CN106550237B - Monitoring video compression method - Google Patents
Monitoring video compression method Download PDFInfo
- Publication number
- CN106550237B CN106550237B CN201510586444.2A CN201510586444A CN106550237B CN 106550237 B CN106550237 B CN 106550237B CN 201510586444 A CN201510586444 A CN 201510586444A CN 106550237 B CN106550237 B CN 106550237B
- Authority
- CN
- China
- Prior art keywords
- frame
- vector
- background
- vector field
- macro block
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention relates to the technical field of video processing, in particular to a monitoring video compression method, which comprises the following steps: aiming at the characteristic that the background of the I frame is unchanged for a long time, removing redundant information brought by the I frame which appears repeatedly; p-frames use a selective frame skipping method to reduce the amount of data; for the jumping feeling possibly brought to the video after P frame jumping, the jumping feeling is reduced by using an interpolation frame reconstruction method. The monitoring video compression processing method performs targeted optimization measures aiming at the characteristics of the monitoring video on the basis of the original standard, reduces the size of a code stream as far as possible on the premise of not influencing the video reconstruction quality, and effectively reduces the possibility of data congestion of an indoor monitoring wireless sensor network.
Description
[ technical field ] A method for producing a semiconductor device
The invention relates to the technical field of video processing, in particular to a monitoring video compression method.
[ background of the invention ]
The wireless multimedia sensor network meets the requirements of people on regional information collaborative acquisition, analysis and processing, is different from the traditional low-speed sensor network, has stricter requirements on the information acquisition and processing capacity of the sensor nodes by the high-speed sensor network, and has an increasing requirement on the load capacity capable of being borne by the network. Thus, many key technologies are derived to be urgently researched and solved, such as: 1. high compression ratio multimedia data compression problem; 2. network multimedia data congestion, real-time issues; 3. data storage problems. One of the root causes of these problems is to effectively compress the acquired data.
With the rapid increase of the video codec service demand, the international standards organization performs a series of works on the aspect of the formulation of the unified codec standard, wherein the video coding experts Group VCEG (video coding Expert Group) under the international telecommunication union ITU-T and the motion picture experts Group MPEG (motion picture Expert Group) under the international standardization organization ISO/IEC have the greatest influence in this field. The following are the more important common criteria established by two major organizations:
(1) h.264 standard
H.264 is also known as MPEG-4/AVC, which is the product of a joint study of two large video image organizations. One of the initial goals of its formulation is to meet the increasing demands of network video services, so it has good network adaptability. Compared with the past compression standard, the compression standard has larger compression ratio, so that the method is widely applied to the field of multimedia compression, such as video on demand service, high-definition DVD compression and video compression in a wireless environment. Compared with the prior compression standard, the method has the advantages of two aspects:
low code rate: due to the adoption of a more flexible macro block dividing means and the introduction of a more mature entropy coding mechanism CABAC, the prediction accuracy and the residual error coding efficiency of H.264 are greatly improved compared with the prior standard. Experiments show that the compression ratio of H.264 can reach more than one time of MPEG-4 under the same video effect. In actual use, network bandwidth resources can be greatly saved.
The fault tolerance capability is strong: the H.264 video coding and decoding standard is divided into a video coding layer and a network extraction layer, so that the adaptability to various channels is enhanced, error codes and packet loss errors can be better solved, and the fault-tolerant capability of H.264 is doubled compared with that of MPEG-4.
(2) AVS Standard
The AVS (Audio Video coding Standard) Standard is a Video compression Standard with the proprietary property rights of China, is a short for a series of standards, and is still in the process of continuous establishment and perfection nowadays. The AVS has the following characteristics: the method can be applied to various video compression technology application fields, the compression performance of the method is equivalent to that of H.264, and the method has obvious advantages compared with compression algorithms such as H.261, H.262, H.263, MPEG-1, MPEG-2 and the like. In addition to having the same general compression standard characteristics as h.26x and MPEG-X, AVS also makes independent standards in this particular aspect of surveillance video coding. Aiming at the compression of the surveillance video, the AVS revision AVS-S2 is different from the H.265 which adds a more flexible macro block dividing method and a macro block prediction mechanism to improve the compression ratio, the time redundancy caused by the long-term invariance of the background scene of the surveillance video is eliminated, and the coding time and the compression ratio are improved by more than one time through the learning modeling of the background and the foreground.
(3)H.265
With the rise of 4K technology, the original h.264 technology has not been able to meet the requirement of high-definition video compression transmission such as 1080P. H.265 is a new video coding standard following h.264, and hua cheng corporation participated in the establishment of its standard, with most of its patents. The H.265 standard reserves some original technologies of H.264, keeps the encoding process thereof, comprises modules of predictive encoding, residual error conversion, quantization, entropy encoding and the like, and improves the technologies of a block matching search algorithm, macro block division, intra-frame prediction direction and the like. The use of new techniques allows a significant improvement in compression ratio and coding quality, but with a consequent increase in the complexity of the algorithm and the consumption of resources by the computation. Experiments show that under the same coding condition, the H.265 can reduce the code stream by 39-44% compared with the H.264, but the consumption of the power consumption is also increased by four times.
(4) Distributed encoding
The traditional video compression technology needs time-consuming steps such as coding mode selection, code stream control and the like at a coding end, so that the decoding is simple and the coding is difficult. The encoding mode is suitable for the occasions of one-time encoding and multiple-time decoding, such as optical disc storage and the like, but the encoding standard does not conform to the characteristic of low power consumption of the sensor nodes in the wireless sensor network. Distributed video coding arises for compression of sensor network data. The coding frame places the complexity of coding and decoding at a decoding end instead of a coding end, has the characteristics of complex decoding and simple coding, has satisfactory compression ratio and is very suitable for a wireless multimedia sensor network. Such video coding concepts were first proposed by lepian, Wolf, Wyner, Ziv, and are being studied by an increasing number of scholars. But in the stage of just starting research, theories and algorithms are not mature, but have great research and commercial values.
[ summary of the invention ]
In view of the above, there is a need for a high-energy efficient surveillance video compression method.
A surveillance video compression method is characterized by comprising the following steps:
A. judging the background of the monitoring video, and further removing I frame redundant information on the basis of the traditional I frame coding if the background area of the monitoring video is in an unchanged state for a long time;
B. selectively starting a frame skipping mechanism of the P frame, judging the relevance of the previous frame and the next frame, and skipping the frame if the relevance of the previous frame and the next frame is greater than a preset value;
C. in view of the influence of the skip feeling of the skip frame on the video quality, an interpolation frame reconstruction technology is used at a decoding end to reconstruct the skipped frame.
2. The indoor surveillance video compression method according to claim 1, wherein the further removal of I-frame redundant information adopted in step 1) comprises the steps of:
a. extracting and updating a monitoring video background of a scene;
b. the background macro block is represented by a proper interface, the traditional compression coding mode is abandoned for the macro block of the type, and only one mark of the background macro block is transmitted, and the mark is used for telling a decoding end: the macroblock is a background macroblock;
c. when the decoding end receives the background macro block mark, the background data which is cached is directly copied and filled to reconstruct the background data.
Further, the P frame adopted in the step (2) is a standard for selectively performing frame skipping coding, i.e.
Wherein, V (j) represents the residual space of the buffer zone at the encoder end when the j frame is encoded, f represents the encoded frame rate, and BD (j) represents the current time and the channel bandwidth; b (i) represents the actual bit number of the ith frame after being coded, and N (j) is the number of the residual frames of the current estimated moving object passing through the camera area; r (j) reflects the average number of coded bits per frame now; b (j) is an estimate of whether there is a value within N (j) frames that would make buffer overflow possible under current code rate and network transmission conditions;
if B (j) < ═ 0, it indicates that there is moving object not leaving the monitoring area, the buffer has already begun to overflow, the selective frame skipping mechanism starts;
after each frame has been encoded, the buffer and N (j) need to be updated according to the display, i.e.
N(j)=N(j)-1。
Still further, the step (3) of reconstructing the skipped frame by using the frame interpolation reconstruction technique at the decoding end specifically includes the following steps:
I. vector field confidence partitioning
Firstly, determining the reliability of a vector field obtained from an H.264 code stream, wherein a motion vector of an I macro block in a P frame initializes zero; dividing a vector field according to residual energy of a corresponding macro block in an H.264 code stream, and forcibly classifying a vector corresponding to an I-type coding macro block appearing in a P frame into an untrusted vector, namely the vector
Wherein R isY(i,j),RCr(i, j) represent the energy difference between the reconstructed macroblock and the real image in chrominance and luminance respectively; then dividing the vector field of the whole image according to a set threshold value, namely
Wherein L3 is a trusted vector field and L1 is an untrusted vector field;
II. Vector field selection
Vector correction is performed on the untrusted vector selected in the previous step, and the untrusted vector is corrected by using the trusted vectors surrounding it, i.e. the untrusted vector is corrected by using the trusted vectors
Wherein
B represents the macroblock corresponding to the current unreliable vector, where the range S of the candidate vectors is selected to be all vectors in eight 16X16 macroblocks around the 16X16 macroblock where the current vector is located,
III, vector field Re-judgment
The corrected vector field is subdivided in accordance with the method of the first step, i.e.
Where L3 is the trusted vector field, L2 is the untrusted vector field,
IV, vector field reselection
And C, carrying out full search algorithm matching in a certain limited range S on the macro block corresponding to the L2 vector judged in the step III to carry out vector assignment, namely carrying out vector assignment
Wherein
S=(sx,sy)where-sx<=vx<=sx,-sy<=vy<=sy。
The invention relates to an indoor monitoring video compression method based on H.264 standard, which removes redundant information brought by I frames which appear repeatedly by aiming at the characteristic that the background of the I frames is not changed for a long time; p-frames use a selective frame skipping method to reduce the amount of data; aiming at the jumping sense possibly brought to the video after the P frame is skipped, the jumping sense is reduced by using an interpolation frame reconstruction method, so that the compression of the monitoring video is completed. The invention provides a simpler processing method aiming at the problems that the indoor monitoring video compression based on the H.264 standard is easy to have unstable code rate and the data volume is increased rapidly periodically. And the complexity of coding and decoding is put at a decoding end as much as possible, so that the method can be suitable for compressing use scenes on wireless equipment such as a wireless sensor network and the like.
[ description of the drawings ]
FIG. 1 is a schematic diagram of an encoding process of a surveillance video compression method according to the present invention;
FIG. 2 is a schematic diagram illustrating a decoding process of a surveillance video compression method according to the present invention;
fig. 3 is a schematic diagram of I-frame background extraction, compression and encoding in the surveillance video compression method of the present invention.
[ detailed description ] embodiments
In order to make the objects, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In the embodiment, the surveillance video compression method is an indoor surveillance video compression method based on the H.264 standard; the method comprises the following steps: 1. aiming at the characteristic that the background of the I frame is unchanged for a long time, removing redundant information brought by the I frame which appears repeatedly; 2. p-frames use a selective frame skipping method to reduce the amount of data; 3. for the jumping feeling possibly brought to the video after P frame jumping, the jumping feeling is reduced by using an interpolation frame reconstruction method.
As shown in fig. 1 and fig. 2, in a preferred embodiment, removing redundant information from repeated I frames includes the following steps: by adopting the background extraction technology, the background macro block is represented by using a proper interface and is specially processed, compression coding is abandoned for the macro block of the type, and when the decoding end encounters the background macro block, the buffered background data is directly copied to fill the macro block data for decoding.
In a preferred embodiment, the background extraction formula is:
y (x, Y, t) represents the luminance value of the pixel at the (x, Y) position at time t, and the entire video length is H. The video is divided into a plurality of periods, and each period contains L data. The entire video sequence is divided into I segments, of which
Determining an expectation of pixel brightness for each small time period
I1, 2,3
Calculating the variance of the pixels in each time segment
I1, 2,3
The time period with the shortest variance is selected, the pixel is considered to be in a background state in the time period, and the brightness in the time period is expected to be the real background brightness.
The chroma value of the background image pixel is extracted according to the method.
In a preferred embodiment, the context update mechanism is as follows:
the result of the H.264 code stream reflects the change of the current environment to a certain extent. When the environment changes dramatically, the size of the X264 code stream is maintained in a higher range. When the environment is stable and no change occurs or the change is completed, the current state is stable, and the size of the X264 code stream is maintained at a lower level. When moving objects appear in the monitoring range, the objects are in a slow motion state in time, and the size of the code stream still shows a larger difference compared with that in an unmanned state.
The method only adopts the frame with smaller code stream after compression to update the background image. The updating algorithm basically adopts a multi-frame averaging algorithm, but adds some limiting conditions. Therefore, the method can effectively avoid the ghost phenomenon of the traditional multi-frame averaging method, and removes the influence on the updating result caused by improper threshold setting.
Where CR (i) represents the compression ratio of the ith frame
if|∑Bm(x,y)-∑Bm-1(x,y)|>T,changed
if|∑Bm(x,y)-∑Bm-1(x,y)|<=T,unchanged (5)
The threshold T can be set to be smaller, and because of the limited condition, the influence of undersize threshold selection and noise on the extraction of the real background can be greatly reduced.
When the authenticity change of the background is judged, the system adopts an I frame coding mode specified by H.264 and a background image after synchronous change of a decoding end, and the decoding end caches background data and carries out subsequent I frame reconstruction. In a preferred embodiment, the presentation interface selection and special handling of the background macroblock is as follows: i _ PCM (Pulse Code Modulation) is a special intra coding mode. The macro block of the coding mode is selected, the traditional coding processes of prediction, quantization and entropy coding are not adopted, and pixel values in the macro block are directly written into a code stream. But in the surveillance video stream, we never found this type of macroblock to be present, so we have selected this macroblock type to represent the background macroblock. In a video surveillance sequence, background macroblocks often appear in a regional aggregate. In the above proposed algorithm, each background macroblock occupies an I _ PCM, and in each macroblock data in the code stream, the macroblock header occupies a certain data space. This in turn creates new data redundancy. For these repeatedly occurring I _ PCM macroblocks, we use run-length coding to eliminate the redundancy. In encoding, for background macroblocks that appear continuously, we do not encode directly, but count the number of times they appear continuously until a foreground macroblock is encountered. In the background macroblock, we fill in the number of statistics in the PCM macroblock. As shown in fig. 3:
in a preferred embodiment, the criteria for selective frame skipping of P frames are as follows:
wherein, V (j) represents the residual space of the buffer area at the encoding end when the j frame is encoded, f represents the encoding frame rate, and BD (j) represents the current time and the channel bandwidth. b (i) represents the actual number of coded bits of the ith frame, and N (j) is the number of residual frames of the current estimated moving object passing through the camera area. R (j) reflects the average number of coded bits per frame now. B (j) is an estimate of whether it is possible to overflow the buffer within N (j) frames under current code rate and network transmission conditions. N is typically initialized to 75 because the code rate is 15-20 frames per second, and the time for a typical person to pass through the surveillance area is within five seconds.
Therefore, if B (j) < ═ 0, it indicates that there is a possibility that the buffer has already started overflowing when the moving object does not leave the monitored area, so the selective frame skipping mechanism starts.
After each frame has been encoded, the buffer and N (j) need to be updated as displayed:
N(j)=N(j)-1 (8)
in a preferred embodiment, the method for reconstructing the P-frame skipped frame interpolation is as follows:
1) vector field confidence partitioning
First, the confidence level of the vector field obtained from the h.264 code stream is determined, where the motion vector of the I macroblock in the P frame initializes zero. According to experiments, the macro block with accurate motion vector is found, and compared with a real frame, residual energy is very low after reconstruction. And in the places with high residual energy, such as the edge of a moving object, the motion vector is often complex and the error is easy to occur. In these places, the probability of occurrence of I-type macroblocks is also high. The vector field is therefore divided from the residual energy:
wherein R isY(i,j),RCrAnd (i, j) respectively represent the energy difference of the reconstructed macro block and the real image in chrominance and luminance. Then, dividing the vector field of the whole image according to a set threshold value:
where L3 is the trusted vector field and L1 is the untrusted vector field. And the I-type coded macro blocks appearing in the P frame are classified as an unreliable vector field.
MV(m,n)={L3,if Blocktype=I} (11)
2) Vector field selection
And carrying out vector correction on the unreliable vector selected in the last step. In addition to having a large correlation in time, moving objects also have a strong correlation in space. The motion vector of a macroblock tends to be very similar to the surrounding motion vectors. Therefore, trusted vectors around the untrusted vector are selected to be corrected:
wherein
And B represents the macro block corresponding to the current unreliable vector. The range S of candidate vectors is selected here as all vectors within eight 16X16 macroblocks around the 16X16 macroblock where the current vector is located.
3) Vector field re-determination
The previous step corrects the untrusted vector field, but the corrected result does not represent the true motion vector, and the corrected vector field is again divided and corrected according to the method of the first step.
Where L3 is the trusted vector field and L2 is the untrusted vector field.
4) Vector field reselection
And (3) carrying out full search algorithm matching in a certain limited range S on the macro block corresponding to the L2 vector judged in the previous step to carry out vector assignment:
wherein
S=(vx,vy) where -Sx<=vx<=Sx,Sy<=vy<=Sy(14)
The principle of the invention mainly lies in that the reasonable reuse of an H.264 interface is carried out on an I frame, the frame skipping processing is carried out on a P frame aiming at the reasonable prediction of a buffer area of an encoder and the motion intensity of an object, and the vector reconstruction is carried out based on the residual energy of the area. It will be apparent to those skilled in the art that various modifications and enhancements can be made without departing from the principles of the invention, and such modifications and enhancements are intended to be included within the scope of the invention.
Claims (3)
1. A surveillance video compression method is characterized by comprising the following steps:
1) judging the background of the monitoring video, and further removing I frame redundant information on the basis of the traditional I frame coding if the background area of the monitoring video is in an unchanged state for a long time;
2) selectively starting a frame skipping mechanism of the P frame, judging the relevance of the previous frame and the next frame, and skipping the frame if the relevance of the previous frame and the next frame is greater than a preset value;
3) the skipped frames are reconstructed at a decoding end by using an inserted frame reconstruction technology aiming at the influence of the skipping sense brought by the skipping frames on the video quality;
wherein the P frames taken in step 2) are selectively subjected to frame skipping coding, i.e.
Wherein, V (j) represents the residual space of the buffer area at the encoder end when the j frame is encoded, f represents the encoded frame rate, and BD (j) represents the channel bandwidth at the moment; b (i) represents the actual bit number of the ith frame after being coded, and N (j) is the number of the residual frames of the current estimated moving object passing through the camera area; r (j) reflects the average number of coded bits per frame now; b (j) is an estimate of whether there is a buffer overflow value within the N (j) frame under current code rate and network transmission conditions;
if B (j) < ═ 0, it indicates that there is moving object not leaving the monitoring area, the buffer has already begun to overflow, the selective frame skipping mechanism starts;
after each frame is encoded, the remaining space sum N (j) of the buffer area needs to be updated according to the display, namely
N(j)=N(j)-1。
2. The surveillance video compression method according to claim 1, wherein the further removal of I-frame redundant information taken in step 1) comprises the steps of:
a. extracting and updating a monitoring video background of a scene;
b. the indicating interface of the background macro block adopts I _ PCM intra-frame coding mode, and abandons the traditional compression coding mode for the type macro block, and only transmits an indication of the background macro block, and the indication is used for informing the decoding end of: the macroblock is a background macroblock;
c. when the decoding end receives the background macro block mark, the background data which is cached is directly copied and filled to reconstruct the background data.
3. The surveillance video compression method according to claim 1, wherein the surveillance video compression method is specifically an indoor surveillance video compression method based on the h.264 standard; the step 3) of reconstructing the skipped frame at the decoding end by using an interpolated frame reconstruction technique specifically includes the following steps:
I. vector field confidence partitioning
Firstly, determining the reliability of a vector field obtained from an H.264 code stream, wherein a motion vector of an I macro block in a P frame initializes zero; dividing a vector field according to residual energy of a corresponding macro block in an H.264 code stream, and forcibly classifying a vector corresponding to an I-type coding macro block appearing in a P frame into an untrusted vector, namely the vector
Wherein R isY(i,j),RCr(i, j) represent the energy difference between the reconstructed macroblock and the real image in chrominance and luminance respectively; then dividing the vector field of a plurality of whole images according to a set threshold value, namely
Wherein L3 is a trusted vector field and L1 is an untrusted vector field;
II. Vector field selection
Vector correction is performed on the untrusted vector selected in the previous step, and the untrusted vector is corrected by using the trusted vectors surrounding it, i.e. the untrusted vector is corrected by using the trusted vectors
Wherein
B represents the macroblock corresponding to the current unreliable vector, where the range S of the candidate vectors is selected to be all vectors in eight 16X16 macroblocks around the 16X16 macroblock where the current vector is located,
III, vector field Re-judgment
The corrected vector field is subdivided in accordance with the method of the first step, i.e.
Where L3 is the trusted vector field, L2 is the untrusted vector field,
IV, vector field reselection
And carrying out full search algorithm matching in a certain limited range S on the macro block corresponding to the L2 vector judged in the last step to carry out vector assignment, namely carrying out vector assignment
Wherein
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510586444.2A CN106550237B (en) | 2015-09-16 | 2015-09-16 | Monitoring video compression method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510586444.2A CN106550237B (en) | 2015-09-16 | 2015-09-16 | Monitoring video compression method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106550237A CN106550237A (en) | 2017-03-29 |
CN106550237B true CN106550237B (en) | 2020-05-19 |
Family
ID=58362322
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510586444.2A Active CN106550237B (en) | 2015-09-16 | 2015-09-16 | Monitoring video compression method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106550237B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107592541B (en) * | 2017-09-13 | 2022-03-04 | 中国科学院半导体研究所 | Image decompression method and system |
CN108063914B (en) * | 2017-11-22 | 2020-10-16 | 国政通科技股份有限公司 | Method and device for generating and playing monitoring video file and terminal equipment |
CN110113602A (en) * | 2019-04-22 | 2019-08-09 | 西安电子科技大学 | A kind of H.264 code rate control frame-skipping optimization method |
CN111586418A (en) * | 2020-05-09 | 2020-08-25 | 北京电信易通信息技术股份有限公司 | Video compression method and device |
CN113747154A (en) * | 2021-08-27 | 2021-12-03 | 杭州当虹科技股份有限公司 | Code rate control method and encoder |
WO2023076817A1 (en) * | 2021-10-25 | 2023-05-04 | Sensormatic Electronics, LLC | Hierarchical surveillance video compression repository |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101064849A (en) * | 2006-04-29 | 2007-10-31 | 鲁海宁 | Dynamic image coding method, apparatus and computer readable record medium |
CN101621689A (en) * | 2009-07-28 | 2010-01-06 | 天津大学 | MPEG-to-H.264/AVC video coding conversion system |
JP2010081227A (en) * | 2008-09-25 | 2010-04-08 | Toshiba Corp | Moving image decoder |
CN101729902A (en) * | 2008-10-15 | 2010-06-09 | 深圳市融创天下科技发展有限公司 | Video compression method |
CN103227963A (en) * | 2013-03-20 | 2013-07-31 | 西交利物浦大学 | Static surveillance video abstraction method based on video moving target detection and tracing |
CN103269436A (en) * | 2013-05-20 | 2013-08-28 | 山东大学 | Key frame selection method in 2D-3D video conversion |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070292108A1 (en) * | 2006-06-15 | 2007-12-20 | Thales Avionics, Inc. | Method and system for processing digital video |
-
2015
- 2015-09-16 CN CN201510586444.2A patent/CN106550237B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101064849A (en) * | 2006-04-29 | 2007-10-31 | 鲁海宁 | Dynamic image coding method, apparatus and computer readable record medium |
JP2010081227A (en) * | 2008-09-25 | 2010-04-08 | Toshiba Corp | Moving image decoder |
CN101729902A (en) * | 2008-10-15 | 2010-06-09 | 深圳市融创天下科技发展有限公司 | Video compression method |
CN101621689A (en) * | 2009-07-28 | 2010-01-06 | 天津大学 | MPEG-to-H.264/AVC video coding conversion system |
CN103227963A (en) * | 2013-03-20 | 2013-07-31 | 西交利物浦大学 | Static surveillance video abstraction method based on video moving target detection and tracing |
CN103269436A (en) * | 2013-05-20 | 2013-08-28 | 山东大学 | Key frame selection method in 2D-3D video conversion |
Non-Patent Citations (2)
Title |
---|
A novel method of background subtraction for indoor surveillance;Ganfeng Qiang ET AL;《2014 4th IEEE International Conference on Information Science and Technology》;20141013;全文 * |
An region of interest based video compression for indoor surveillance;Ganfeng Qiang ET AL;《Proceedings of 2nd International Conference on Information Technology and Electronic Commerce》;20150514;参见A背景提取小节、ROI压缩小节 * |
Also Published As
Publication number | Publication date |
---|---|
CN106550237A (en) | 2017-03-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106550237B (en) | Monitoring video compression method | |
CN103248893B (en) | From H.264/AVC standard to code-transferring method and transcoder thereof the fast frame of HEVC standard | |
KR101055738B1 (en) | Method and apparatus for encoding/decoding video signal using prediction information of intra-mode macro blocks of base layer | |
US6959044B1 (en) | Dynamic GOP system and method for digital video encoding | |
CN100499813C (en) | Advanced video coding (AVC) intra-frame prediction system and method | |
US9071841B2 (en) | Video transcoding with dynamically modifiable spatial resolution | |
US9521411B2 (en) | Method and apparatus for encoder assisted-frame rate up conversion (EA-FRUC) for video compression | |
JP5289440B2 (en) | Image encoding device, image decoding device, image encoding method, and image decoding method | |
CN101982977B (en) | For the method and apparatus carrying out data alignment with time domain border | |
US9055301B2 (en) | Changing motion estimation precision to maintain encoding within desired time | |
JP3609965B2 (en) | Method for processing digital video data in compressed format | |
US20060204115A1 (en) | Video encoding | |
JP2006519565A (en) | Video encoding | |
CN101087408A (en) | Image coding/recording apparatus and image coding/recording method | |
KR20050112130A (en) | Video transcoding | |
KR20060043051A (en) | Method for encoding and decoding video signal | |
US20220239911A1 (en) | Video Encoding and Decoding Methods and Apparatuses, Electronic Device, and Storage Medium | |
CN114913250A (en) | Video coding and decoding system and method based on semantic segmentation and Hungarian algorithm | |
CN115442618A (en) | Time domain-space domain self-adaptive video compression based on neural network | |
KR100883591B1 (en) | Method and apparatus for encoding/decoding video signal using prediction information of intra-mode macro blocks of base layer | |
Minallah et al. | Performance analysis of H. 265/HEVC (high-efficiency video coding) with reference to other codecs | |
CN113163199B (en) | H265-based video rapid prediction method, rapid coding method and system | |
KR20060043050A (en) | Method for encoding and decoding video signal | |
CN111212288B (en) | Video data encoding and decoding method and device, computer equipment and storage medium | |
CN110446042B (en) | Coding method for improving P frame quality in H.264 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |