CN109688407A

CN109688407A - Reference block selection method, device, electronic equipment and the storage medium of coding unit

Info

Publication number: CN109688407A
Application number: CN201710978417.9A
Authority: CN
Inventors: 张贤国; 朱政; 金星; 张二丽; 范娟婷
Original assignee: Beijing Kingsoft Cloud Network Technology Co Ltd; Beijing Kingsoft Cloud Technology Co Ltd
Current assignee: Beijing Kingsoft Cloud Network Technology Co Ltd; Beijing Kingsoft Cloud Technology Co Ltd
Priority date: 2017-10-18
Filing date: 2017-10-18
Publication date: 2019-04-26
Anticipated expiration: 2037-10-18
Also published as: CN109688407B; WO2019076344A1

Abstract

The embodiment of the invention provides a kind of selection method of the reference block of coding unit, device, electronic equipment and storage mediums, method includes: to obtain the object reference frame set of target code unit said target video frame, wherein, the quantity of the included reference frame of object reference frame set and the reference frame is arranged according to the frame type and frame-layer grade of target video frame；From the reference frame that object reference frame set includes, the reference block of selection target coding unit.In technical solution provided by the embodiment of the present invention, compared with the prior art, adaptively the quantity of reference frame and reference frame in object reference frame set can be limited according to frame type and frame-layer grade, with reduce target code unit reference block selection calculation amount, so as to improve coding unit reference block efficiency of selection.

Description

Reference block selection method, device, electronic equipment and the storage medium of coding unit

Technical field

The present invention relates to technical field of video coding, reference block selection method, dress more particularly to a kind of coding unit It sets, electronic equipment and storage medium.

Background technique

With the continuous development of multimedia digital Video Applications, the increase of original video information source data volume is so that existing biography Defeated network bandwidth and storage resource are unbearable, currently, often by between Video coding (also referred to as video compress) elimination vision signal Existing redundancy alleviates the pressure of transmission network bandwidth and storage resource.So far, national and foreign standards tissue is successive A variety of different video encoding standards are formulated, currently, the video encoding standard of mainstream is all made of " block-based prediction and change Change " hybrid encoding frame, video frame is predicted especially by encoder, converts, quantify, scanning, a system such as entropy coding Column processing technology obtains output bit flow, in case transmission or storage.

Predicting Technique involved in Video coding is generally divided into two kinds, i.e. intra prediction and inter-prediction.Wherein, interframe is pre- It surveys for reducing temporal redundancy, carries out predictive coding using the correlation between adjacent video frames, each video frame can be used Encoded and reconstruct video frame is as reference frame.For target code unit to be encoded, firstly, giving target code list in advance Target video frame belonging to member configures available reference frame, and the reference frame configured is to encode and weighed before target video frame The video frame of structure.Then, the reference block that selection prediction uses in the reference frame configured in advance to target video frame.Due to adjacent There are regular hour correlations between video frame, and therefore, inter-prediction can search more matched reference block, so that residual error Value is smaller, can not only use less coded-bit encoded video frame in this way, moreover it is possible to guarantee the precision of Video coding.

Can be more however, completing the possible quantity of the video frame for encoding and being reconstructed before target video frame, based on this choosing Calculation amount is larger when selecting the reference block of target code unit, causes the reference block efficiency of selection for target code unit low.

Summary of the invention

A kind of selection method of the reference block for being designed to provide coding unit of the embodiment of the present invention, device, electronics are set Standby and computer readable storage medium, to improve the efficiency of selection of the reference block of coding unit.Specific technical solution is as follows:

In a first aspect, the embodiment of the invention provides a kind of reference block selection methods of coding unit, which comprises

Obtain the object reference frame set of target code unit said target video frame, wherein the object reference frame collection The quantity for closing included reference frame and the reference frame is arranged according to the frame type and frame-layer grade of the target video frame；

From the reference frame that the object reference frame set includes, the reference block of the target code unit is selected.

Optionally, from the reference frame that the object reference frame set includes, the reference of the target code unit is selected The step of block, comprising:

According to the correlation between coding unit, target candidate set is determined from the object reference frame set；

From the reference frame that the target candidate set includes, the reference block of the target code unit is selected.

Optionally, the correlation according between coding unit determines target candidate from the object reference frame set The step of set, comprising:

Following any one or any number of union of sets collection are determined as target candidate set:

The target code unit, father's coding unit that motion estimation search is completed candidate collection；

The target code unit, sub- coding unit that motion estimation search is completed candidate collection；

The target code unit, adjacent encoder unit that motion estimation search is completed candidate collection.

Optionally, from the reference frame that the target candidate set includes, the reference of the target code unit is selected Before the step of block, the method also includes:

Judge whether the target code unit meets the first preset condition, if it is satisfied, by the target candidate set In reference frame be updated to the reference frame in the object reference frame set；

Wherein, first preset condition include it is following at least one:

The target code unit is located at the boundary of the target video frame；

In the encoded adjacent encoder unit of the target code unit, the sum of existing intraprediction unit is greater than First preset threshold；

Encoded father's coding unit of the target code unit, sub- coding unit, in adjacent encoder unit, using frame Between predict coding unit sum less than the second preset threshold.

Optionally, from the reference frame that the target candidate set includes, the reference block of the target code unit is selected The step of, comprising:

Based on the target code unit, whole pixel is carried out on the reference frame in the target candidate set one by one and is searched Rope determines the reference frame of the target code unit according to search result from the reference frame that the target candidate set includes；

Based on the target code unit, a point pixel search is carried out on the reference frame of the target code unit, is obtained The optimal reference block of acquisition is determined as the reference block of the target code unit by optimal reference block.

Optionally, described according to search result, the target is determined from the reference frame that the target candidate set includes The step of reference frame of coding unit, comprising:

At least one of according to search result, in the following way, the reference frame for including from the target candidate set The reference frame of the middle determination target code unit:

The reference frame for the target minimum distortion threshold value that the distortion of whole pixel search is less than the preset target video frame is true It is set to the reference frame of the target code unit；

In the reference frame for including by the target candidate set, whole pixel distortion is smaller than the whole pixel distortion of other reference frames , and small amplitude is determined as the mesh less than the reference frame of the whole pixel distortion threshold value of target of the preset target video frame Mark the reference frame of coding unit；

The whole the smallest reference frame of pixel distortion in object reference frame that the target candidate set includes is determined as described The reference frame of target code unit.

Optionally, the target minimum distortion threshold value and the whole pixel distortion threshold value of target are according to the target video frame Frame type and frame-layer grade it is preset.

Judge whether the target code unit meets the second preset condition, second preset condition includes: based on institute Target code unit is stated, the minimum predicted distortion searched on each reference frame in the target candidate set is more than preset institute State the target maximum distortion threshold value of target video frame；

If it is satisfied, the reference frame in the target candidate set is updated to the reference in the object reference frame set Frame.

Optionally, in the case where the target code unit meets second preset condition, from the target candidate In the reference frame that set includes, the step of selecting the reference block of the target code unit, comprising:

Based on the target code unit, one by one on the reference frame in the difference set before and after target candidate set update It scans for, according to the search result in target candidate set in the updated on each reference frame, from the target candidate collection In the reference frame that conjunction includes, the reference block of the target code unit is selected；Or according on the reference frame in the difference set Search result selects the reference block of the target code unit from the reference frame that the difference set includes；

In the case where the target code unit is unsatisfactory for second preset condition, from the target candidate set packet In the reference frame contained, the step of selecting the reference block of the target code unit, comprising:

According to being based on the target code unit, search result in the target candidate set on each reference frame, Select the reference block of the target code unit.

Optionally, the target maximum distortion threshold value is default according to the frame type and frame-layer grade of the target video frame 's.

Second aspect, the embodiment of the invention provides a kind of reference block selection device of coding unit, described device includes:

Module is obtained, for obtaining the object reference frame set of target code unit said target video frame, wherein described The quantity of the included reference frame of object reference frame set and the reference frame be according to the frame type of the target video frame and The setting of frame-layer grade；

Selecting module, for selecting the target code unit from the reference frame that the object reference frame set includes Reference block.

Optionally, the selecting module includes:

Submodule is determined, for determining target from the object reference frame set according to the correlation between coding unit Candidate collection；

Submodule is selected, for selecting the target code unit from the reference frame that the target candidate set includes Reference block.

Optionally, the determining submodule is specifically used for:

Optionally, the selecting module further include:

First processing submodule, for executing the reference frame for including from the target candidate set in the selection submodule In, before the step of selecting the reference block of the target code unit, it is pre- to judge whether the target code unit meets first If condition, if it is satisfied, the reference frame in the target candidate set is updated to the reference in the object reference frame set Frame；

Wherein, first preset condition include it is following at least one:

The target code unit is located at the boundary of the target video frame；

In the encoded adjacent encoder unit of the target code unit, the sum of existing intraprediction unit is less than First preset threshold；

Optionally, the selection submodule includes:

Search unit, for being based on the target code unit, one by one on the reference frame in the target candidate set Carry out whole pixel search；

First determination unit, for determining institute from the reference frame that the target candidate set includes according to search result State the reference frame of target code unit；

Second determination unit, it is enterprising in the reference frame of the target code unit for being based on the target code unit Row divides pixel search, obtains optimal reference block, the optimal reference block of acquisition is determined as to the reference block of the target code unit.

Optionally, first determination unit is specifically used for:

Optionally, the selecting module further include:

Second processing submodule, for executing the reference frame for including from the target candidate set in the selection submodule In, before the step of selecting the reference block of the target code unit, it is pre- to judge whether the target code unit meets second If condition, if it is satisfied, the reference frame in the target candidate set is updated to the reference in the object reference frame set Frame；

Second preset condition includes: each ginseng in the target candidate set based on the target code unit Examine the target maximum distortion threshold value that the minimum predicted distortion searched on frame is more than the preset target video frame.

Optionally, the selection submodule includes:

Third determination unit is to wait one by one in target when meeting for the judging result in the second processing submodule Selected works close to be scanned on the reference frame in the difference set for updating front and back, according to each reference in target candidate set in the updated The search result of frame, or according to the search result in the difference set on reference frame, select the reference of the target code unit Block；

4th determination unit is when being unsatisfactory for, according to based on institute for the judging result in the second processing submodule Target code unit is stated, the search result in the target candidate set on each reference frame selects the target code list The reference block of member.

The third aspect, the embodiment of the invention provides a kind of electronic equipment, including processor and memory,

The memory, for storing computer program；

The processor when for executing the program stored on memory, realizes the reference of any of the above-described coding unit Method and step described in the selection method of block.

Fourth aspect, the embodiment of the invention provides a kind of computer readable storage medium, the computer-readable storage Dielectric memory contains computer program, computer program reference block of any of the above-described coding unit when being executed by processor Method and step described in selection method.

5th aspect, the embodiment of the invention provides a kind of computer program products comprising instruction, when it is in computer When upper operation, so that computer executes method and step described in the selection method of the reference block of any of the above-described coding unit.

6th aspect, the embodiment of the invention provides a kind of computer programs, when run on a computer, make to succeed in one's scheme Calculation machine executes method and step described in the selection method of the reference block of any of the above-described coding unit.

Selection method, device, electronic equipment and the storage of a kind of reference block of coding unit provided in an embodiment of the present invention In medium, method includes: to obtain the object reference frame set of target code unit said target video frame, wherein object reference The quantity of the included reference frame of frame set and the reference frame is arranged according to the frame type and frame-layer grade of target video frame；From In the reference frame that object reference frame set includes, the reference block of selection target coding unit.Skill provided by the embodiment of the present invention It, compared with the prior art, can be adaptively according to frame type and frame-layer grade to the ginseng in object reference frame set in art scheme The quantity for examining frame and reference frame is limited, to reduce the calculation amount of the reference block selection of target code unit, so as to mention The efficiency of selection of the reference block of high coding unit.

Detailed description of the invention

In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with It obtains other drawings based on these drawings.

Fig. 1 is the referring-to relation schematic diagram of an image group unit in the prior art；

Fig. 2 is a kind of flow diagram of the reference block selection method of coding unit provided in an embodiment of the present invention；

Fig. 3 is the structural schematic diagram for the coding unit that an embodiment of the present invention provides；

Selection target coding in the reference frame that Fig. 4 includes for the slave object reference frame set that an embodiment of the present invention provides The flow diagram of the reference block of unit；

Selection target coding in the reference frame that Fig. 5 includes for the slave target candidate set that another embodiment of the present invention provides The flow diagram of the reference block of unit；

Fig. 6 is a kind of structural schematic diagram of the reference block selection device of coding unit provided in an embodiment of the present invention；

Fig. 7 is the structural schematic diagram for the selecting module that an embodiment of the present invention provides；

Fig. 8 is the structural schematic diagram for the selection submodule that an embodiment of the present invention provides；

Fig. 9 is the structural schematic diagram of a kind of electronic equipment provided in an embodiment of the present invention.

Specific embodiment

Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other Embodiment shall fall within the protection scope of the present invention.

In the prior art, the possible quantity of the video frame for encoding and being reconstructed is completed before target video frame can be more, base Calculation amount is larger when the reference block of this selection target coding unit, leads to the reference block efficiency of selection for target code unit It is low.In order to improve the reference block efficiency of selection of target code unit, the embodiment of the present invention proposes a kind of coding unit Reference block selection method, device, electronic equipment and computer readable storage medium.

In reference block selection method, device, electronic equipment and the meter of a kind of coding unit proposed to the embodiment of the present invention Before calculation machine readable storage medium storing program for executing describes in detail, first to present document relates to portion of techniques terms simply to be introduced.

Video frame type: video frame type includes three kinds: I frame, P frame and B frame；

Wherein, I frame: the coding of all coding units all uses intra prediction in I frame, when encoding and decoding just using this frame data Complete image can be reconstructed；P frame: it is encoded with reference to the encoded and reconstructed video frame of front, front refers to According to display order, before current P frame；B frame: with reference to video frame that is front and subsequent encoded and having reconstructed into Row coding, referring to according to display order for front is subsequent to refer to according to display order before current B frame, in current B After frame.

Image group unit: refer to one group of video frame (including the latter key continuous and positioned at adjacent key frame between Frame does not include previous key frame), wherein key frame refers to I frame or P frame.It is last in this group of image according to display order One frame is without reference to other images in organizing.It should distinguish, image group unit is different from image group (GOP, Group of Pictures), an image group may include multiple images group unit.

As shown in Figure 1, multiple video frames shown in fig. 1 form an image group unit, the corresponding frame of this multiple video frame It number is respectively 8K-7,8K-6,8K-5,8K-4,8K-3,8K-2,8K-1 and 8K according to display order, it is convenient for statement, hereafter will Use P_kIndicate that frame number is the video frame of k, for example, P_8K-7Indicate that frame number is the video frame of 8K-7.

Frame-layer grade: i.e. video frame level refers to the mutual referring-to relation of image group unit of the corresponding video frame belonging to it On level.In same image group unit, the higher video frame of frame-layer grade, the priority of coding is higher.Therefore, frame-layer grade is lower Video frame can be encoded with the higher video frame of reference frame level.

As shown in Figure 1, multiple video frames shown in Fig. 1 form an image group unit, P_8K-7、P_8K-5、P_8K-3、P_8K-1 Positioned at the 4th level of the image group unit, P_8K-6And P_8K-2Positioned at the third level of the image group unit, P_8K-4Positioned at the image Second level of group unit, P_8KPositioned at the first level of the image group unit, from the first level to the 4th level, level successively becomes It is low.

The coded sequence of video frame be according to the first level, the second level, the sequence of third level to the 4th level successively It carries out, therefore, the video frame of the first level can only be encoded with reference to the video frame of the image group unit of other priority encodings Or it is encoded using intra prediction；The video frame of second level can be with reference to the video frame of the first level and other are preferential The video frame etc. of the image group unit of coding is encoded；The video frame of third level can refer to the first level, the second level Video frame and the video frames etc. of image group unit of other priority encodings encoded；The video frame of 4th level can be with With reference to the first level, the second level, the video frame of third level and video frame of the image of other priority encodings group unit etc. It is encoded.

It describes in detail below by specific embodiment to the present invention.

Referring to Fig. 2, Fig. 2 is a kind of process signal of reference block selection method of coding unit provided in an embodiment of the present invention Figure, this method comprises:

S101, the object reference frame set for obtaining target code unit said target video frame, wherein object reference frame collection The quantity for closing included reference frame and the reference frame is arranged according to the frame type and frame-layer grade of target video frame；

Target video frame refers to current video frame to be encoded, and target code unit refers to according to coded sequence, target view Current coding unit to be encoded in frequency frame, object reference frame set are for the set of the reference frame of target video frame setting.It can With understanding, if target video frame is I frame, object reference frame set should be empty set, directly progress intraframe predictive coding. If target video frame is B frame or P frame, object reference frame collection is combined into nonempty set, at this point, for appointing in target video frame One coding unit, the final used reference block of coding can be to be obtained from the video frame that object reference frame set is included 's.

In this step, according to the frame type of target video frame and frame-layer grade, the reference frame quantity of target video frame is set And reference frame can be specifically configured in the following way with obtaining object reference frame set:

Firstly, obtaining the frame type and frame-layer grade of target video frame；

Then, reference frame setting is carried out according to the frame type of target video frame and frame-layer grade, obtains the mesh of target video frame Reference frame set is marked, is specifically configured according to following two principle:

If the principle based on video frame type include: target video frame be P frame, from according to display order in target video It is before frame and before the target video frame coding in the encoded and video frame that has reconstructed, for the selection setting of target video frame Reference frame；If target video frame be B frame, from according to display order before target video frame and target video frame compile The encoded and video frame that has reconstructed before code, and according to display order after target video frame and in target video Frame encodes in video frame that is encoded in the past and having reconstructed, for target video frame selection setting reference frame.

Principle based on video frame frame-layer grade includes: in the image group unit belonging to target video frame, compared to than mesh The high video frame of the frame-layer grade of reference frame set is marked, the quantity of reference frame included by object reference frame set is not more than the ginseng of the latter Examine the quantity of reference frame included by frame set.It may be implemented in same image group unit based on this principle, frame-layer grade is higher Video frame use more reference frame, the lower video frame of frame-layer grade use less reference frame.Based on realize this purpose, In other embodiments, this principle can also be other concrete forms, be not specifically limited herein.

It will be appreciated by persons skilled in the art that frame-layer grade is higher prior to frame-layer grade in an image group unit Lower video frame is encoded, that is to say, that the higher video frame of frame-layer grade can be referred to by the lower video frame of frame-layer grade. For the higher video frame of frame-layer grade, more reference frame is set, it is ensured that the encoding precision of the higher video frame of frame-layer grade compared with Height, and then the lower video frame of the frame-layer grade encoding precision with higher with reference to it can also be made.Meanwhile it is lower for frame-layer grade Video frame less reference frame is set, can also reduce selection coding unit reference block when calculation amount, to improve volume The efficiency of selection of the reference block of code unit.

Further, it is also contemplated that greater number of reference frame is set for P frame, less reference is set for B frame Frame.

It will be appreciated by persons skilled in the art that comparatively, P frame is carried out prior to B frame in same image group unit Coding, P frame can be used as the reference frame of B frame, and greater number of reference frame is arranged if it is P frame, it is ensured that the coding essence of P frame Degree is higher, to can also make the B frame encoding precision with higher with reference to it.Meanwhile less reference frame is set for B frame, The calculation amount when reference block of selection coding unit can also be reduced, to improve the efficiency of selection of the reference block of coding unit. For example, being directed to P frame, using three frame forward reference frames, a reference frame is used to each for before and after B frame.

S102, from the reference frame that object reference frame set includes, the reference block of selection target coding unit.

From the reference frame that object reference frame set includes, target code unit is selected especially by the mode of search Reference block, for being used when target code cell encoding.This step can be realized by two ways:

The first: being based on target code unit, directly scan on the reference frame that object reference frame set is included, To obtain the reference block met the requirements as the reference block of target code unit.For example, directly against object reference frame set institute Include is searched for one by one all referring to frame, and the reference block of actual demand is met until obtaining, or all search is completed to obtain Optimal reference block, using reference block obtained as the reference block of target code unit.Alternatively, directly against object reference frame collection Close included scanned for all referring to frame, search after the completion of obtain optimal reference block, using the optimal reference block of acquisition as The reference block of target code unit.

Second: according to actual needs, further reducing the range of target video frame reference frame, that is, obtain a smaller model The target candidate set enclosed, this target candidate set are specifically the subset of object reference frame set.Then, according to the first side Formula is based on target code unit, the reference block of selection target coding unit from target candidate set.

Using the second way, while meeting code requirement, the reference of target code unit can be further decreased The calculation amount of block selection, so as to improve the reference block efficiency of selection of coding unit.

It, compared with the prior art, can be adaptively according to frame type and frame-layer grade to target using embodiment illustrated in fig. 2 The quantity of reference frame and reference frame in reference frame set is limited, and the selection of target code elements reference block can be also reduced Calculation amount, so as to improve coding unit reference block efficiency of selection.

In a kind of embodiment of the embodiment of the present invention, referring to Fig. 4, from the reference frame that object reference frame set includes, The step of reference block of selection target coding unit, comprising:

S201, according to the correlation between coding unit, target candidate set is determined from object reference frame set；

Wherein, target candidate set is exactly the set for the reference frame that may be referred to when encoding to target code unit, The difference set of object reference frame set and target candidate set refers to from object reference frame set, exclusion, encoding target coding The set for the reference frame that will not be referred to when unit.

Specifically, in same video frame there is certain spatial coherence, the ginseng of different coding unit between coding unit Examining between frame also just has certain similitude, for example, the candidate collection of coding unit a is A in same video frame, for The relevant coding unit b of coding unit a, the candidate collection of coding unit b are similar to set candidates A, that is to say, that can select Candidate collection of the candidate collection A as coding unit b.

In addition, the spatial coherence between coding unit is higher, the similitude between their reference frame is higher, and And in same video frame, compared between other coding units, between adjacent encoder unit, sub- coding unit and father encode it is single Spatial coherence between member is higher.For example, the candidate collection of coding unit a ' is A ', coding unit b ' in same video frame Candidate collection be B ', coding unit b " is adjacent with coding unit b ', both non-conterminous with coding unit a ', nor father and son or son Father's relationship, then for candidate collection A ' compared with candidate collection B ', the candidate collection and candidate collection B ' of coding unit b " is more like, Candidate collection B ' can be selected as the candidate collection of coding unit b ".

S202, from the reference frame that target candidate set includes, the reference block of selection target coding unit.

Relative to object reference frame set, the range of target candidate set is reduced, and more targetedly, therefore, Again from target candidate set, the reference block of selection target coding unit, so that it may the calculation amount of reference block selection is substantially reduced, It is thus possible to improve the efficiency of reference block selection.

Further, in a kind of embodiment of the embodiment of the present invention, according to the correlation between coding unit, join from target Examine the step of target candidate set is determined in frame set, comprising:

Any one following set or any two following above union of sets collection are determined as target candidate set:

Target code unit, father's coding unit that motion estimation search is completed candidate collection (hereinafter referred to as first Set)；

Target code unit, sub- coding unit that motion estimation search is completed candidate collection (hereinafter referred to as second Set)；

Target code unit, adjacent encoder unit that motion estimation search is completed candidate collection (hereinafter referred to as Three set).

Wherein, above-mentioned first set, second set and third set both can be separately as target candidate set, can also be with Intercombination forms target candidate set.Each set is preparatory corresponding record with corresponding coding unit, for example, for every A candidate collection records the title of the candidate collection, the frame number of its reference frame for being included and corresponding coding unit, leads to The each set of overwriting can form index, find corresponding candidate collection according to index so as to subsequent.

In above-mentioned first set, second set and third set, when specifically used, compiled for the whole of video Same setting can be used in code unit, for example, using target code unit, adjacent volume that motion estimation search is completed The candidate collection of code unit is as target candidate collection.It is of course also possible to be set according to the specific code requirement of different video frame It sets, which union of sets collection realization is adaptively adjusted using as target candidate set, for example, P frame and B frame are comparatively, Guarantee that the encoding precision of P frame is more important, if target video frame is P frame, using first set, second set and third set Union is as target candidate set, if target video frame is B frame, using third set as target candidate set.

As shown in figure 3, the coding unit that video encoding standard uses size can be changed, general range is from minimum coding unit 8 × 8 arrive maximum coding unit 64 × 64, therefore, because the difference of the coding unit zoned format of video frame, in any two videos In frame, the coding unit of same location, the quantity of respective adjacent encoder unit, size, relative position are likely to difference.Its In, 64 × 64 coding unit and 32 × 32 both coding units in, the former can be father's coding unit of the latter, Hou Zheke Think the former sub- coding unit；32 × 32 coding unit and 16 × 16 both coding units in, the former can be the latter Father's coding unit, the latter can be the former sub- coding unit；Both 16 × 16 coding unit and 8 × 8 coding unit In, the former can be father's coding unit of the latter, and the latter can be the former sub- coding unit.

According to the difference of coding unit zoned format, target code unit, adjacent volume that motion estimation search is completed Code unit may include:

The left adjacent encoder unit of target code unit, upper adjacent encoder unit, upper left adjacent encoder unit, upper right are adjacent Coding unit and other the adjacent encoder unit of motion estimation search may be carried out before target code unit.

It should be noted that target code unit, adjacent encoder unit that motion estimation search is completed Candidate Set It closes, can only include the candidate collection of any one of aforementioned various adjacent encoder units, also may include aforementioned various adjacent The union of any a variety of candidate collection in coding unit.

Illustratively, as shown in figure 3, target code unit is coding unit composed by c, d, i and j, target code list The adjacent encoder unit first, motion estimation search is completed includes: coding unit f and e.The father of target code unit encodes single Member be c, d, i, j, f, e and g composed by coding unit, the sub- coding unit of target code unit include: coding unit c, d, i,j.It is understood that determining target candidate set from object reference frame set according to the correlation between coding unit The step of before, father's coding unit of target code unit may complete motion estimation search, it is also possible to not complete to transport Dynamic estimating searching.Similarly, before this step, the sub- coding unit of target code unit may complete estimation and search Rope, it is also possible to not complete motion estimation search.

By the setting of this step, identified target candidate set has more specific aim, not only can improve coding The efficiency of selection of the reference block of unit, it can also be ensured that certain encoding precision.

It should be noted that determining that target is waited from object reference frame set by according to the correlation between coding unit Selected works close step, and identified target candidate set is also possible to as empty set, and allowing in this way can not be by target candidate set The reference block for choosing target code unit, for reference to realize coding.In order to go on smoothly coding, can specify that From target candidate set, in the case where the reference block for selecting fall short coding unit, target code unit is switched in frame Predictive coding.It is of course also possible to be modified by other means to target candidate set, to guarantee smoothly to wait from target Selected works choose the reference block of target code unit in closing.

In addition, when identified target candidate set is not empty set, it is also possible to comprising less reference frame, in this way In order to guarantee encoding precision, it can specify that target code unit is encoded using intra prediction mode；Or pass through other Mode is modified target candidate set.

In a kind of embodiment of the embodiment of the present invention, target candidate set can be modified in the following manner, On the one hand can guarantee subsequent can choose the reference block of target code unit, separately from target candidate set adjusted On the one hand it is also possible that improving encoding precision while accelerating the selection course of reference block of target code unit.Specifically Modification mode include:

From the reference frame that target candidate set includes, the step of the reference block of selection target coding unit before, mesh Mark the reference block selection method of coding unit further include:

Judge whether target code unit meets the first preset condition, if it is satisfied, by the reference in target candidate set Frame is updated to the reference frame in object reference frame set；

Wherein, the first preset condition include it is following at least one:

Target code unit is located at the boundary of target video frame；

In the encoded adjacent encoder unit of target code unit, the sum of existing intraprediction unit is greater than first Preset threshold；Wherein, the first preset threshold be not less than integer.

Encoded father's coding unit of target code unit, sub- coding unit, in adjacent encoder unit, it is pre- using interframe The sum of the coding unit of survey is less than the second preset threshold, wherein the second preset threshold is the integer greater than 0.

If target code unit meets any of the above-described condition, by joining from target according to the correlation between coding unit It examines and determines that identified target candidate set is all easy for empty set or comprising less the step for target candidate set in frame set Reference frame, be easy to cause the subsequent reference block that can not choose target code unit in this way, or the reference block of selection can not be protected Demonstrate,prove the encoding precision of target code unit.Therefore, by this step by target candidate set adjust to object reference frame set Identical range, it is ensured that the precision of target code unit.

For example, according to the correlation between coding unit, from object reference frame set the step of determining target candidate set, Third set is specifically determined as to the target candidate set of target code unit by mesh.If target code unit is located at target The boundary of video frame, then the reference frame quantity that determining target candidate set is included in this way may be considerably less. When target code unit is located at left margin or the coboundary of video frame, such case is showed particularly evident, that is to say, that institute A possibility that range of determining target candidate set is especially small is higher.At this point, if being adhered to again using originally determining target Candidate collection, therefrom if the reference block of selection target coding unit, the problem that is easy to cause encoding precision not high.

For example, according to the correlation between coding unit, from object reference frame set the step of determining target candidate set, Specifically: third set is determined as to the target candidate set of target code unit.If target code unit is encoded In adjacent encoder unit, the sum of existing intraprediction unit is greater than the first preset threshold, then determines in this way The quantity for the reference frame that target candidate set is included may be considerably less.At this point, if being adhered to again using originally determining target Candidate collection, therefrom if the reference block of selection target coding unit, the problem that is easy to cause encoding precision not high.

For another example determining the step of target candidate set from object reference frame set according to the correlation between coding unit Suddenly, specifically: first set, second set and third union of sets collection are determined as target candidate set.If target code Encoded father's coding unit of unit, sub- coding unit, in adjacent encoder unit, using inter-prediction coding unit it is total Number is less than the second preset threshold, then determining target candidate range of convergence may be especially small in this way.At this point, such as Fruit adheres to being easy to cause therefrom if the reference block of selection target coding unit using originally determining target candidate set again The not high problem of encoding precision.

In a kind of embodiment of the embodiment of the present invention, referring to Fig. 5, from the reference frame that target candidate set includes, choosing The step of selecting the reference block of target code unit, comprising:

S301, it is based on target code unit, carries out whole pixel search, root on the reference frame in target candidate set one by one According to search result, the reference frame of target code unit is determined from the reference frame that target candidate set includes；

The process that this step is searched for one by one can be the search all referring to frame completed in target candidate set, can also To only complete the search of part reference frame.Then according to whole search result or partial search result, from target candidate collection The reference frame of target code unit is determined in the reference frame that conjunction includes.It is specifically searched for using which kind of mode, can be set and searching for In the process, according to actual needs and the adaptive selection of actual search result continues searching or terminates search, thus The whole pixel search of some reference frames can be skipped, so as to improve coding unit reference block efficiency of selection.

S302, it is based on target code unit, a point pixel search is carried out on the reference frame of target code unit, obtained best The optimal reference block of acquisition is determined as the reference block of target code unit by reference block.

A point pixel search only is carried out to the reference frame of target code unit by this step, skips its in target candidate set His reference frame divides pixel search, can also improve the efficiency of selection of the reference block of target code unit.Wherein, a point pixel is searched Rope may include following at least one: half pixel searching, a quarter pixel search, eight/once pixel search.

In a kind of embodiment of the embodiment of the present invention, according to search result, the reference frame that includes from target candidate set The step of reference frame of middle determining target code unit, comprising:

At least one of according to search result, in the following way, from the reference frame that target candidate set includes really Set the goal the reference frame of coding unit:

Mode one, the reference frame by the distortion of whole pixel search less than the target minimum distortion threshold value of preset target video frame It is determined as the reference frame of target code unit；Wherein, target minimum distortion threshold value is the minimum distortion for the setting of target video frame Threshold value.

In mode two, the reference frame for including by target candidate set, whole pixel distortion loses than the whole pixel of other reference frames The reference frame that very small and small amplitude is greater than the whole pixel distortion threshold value of target of preset target video frame is determined as target volume The reference frame of code unit；Wherein, the whole pixel distortion threshold value of target is the whole pixel distortion threshold value for the setting of target video frame.

The whole the smallest reference frame of pixel distortion in object reference frame that target candidate set includes is determined as mesh by mode three Mark the reference frame of coding unit.

When specific implementation, if mode one is used only, whole pixel is carried out on the reference frame in target candidate set one by one In search process, if encountering the reference frame that whole pixel search distortion is less than target minimum distortion threshold value, it can stop search, Just it is to skip the search that other in target candidate set do not carry out the reference frame of whole pixel search.In addition, if being waited for target After having been completely finished whole pixel search all referring to frame in selected works conjunction, whole pixel search distortion is also encountered and is less than preset target The reference frame of the target minimum distortion threshold value of video frame then can carry out intraframe predictive coding to target code unit.

If mode two is used only, one by one during carrying out whole pixel search on the reference frame in target candidate set, It can monitor in real time in the reference frame for having carried out whole pixel search, if it is more whole than other reference frames that there are whole pixel distortions Pixel distortion is small, and small amplitude is greater than the reference frame of the whole pixel distortion threshold value of target, if there is can stop search, Just it is to skip the search that other in target candidate set do not carry out the reference frame of whole pixel search.In addition, if being waited for target After having been completely finished whole pixel search all referring to frame in selected works conjunction, whole picture of the whole pixel distortion than other reference frames is not found yet Element is distorted the reference frame that small and small amplitude is greater than the whole pixel distortion threshold value of target, then can carry out to target code unit Intraframe predictive coding.

If mode three is used only, the whole the smallest reference of pixel distortion in the object reference frame for including by target candidate set Frame is determined as the reference frame of target code unit.In target candidate set all referring to all having carried out whole pixel search on frame, All whole pixel distortions are compared, the smallest whole pixel distortion can be found, may thereby determine that out the ginseng of target code unit Examine frame.

Above-mentioned three kinds of modes can be combined with each other use, complement each other.In the process of whole pixel search, in real time according to group Each mode is monitored in conjunction, when determining the reference frame of target code unit, or after having been completely finished whole pixel search Stop whole pixel search process.

For example, determining that the process of the reference frame of target code unit specifically uses the combination of above-mentioned three kinds of modes, in whole picture In plain search process, implement to monitor whether to be distorted the reference frame for being less than target minimum distortion threshold value there are whole pixel search, if There are the whole pixel distortion amplitudes smaller and small than the whole pixel distortion of other reference frames to be greater than the whole pixel distortion threshold value of target Reference frame.It is understood that both reference frames may be entirely absent, it is also possible to one of which is only existed, in addition, also May all exist, but be present in same reference frame.If in searching target candidate set before last reference frame Any reference frame defines the reference frame of a target code unit by first two mode, then can stop search, also It is to skip the search that other in target candidate set do not carry out the reference frame of whole pixel search.If from first frame to reciprocal the It is not all determined in two frames, then after in target candidate set all referring to frame progress complete pixel search, at least It can be determined as target code by the whole the smallest reference frame of pixel distortion in the object reference frame that includes by target candidate set The step of reference frame of unit, determines the reference frame of target code unit.

Further, target minimum distortion threshold value is default according to the frame type and frame-layer grade of target video frame.

By the setting of this step, it can be adaptively adjusted target minimum distortion threshold value, be conducive to improve binary encoding Precision and code efficiency.

Specifically, " distortion of whole pixel search is being less than preset target video frame based on target minimum distortion threshold value The reference frame of target minimum distortion threshold value is determined as the reference frame of target code unit " effect played in step considers, it can Minimum distortion threshold value is arranged as target video frame based on following principle: relatively small minimum distortion threshold value is arranged for P frame, is Relatively large minimum distortion threshold value is arranged in B frame；Lesser minimum distortion threshold value is set for the lower video frame of frame-layer grade, is frame Biggish minimum distortion threshold value is arranged in the higher video frame of level.Setting is primarily due in this way: using lesser minimum distortion Threshold value so that being possible to determine the reference frame of coding unit in relatively multi-reference frame in candidate collection, and is determined The reference frame of coding unit there is lesser whole pixel search distortion, this makes it possible to improve P frame and the higher view of frame-layer grade The encoding precision of frequency frame.On the contrary, using biggish minimum distortion threshold value, so that just having in less reference frame in candidate collection It may determine out the reference frame of coding unit, be directed to B frame and the lower video frame of frame-layer grade in this way, it will be able to improve coding unit Reference block selection speed, to improve code efficiency.

Again due to by Such analysis it is found that in same image group unit, if the encoding precision of P frame is higher, thus can also So that with reference to its B frame encoding precision with higher.If the encoding precision of the higher video frame of frame-layer grade is higher, thus It can make the lower video frame of the frame-layer grade encoding precision with higher with reference to it.

It follows that relatively small minimum distortion threshold value is arranged for P frame, relatively large minimum distortion is set for B frame Threshold value；Lesser minimum distortion threshold value is set for the lower video frame of frame-layer grade, is that the higher video frame setting of frame-layer grade is larger Minimum distortion threshold value, whole encoding precision not only can be improved, but also whole code efficiency can be improved.

In other embodiments, it may recognize that in same image group unit, the minimum B frame of frame-layer grade, and be it Biggish minimum distortion threshold value is set, is that lesser minimum distortion threshold value is arranged in the video frame of other frame-layer grades.

In addition, the whole pixel distortion threshold value of target be also possible to it is preset according to the frame type and frame-layer grade of target video frame.

In a kind of embodiment of the embodiment of the present invention, based on the whole pixel distortion threshold value of target " by target candidate set In the reference frame for including, the whole pixel distortion amplitude smaller and small than the whole pixel distortion of other reference frames is greater than preset mesh The reference frame for marking the whole pixel distortion threshold value of target of video frame is determined as the reference frame of target code unit " played in step Effect considers, can be that whole pixel distortion threshold value is arranged in target video frame based on following principle, can be arranged for P frame relatively large The whole pixel distortion threshold value of target, for B frame, the relatively small whole pixel distortion threshold value of target is set；For the higher video of frame-layer grade The biggish whole pixel distortion threshold value of target is arranged in frame, is that the lesser whole pixel distortion threshold of target is arranged in the lower video frame of frame-layer grade Value.Setting is primarily due in this way: biggish whole pixel distortion threshold value is used, so that in the relatively multi-reference frame in candidate collection The reference frame for the coding unit for being possible to determine the reference frame of target code unit, and determining has lesser whole picture Element distortion, this makes it possible to the encoding precisions for improving P frame and the higher video frame of frame-layer grade.Similarly, using lesser whole pixel Distortion threshold, so that being possible to determine the reference frame of coding unit, such needle in less reference frame in candidate collection To B frame and the lower video frame of frame-layer grade, it will be able to the selection speed of the reference block of coding unit is improved, to improve coding effect Rate.

It follows that relatively large whole pixel distortion threshold value is arranged for P frame, relatively small whole pixel is set for B frame Distortion threshold；Biggish whole pixel distortion threshold value is set for the higher video frame of frame-layer grade, is set for the lower video frame of frame-layer grade Lesser whole pixel distortion threshold value is set, whole encoding precision not only can be improved, but also whole code efficiency can be improved.

In the another embodiment of the embodiment of the present invention, target candidate set can also be repaired in the following manner Just, on the one hand can guarantee it is subsequent can choose the reference block of target code unit from target candidate set adjusted, On the other hand it is also possible that accelerate target code unit reference block selection course simultaneously, improve encoding precision.Specifically Modification mode include:

From the reference frame that target candidate set includes, the step of the reference block of selection target coding unit before, compile Code unit reference block selection method can also include:

Judge whether target code unit meets the second preset condition, the second preset condition includes: based on target code list Member, the minimum predicted distortion searched on each reference frame in target candidate set be more than preset target video frame target most Big distortion threshold；Wherein, target maximum distortion threshold value can be preset according to the frame type and frame-layer grade of target video frame.Such as Fruit meets, and the reference frame in target candidate set is updated to the reference frame in object reference frame set.If conditions are not met, can be with Any operation is not done to target candidate set.

Specifically, the step of whether target code unit meets the second preset condition judged, may include:

It based on target code unit, is searched on each reference frame in target candidate set, and calculates minimum pre- dendrometry Very；

Judge whether the minimum predicted distortion being calculated is more than target maximum distortion threshold value, if it does, representing target Coding unit meets the second preset condition and represents target code unit if be no more than and be unsatisfactory for the second preset condition.In addition, The method for calculating predicted distortion belongs to the prior art, and this will not be repeated here.

It is understood that being based on target code unit, the minimum searched on each reference frame in target candidate set Predicted distortion is more than target maximum distortion threshold value, it is believed that minimum predicted distortion is larger, if being based on this target candidate collection It closes included reference frame and carries out reference encoder, may result in the low problem of encoding precision.Therefore, it is waited subsequently through by target The step of reference frame in selected works conjunction is updated to the reference frame in object reference frame set, may be implemented to target candidate set model The adjustment enclosed, so as to which the reference block that can make the more accurate coding of target code unit can be found.

When judging whether target code unit meets the second preset condition, target code unit is had been based on, original The reference frame that determining target candidate set is included is searched for, and therefore, subsequent when selecting reference block, there is no need to again Repeat search is carried out to the reference frame of this part.

In the case where target code unit meets the second preset condition, from the reference frame that target candidate set includes, The step of reference block of selection target coding unit, comprising:

Based on target code unit, searched on the reference frame in the difference set before and after the update of target candidate set one by one Rope, according to the search result in target candidate set in the updated on each reference frame, the ginseng that includes from target candidate set It examines in frame, the reference block of selection target coding unit；Or according to the search result on the reference frame in difference set, include from difference set Reference frame in, the reference block of selection target coding unit；

It should be noted that the range of target candidate set may if target code unit meets the second preset condition Effectively adjustment is based on target code unit, has scanned on the reference frame that target candidate set before the update is included, It is scanned on the reference frame in the difference set before and after the update of target candidate set one by one in this step.If according in the updated Target candidate set in search result on each reference frame, the reference block of selection target coding unit, so that it may preferably exist Best matching blocks in this complete or collected works of object reference frame set are provided with are conducive to improve encoding precision in this way.If according in difference set In reference frame on search result, it is equal thus to exclude minimum predicted distortion the reference block of selection target coding unit in advance Greater than the part reference frame of target maximum distortion threshold value, can from difference set selection target coding unit reference block, in this way may be used Calculation amount when reference block is selected to reduce, and efficiency of selection can be improved.

In addition, with greater need for explanation, it is based on target code unit, one by one the difference set before and after the update of target candidate set In reference frame on scan for, according to the search result on the reference frame in difference set, the step of selection target coding unit, It can specifically include:

Firstly, being based on target code unit, the reference frame in the difference set before and after the update of target candidate set is enterprising one by one The whole pixel search of row, according to the whole pixel search on the reference frame in difference set as a result, determining target code list from the difference set The reference frame of member；

Then, it is based on target code unit, a point pixel search is carried out on the reference frame of target code unit, is obtained best The best matching blocks of acquisition are determined as the reference block of target code unit by match block.

In this step, a point pixel search only is carried out to the reference block of target code unit, has skipped other reference frames Divide pixel search, the efficiency of the reference block of selection target coding unit thus can be improved.Wherein, pixel search is divided to can wrap Include following at least one: half pixel searching, a quarter pixel search, eight/once pixel search.

Further, according to the whole pixel search on the reference frame in difference set as a result, determining target from the difference set The step of reference frame of coding unit, may include:

According to the whole pixel search on the reference frame in difference set as a result, at least one of in the following way, from this The reference frame of target code unit is determined in difference set:

The reference frame for the target minimum distortion threshold value that the distortion of whole pixel search is less than preset target video frame is determined as The reference frame of target code unit；

In the reference frame for including by target candidate set, whole pixel distortion is smaller than the whole pixel distortion of other reference frames, And small amplitude is determined as target code unit less than the reference frame of the whole pixel distortion threshold value of target of preset target video frame Reference frame；

The whole the smallest reference frame of pixel distortion in object reference frame that target candidate set includes is determined as target code The reference frame of unit.

Determine that the mode of the reference frame of target code unit is determined from target candidate set with above-mentioned in present embodiment The mode of the reference frame of target code unit is identical, is only that the maximum magnitude of search is different, the former is that target candidate set updates The difference set of front and back, the latter are object reference frame set, and therefore, details are not described herein.

In the case where target code unit is unsatisfactory for the second preset condition, the reference frame that includes from target candidate set In, the step of the reference block of selection target coding unit, comprising:

According to target code unit is based on, the search result in target candidate set on each reference frame is obtained best The optimal reference block of acquisition is determined as the reference block of target code unit by match block.

Using above method embodiment, compared to traditional coding mode, coding rate is promoted, specifically such as 1 institute of table Show.

Table 1

In table 1, a column of visual classification, corresponding movie, 1080p, 240p, 720p, cif and 480p are from different angles Spend the type divided, wherein movie is particular set, and the image scene switching in the set is relatively more, remaining be according to The clarity of video is distinguished.The column of coding rate one are corresponding with the column of visual classification one, indicate the coding rate of each video；It is right Four column of side are corresponding with the column of visual classification one respectively, indicate that code rate of each video under Y, U, V and YUV merged-quality increases Add；Average correspond to row each value be each column average value.

On the basis of above method embodiment, the embodiment of the invention also provides a kind of selections of the reference block of coding unit Device, referring to Fig. 6, which includes:

Module 61 is obtained, for obtaining the object reference frame set of target code unit said target video frame, wherein mesh The quantity for marking the included reference frame of reference frame set and the reference frame is set according to the frame type and frame-layer grade of target video frame It sets；

Selecting module 62, for from the reference frame that object reference frame set includes, the reference of selection target coding unit Block.

It, compared with the prior art, can be adaptively according to frame type and frame-layer grade to target using embodiment illustrated in fig. 6 The quantity of reference frame and reference frame in reference frame set is limited, and the meter when reference block selection of coding unit can be reduced Calculation amount, so as to improve coding unit reference block efficiency of selection.

In a kind of embodiment of the embodiment of the present invention, referring to Fig. 7, selecting module 62 includes:

Submodule 621 is determined, for determining that target is waited from object reference frame set according to the correlation between coding unit Selected works close；

Select submodule 622, for from the reference frame that target candidate set includes, the reference of selection target coding unit Block.

In a kind of embodiment of the embodiment of the present invention, determine that submodule 621 is specifically used for:

Target code unit, father's coding unit that motion estimation search is completed candidate collection；

Target code unit, sub- coding unit that motion estimation search is completed candidate collection；

Target code unit, adjacent encoder unit that motion estimation search is completed candidate collection.

In a kind of embodiment of the embodiment of the present invention, selecting module 62 further include:

First processing submodule, for being executed from the reference frame that target candidate set includes in selection submodule 622, choosing Before the step of selecting the reference block of target code unit, judge whether target code unit meets the first preset condition, if full Foot, is updated to the reference frame in object reference frame set for the reference frame in target candidate set；

Wherein, the first preset condition include it is following at least one:

Target code unit is located at the boundary of target video frame；

In the encoded adjacent encoder unit of target code unit, the sum of existing intraprediction unit is greater than first Preset threshold；

Encoded father's coding unit of target code unit, sub- coding unit, in adjacent encoder unit, it is pre- using interframe The sum of the coding unit of survey is less than the second preset threshold.

Further, in a kind of embodiment of the embodiment of the present invention, referring to Fig. 8, selection submodule 622 includes:

Search unit 6221 carries out on the reference frame in target candidate set one by one for being based on target code unit Whole pixel search；

First determination unit 6222, for determining mesh from the reference frame that target candidate set includes according to search result Mark the reference frame of coding unit；

Second determination unit 6223 is divided on the reference frame of target code unit for being based on target code unit Pixel search obtains optimal reference block, the optimal reference block of acquisition is determined as to the reference block of target code unit.

Further, the first determination unit 6222 is specifically used for:

Specifically, target minimum distortion threshold value and the whole pixel distortion threshold value of target are the frame types according to target video frame It is preset with frame-layer grade.

In a kind of embodiment of the embodiment of the present invention, selecting module 62 further includes that second processing submodule (does not show in figure Out)；

Second processing submodule, for being executed from the reference frame that target candidate set includes in selection submodule 622, choosing Before the step of selecting the reference block of target code unit, judge whether target code unit meets the second preset condition, if full Foot, is updated to the reference frame in object reference frame set for the reference frame in target candidate set；

Second preset condition includes: to be searched on each reference frame in target candidate set based on target code unit Minimum predicted distortion is more than the target maximum distortion threshold value of preset target video frame；Wherein, target maximum distortion threshold value is root It is preset according to the frame type and frame-layer grade of target video frame.

In a kind of embodiment of the embodiment of the present invention, selecting submodule 622 includes: that third determination unit and the 4th determine Unit (not shown)；

Third determination unit is when meeting, one by one in target candidate collection for the judging result in second processing submodule It closes and is scanned on the reference frame in the difference set for updating front and back, according to each reference frame in target candidate set in the updated Search result, or according to the search result in difference set on reference frame, the reference block of selection target coding unit；

4th determination unit is when being unsatisfactory for, to compile according to based on target for the judging result in second processing submodule Code unit, the search result in target candidate set on each reference frame, the reference block of selection target coding unit.

The embodiment of the invention also provides a kind of electronic equipment, as shown in figure 9, including processor 91 and memory 92；

Memory 92, for storing computer program；

Processor 91 when for executing the program stored on memory 92, realizes following steps:

Obtain the object reference frame set of target code unit said target video frame, wherein object reference frame set institute Quantity comprising reference frame and the reference frame is arranged according to the frame type and frame-layer grade of target video frame；

From the reference frame that object reference frame set includes, the reference block of selection target coding unit.

The specific implementation of each step performed by processor 91 and relevant explanation content may refer to above-mentioned side Method embodiment, this will not be repeated here.

Above-mentioned electronic equipment can have the communication interface realized and communicated between above-mentioned electronic equipment and other equipment.

In above-mentioned electronic equipment, memory 92 and processor 91 can complete mutual communication by communication bus, Communication bus mentioned herein can be Peripheral Component Interconnect standard (Peripheral Pomponent Interconnect, letter Claim PCI) and bus or expanding the industrial standard structure (Extended Industry Standard Architecture, referred to as EISA) bus etc..The communication bus can be divided into address bus, data/address bus, control bus etc..For convenient for indicating, in figure only It is indicated with a thick line, it is not intended that an only bus or a type of bus.

Above-mentioned memory 92 may include random access memory (Random Access Memory, abbreviation RAM), It may include nonvolatile memory (non-volatile memory), for example, at least a magnetic disk storage.Optionally, it deposits Reservoir can also be that at least one is located remotely from the storage device of aforementioned processor.

Above-mentioned processor 91 can be general processor, including central processing unit (Central Processing Unit, abbreviation CPU), network processing unit (Ne twork Processor, abbreviation NP) etc.；It can also be digital signal processor (Digital Signal Processing, abbreviation DSP), specific integrated circuit (Applica tion Specific Integrated Circuit, abbreviation ASIC), field programmable gate array (Field-Programmable Gate Array, Abbreviation FPGA) either other programmable logic device, discrete gate or transistor logic, discrete hardware components.

It, can be adaptively according to frame type and frame-layer grade to object reference frame set using scheme provided in this embodiment In reference frame and the quantity of reference frame limited, can also reduce the calculation amount of target code elements reference block selection, So as to improve coding unit reference block efficiency of selection.

In another embodiment provided by the invention, a kind of computer readable storage medium is additionally provided, which can It reads to be stored with instruction in storage medium, when run on a computer, so that computer executes any volume in above-described embodiment Method and step described in the selection method of the reference block of code unit.

In another embodiment provided by the invention, a kind of computer program product comprising instruction is additionally provided, when it When running on computers, so that computer executes the step of method described in the selection method of the reference block of any of the above-described coding unit Suddenly.

In another embodiment provided by the invention, a kind of computer program is additionally provided, when it runs on computers When, so that computer executes method and step described in the selection method of the reference block of any of the above-described coding unit.

It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that There is also other identical elements in process, method, article or equipment including element.

Each embodiment in this specification is all made of relevant mode and describes, same and similar portion between each embodiment Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for system reality For applying example, since it is substantially similar to the method embodiment, so being described relatively simple, related place is referring to embodiment of the method Part explanation.

The above is merely preferred embodiments of the present invention, it is not intended to limit the scope of the present invention.It is all in this hair Any modification, equivalent replacement, improvement and so within bright spirit and principle, are included within the scope of protection of the present invention.

Claims

1. a kind of reference block selection method of coding unit, which is characterized in that the described method includes:

Obtain the object reference frame set of target code unit said target video frame, wherein object reference frame set institute The quantity of the reference frame and the reference frame that include is arranged according to the frame type and frame-layer grade of the target video frame；

2. the method according to claim 1, wherein from the reference frame that the object reference frame set includes, The step of selecting the reference block of the target code unit, comprising:

3. according to the method described in claim 2, it is characterized in that, the correlation according between coding unit, from the mesh Mark the step of target candidate set is determined in reference frame set, comprising:

4. according to the method described in claim 3, it is characterized in that, from the reference frame that the target candidate set includes, Before the step of selecting the reference block of the target code unit, the method also includes:

Judge whether the target code unit meets the first preset condition, if it is satisfied, by the target candidate set Reference frame is updated to the reference frame in the object reference frame set；

Wherein, first preset condition include it is following at least one:

The target code unit is located at the boundary of the target video frame；

Encoded father's coding unit of the target code unit, sub- coding unit, in adjacent encoder unit, it is pre- using interframe The sum of the coding unit of survey is less than the second preset threshold.

5. according to the described in any item methods of claim 2-4, which is characterized in that the reference for including from the target candidate set In frame, the step of selecting the reference block of the target code unit, comprising:

Based on the target code unit, whole pixel search, root are carried out on the reference frame in the target candidate set one by one According to search result, the reference frame of the target code unit is determined from the reference frame that the target candidate set includes；

Based on the target code unit, a point pixel search is carried out on the reference frame of the target code unit, is obtained best The optimal reference block of acquisition is determined as the reference block of the target code unit by reference block.

6. according to the method described in claim 5, it is characterized in that, described according to search result, from the target candidate set The step of reference frame of the target code unit is determined in the reference frame for including, comprising:

At least one of according to search result, in the following way, from the reference frame that the target candidate set includes really The reference frame of the fixed target code unit:

The reference frame for the target minimum distortion threshold value that the distortion of whole pixel search is less than the preset target video frame is determined as The reference frame of the target code unit；

In the reference frame for including by the target candidate set, whole pixel distortion is smaller than the whole pixel distortion of other reference frames, And small amplitude is determined as the target less than the reference frame of the whole pixel distortion threshold value of target of the preset target video frame The reference frame of coding unit；

The whole the smallest reference frame of pixel distortion in object reference frame that the target candidate set includes is determined as the target The reference frame of coding unit.

7. according to the method described in claim 6, it is characterized in that,

The target minimum distortion threshold value and the whole pixel distortion threshold value of target be according to the frame type of the target video frame and Frame-layer grade is preset.

8. according to the method described in claim 3, it is characterized in that, from the reference frame that the target candidate set includes, Before the step of selecting the reference block of the target code unit, the method also includes:

Judge whether the target code unit meets the second preset condition, second preset condition includes: based on the mesh Coding unit is marked, the minimum predicted distortion searched on each reference frame in the target candidate set is more than the preset mesh Mark the target maximum distortion threshold value of video frame；

If it is satisfied, the reference frame in the target candidate set is updated to the reference frame in the object reference frame set.

9. according to the method described in claim 8, it is characterized in that, meeting the described second default item in the target code unit In the case where part, from the reference frame that the target candidate set includes, the step of the reference block of the target code unit is selected Suddenly, comprising:

Based on the target code unit, carried out on the reference frame in the difference set before and after target candidate set update one by one Search, according to the search result in target candidate set in the updated on each reference frame, from the target candidate set packet In the reference frame contained, the reference block of the target code unit is selected；Or according to the search on the reference frame in the difference set As a result, selecting the reference block of the target code unit from the reference frame that the difference set includes；

In the case where the target code unit is unsatisfactory for second preset condition, include from the target candidate set In reference frame, the step of selecting the reference block of the target code unit, comprising:

According to the target code unit is based on, the search result in the target candidate set on each reference frame is selected The reference block of the target code unit.

10. method according to claim 8 or claim 9, which is characterized in that

The target maximum distortion threshold value is preset according to the frame type and frame-layer grade of the target video frame.

11. a kind of reference block selection device of coding unit, which is characterized in that described device includes:

Module is obtained, for obtaining the object reference frame set of target code unit said target video frame, wherein the target The quantity of the included reference frame of reference frame set and the reference frame is the frame type and frame-layer according to the target video frame Grade setting；

Selecting module, for selecting the ginseng of the target code unit from the reference frame that the object reference frame set includes Examine block.

12. device according to claim 11, which is characterized in that the selecting module includes:

Submodule is determined, for determining target candidate from the object reference frame set according to the correlation between coding unit Set；

Submodule is selected, for selecting the ginseng of the target code unit from the reference frame that the target candidate set includes Examine block.

13. device according to claim 12, which is characterized in that the determining submodule is specifically used for:

By any one following set or it is following any two or more union of sets collection be determined as target candidate set:

14. device according to claim 13, which is characterized in that the selecting module further include:

First processing submodule, for being executed from the reference frame that the target candidate set includes in the selection submodule, Before the step of selecting the reference block of the target code unit, judge whether the target code unit meets the first default item Part, if it is satisfied, the reference frame in the target candidate set is updated to the reference frame in the object reference frame set；

Wherein, first preset condition include it is following at least one:

The target code unit is located at the boundary of the target video frame；

15. the described in any item devices of 2-14 according to claim 1, which is characterized in that the selection submodule includes:

Search unit carries out on the reference frame in the target candidate set one by one for being based on the target code unit Whole pixel search；

First determination unit, for determining the mesh from the reference frame that the target candidate set includes according to search result Mark the reference frame of coding unit；

Second determination unit is divided on the reference frame of the target code unit for being based on the target code unit Pixel search obtains optimal reference block, the optimal reference block of acquisition is determined as to the reference block of the target code unit.

16. device according to claim 15, which is characterized in that first determination unit is specifically used for:

17. device according to claim 16, which is characterized in that

18. device according to claim 13, which is characterized in that the selecting module further include:

Second processing submodule, for being executed from the reference frame that the target candidate set includes in the selection submodule, Before the step of selecting the reference block of the target code unit, judge whether the target code unit meets the second default item Part, if it is satisfied, the reference frame in the target candidate set is updated to the reference frame in the object reference frame set；

Second preset condition includes: each reference frame in the target candidate set based on the target code unit The minimum predicted distortion of upper search is more than the target maximum distortion threshold value of the preset target video frame.

19. device according to claim 18, which is characterized in that the selection submodule includes:

Third determination unit is when meeting, one by one in target candidate collection for the judging result in the second processing submodule It closes and is scanned on the reference frame in the difference set for updating front and back, according to each reference frame in target candidate set in the updated Search result, or according to the search result in the difference set on reference frame, select the reference block of the target code unit；

4th determination unit is when being unsatisfactory for, according to based on the mesh for the judging result in the second processing submodule Coding unit is marked, the search result in the target candidate set on each reference frame selects the target code unit Reference block.

20. device described in 8 or 19 according to claim 1, which is characterized in that

21. a kind of electronic equipment, which is characterized in that including processor and memory,

The memory, for storing computer program；

The processor when for executing the program stored on memory, realizes any method of claim 1-10 Step.

22. a kind of computer readable storage medium, which is characterized in that be stored with computer in the computer readable storage medium Program realizes claim 1-10 any method and step when the computer program is executed by processor.