopted when the packet loss rate is low, and the core reference frame method is adopted when the packet loss rate is high. The video encoder performs real-time encoding of the video sequence to obtain an encoded video stream, as shown in the upper half of FIG. The encoded video stream is then passed to a video decoder for decoding, as shown in the lower half of Figure 4. Due to the certain time delay in the encoder or channel transmission process, the video decoder has a certain time difference from the real video sequence when it receives the encoded video stream. Similarly, due to the certain time delay of the video decoder and or the feedback channel transmission process, the video decoder also has a certain time delay to feed back the current image decoding information to the video encoder. In Fig. 4, it is assumed that the coded encoder encodes the video sequence, after channel transmission, is received and decoded by the decoder, and then the time required to feed back the decoding condition to the encoder is within 2 unit time. That is, after the image frame encoded at time t is received by the decoder, the encoder can obtain the feedback information of the decoder before the time t+2.
'解码器端的核心参考帧的解码状态通过反馈信道传送给编码器， ACK表示该图像帧 能够正确解码， NACK表示该图像帧不能正确解码。 如图 3中 I(t)、 C (t+5)、 C (t+10)、 C (t+15 )等能够正确解码的图像帧通过 ACK反馈至编码器， 而不能正确解码的 P (t+6)、 C ( t+8 )通过 NACK反馈至编码器。 在 t时刻， 编码器对视频序列的第一帧用帧内编码方式编码， 记为 I (t) 。 在 t+1至 t+4时刻，编码器没有收到任何 NACK信息，对当前图像采取以前一帧作为 参考帧的帧间编码方式编码， 记为 P (t+1) 、 P (t+2) 、 P (t+3) 、 P (t+4) 。 在 t+5时刻， 编码器没有接收到任何 NACK信息， 即表明丢包率很低， 对当前帧本 应该采取以前一帧作为参考帧的帧间编码方式编码。 但是由于很长时间内没有指定核心参 考帧， 当前帧和正确核心参考帧的间隔逐渐变远， 一旦发生差错， 需要参考正确核心参考 帧编码， 编码效率将显著降低。 为了提高编码效率， 如图 4所示， 当每连续编码 5个普通 帧后， 不论当前帧参考那一帧编码， 当前帧都指定为核心参考帧。 在 t+6至 t+7时刻， 编码器没有收到任何 NACK信息，对当前图像采取以前一帧作为 参考帧的帧间编码方式编码， 记为 P (t+6) 、 P (t+7) 。 在 t+8时刻， 编码器接收到 NACK信息， P (t+6)不能正确解码。 对当前帧采取核心 参考帧方法进行编码， 即以最近接收的正确核心参考帧 C (t+5)作为参考图像进行帧间编 码， 记为 C (t+8) 。 在 t+9时刻， 编码器没有收到任何 NACK信息， 对当前图像采取以前一帧作为参考 帧的帧间编码方式编码， 记为 P (t+9) 。 在 t+10时刻， 编码器接收到解码器 NACK信息， C (t+8) 不能正确解码， 对当前 图像采取核心参考帧方式进行编码， 即以前一个正确解码的核心参考帧 C (t+5)作为参考 桢进行编码， 记为 C (t+10) 。
在 t+11至 t+14时刻， 编码器没有收到任何 NACK信息，对当前图像采取以前一祯作 为参考桢的桢间编码方式编码， 记为 P (t+11 ) 、 P (t+12) 、 P (t+13) 、 P (t+14) 。 在 t+15时刻， 编码器没有接收到任何 NACK信息， 即表明丢包率很低， 对当前帧本 应该采取以前一帧作为参考帧的帧间编码方式编码。 但是为了提高编码效率， 如图 4所示， 当每连续编码 5个普通帧后， 不论当前桢参考那一帧编码， 当前帧都指定为核心参考帧， 即以 C (t+10)作为参考帧进行帧间编码， 记为 C (t+15) 。 后续时刻的编码方法与前面类似， 在此不再重复。 当然， 本发明还可有其他多种实施例， 在不背离本发明精神及其实质的情况下， 熟悉 本领域的技术人员当可根据本发明作出各种相应的改变和变形， 但这些相应的改变和变形 都应属于本发明所附的权利要求的保护范围。
The decoding state of the core reference frame at the decoder end is transmitted to the encoder through the feedback channel, ACK indicates that the image frame can be correctly decoded, and NACK indicates that the image frame cannot be correctly decoded. As shown in Figure 3, I (t), C (t+5), C (t+10), C (t+15) and other image frames that can be correctly decoded are fed back to the encoder through ACK, but not correctly decoded ( t+6), C (t+8) are fed back to the encoder via NACK. At time t, the encoder encodes the first frame of the video sequence by intra coding, denoted as I(t). At t+1 to t+4, the encoder does not receive any NACK information, and encodes the current frame as the interframe coding mode of the previous frame as P (t+1) and P (t+2). ), P (t+3), P (t+4). At t+5, the encoder does not receive any NACK information, which means that the packet loss rate is very low. The current frame should be encoded by the previous frame as the reference frame. However, since the core reference frame is not specified for a long time, the interval between the current frame and the correct core reference frame is gradually changed. Once an error occurs, the correct core reference frame coding needs to be referred to, and the coding efficiency is significantly reduced. In order to improve the coding efficiency, as shown in FIG. 4, after every 5 consecutive normal frames are encoded, the current frame is designated as the core reference frame regardless of the frame encoding of the current frame. At t+6 to t+7, the encoder does not receive any NACK information, and encodes the current frame as the interframe coding mode of the previous frame as P (t+6) and P (t+7). ). At time t+8, the encoder receives the NACK information and P (t+6) cannot be decoded correctly. The current frame is encoded by the core reference frame method, that is, the recently received correct core reference frame C (t+5) is used as a reference image for inter-frame coding, and is denoted as C (t+8). At time t+9, the encoder does not receive any NACK information, and encodes the current frame as the interframe coding mode of the previous frame as P (t+9). At t+10, the encoder receives the decoder NACK information, C (t+8) cannot be correctly decoded, and encodes the current image in the core reference frame mode, that is, the previous correctly decoded core reference frame C (t+5). ) is coded as a reference ,, denoted as C (t+10). At t+11 to t+14, the encoder does not receive any NACK information, and takes the previous image as the reference code for the current image, which is denoted as P (t+11), P (t+12). ), P (t+13), P (t+14). At t+15, the encoder does not receive any NACK information, which means that the packet loss rate is very low. The current frame should be encoded with the previous frame as the reference frame. However, in order to improve the coding efficiency, as shown in FIG. 4, after every 5 consecutive normal frames are encoded, the current frame is designated as the core reference frame regardless of the current frame coding, that is, C (t+10) is used as a reference. The frame is inter-coded and recorded as C (t+15). The encoding method at the subsequent moment is similar to the previous one and will not be repeated here. Of course, the present invention may be embodied in various other various modifications and changes without departing from the spirit and scope of the invention. Changes and modifications are intended to be included within the scope of the appended claims.