CN103605159A - A method for parallel compression of massive seismic data - Google Patents
A method for parallel compression of massive seismic data Download PDFInfo
- Publication number
- CN103605159A CN103605159A CN201310570935.9A CN201310570935A CN103605159A CN 103605159 A CN103605159 A CN 103605159A CN 201310570935 A CN201310570935 A CN 201310570935A CN 103605159 A CN103605159 A CN 103605159A
- Authority
- CN
- China
- Prior art keywords
- geological data
- data
- block
- piecemeal
- geological
- 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.)
- Pending
Links
Images
Landscapes
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention relates to a method for parallel compression of massive seismic data and belongs to the technical field of petroleum seismic exploration data processing. The method for the parallel compression of the massive seismic data is characterized in that: seismic data cube data is obtained by means of carrying out conversion on three dimensional seismic data through the utilization of a CPU, and the seismic data cube data is subjected to decomposition and secondary decomposition to obtain seismic data sub-blocks and seismic data blocks; then the seismic data is transmitted to a GPU block by block; the parallel compression is carried out through employing a visual perception model, and fusion is carried out on compression results of all the seismic data blocks to obtain a compression result of all the seismic data. According to the method for the parallel compression of the massive seismic data of the invention, the compression quality is guaranteed, and at the same time, the compression speed of the seismic data is substantially raised, so that the parallel compression can be carried out on common computers through the common GPUs. The operation costs are saved. The usage is convenient and the economic benefits are good.
Description
Technical field:
The present invention relates to a kind of method of parallelly compressed mass seismic data, belong to oil seismic exploration technical field of data processing.
Background technology:
At geophysics field, seismic prospecting is one of method important in geophysical survey.It is with artificial method earthquake-wave-exciting, and with seismic prospecting instrument, the vibrations of the earth is recorded on tape, then by the data that computing machine obtains field, processes, thereby obtains information and the hydrocarbon information of relevant underground structure and lithology.As the primary link of petroleum prospecting, the data volume of seismic acquisition is very huge, and its quantity is conventionally all in TB.Geological data is the Large Volume Data of describing underground complex structure, development along with three-dimension high-resolution exploration, the survey line that seismic prospecting is required and road number are more and more, precision is also more and more higher, time sampling interval has reached 0.25 millisecond, and the data of some work areas field acquisition can reach tens trillion hundreds of trillion bytes even; In addition on the one hand, owing to adopting the multiple three-dimensional properties bodies such as coherent body, wave impedance, three winks and numerous in layer seismic properties and reservoir parameter, for Reservoir Description, its seismic interpretation data volume also will increase by 10 times more than on former basis.The record of geological data is used special-purpose tape conventionally, and the quantity of a complete needed data tape of seismic exploration is also very huge, and it is very inconvenient to transport.The transmission that artificial earth satellite is geological data provides a delivery means fast and safely, but the expense of leased satellite circuit is very expensive, and the communication resource taking is excessive, is not economic, practical transmission of seismic data means.
Because the geological data amount that exploration obtains is huge, the calculated amount producing when carrying out mass seismic data compression is also very huge, the compression of mass seismic data in the past all can only be by configuring powerful supercomputer or professional graphics workstation has gone, and supercomputer or professional graphics workstation cost are all more expensive, use also inconvenient, therefore, how huge geological data is carried out to effective, a high proportion of Lossless Compression and transmission is a problem demanding prompt solution.
Summary of the invention:
In order to overcome the deficiencies in the prior art, the object of the present invention is to provide a kind of method of parallelly compressed mass seismic data, 3D seismic data is converted to geological data volume data, and seismic data volume data are decomposed and decomposed, adopt human perceptual model to carry out parallelly compressed with secondary, greatly improved the compression speed of geological data, save operating cost, there is feature easy to use, good in economic efficiency.
The present invention realizes above-mentioned purpose by following technical solution.
The method of a kind of parallelly compressed mass seismic data provided by the present invention, comprises the steps:
1,3D seismic data is converted to geological data volume data:
By cpu central processing unit, 3D seismic data is changed, obtain geological data volume data;
Described changes 3D seismic data, is that the 3D seismic data that each sampled point is obtained converts a voxel to, and each voxel has the amplitude of the 3D seismic data that a corresponding sampled point obtains;
When a region is explored, select a plurality of places in this region to carry out data acquisition, each place is a seismic trace, by seismic prospecting instrument, in this place, carry out data acquisition, use the vibrations of recording the earth perpendicular to the prone a plurality of sampled points of level, and store on tape, each seismic trace is converted into a voxel road, and a plurality of voxels road is geological data volume data;
2, by seismic data volume data decomposition, be geological data sub-block:
According to the size of calculator memory, by seismic data volume data decomposition, be geological data sub-block, and read in calculator memory;
Described is geological data sub-block by seismic data volume data decomposition, that seismic data volume data acquisition is carried out to 2 * 2 * 2 decomposition by the structure of Octree, be divided into geological data sub-block, the corresponding corresponding geological data sub-block of each node that makes Octree, each earthquake data sub-block deposits in disk standby;
When earthquake data sub-block belongs to following 3 kinds of situations, earthquake data sub-block is decomposed again by octree structure;
1. in geological data sub-block, the consistance of voxel is less than the determined critical value of user;
2. the numerical value of geological data sub-block is greater than the numerical value of calculator memory;
3. the numerical value of geological data sub-block is greater than the numerical value of the determined minimum earthquake data sub-block of user;
In described geological data sub-block, the consistance of voxel refers to the similarity of voxel in geological data sub-block;
3, geological data sub-block is decomposed into geological data piecemeal:
According to the size of computing machine video memory, geological data sub-block is decomposed into geological data piecemeal, and is read in computing machine video memory;
Described is decomposed into geological data piecemeal by geological data sub-block, that geological data sub-block data acquisition is carried out to 2 * 2 * 2 decomposition by the structure of Octree, be divided into geological data piecemeal, make the corresponding corresponding geological data piecemeal of each node of Octree, to obtain less data block, and read in computing machine video memory;
When earthquake deblocking belongs to following 3 kinds of situations, earthquake deblocking is decomposed again by octree structure;
1. in geological data piecemeal, the consistance of voxel is less than the determined critical value of user;
2. the numerical value of geological data piecemeal is greater than the numerical value of computing machine video memory;
3. the numerical value of geological data piecemeal is greater than the numerical value of the determined minimum earthquake deblocking of user;
In described geological data piecemeal, the consistance of voxel refers to the similarity of voxel in geological data piecemeal;
4, adopt human perceptual model to process geological data piecemeal:
In video memory, adopt human perceptual model, geological data piecemeal is decomposed into the part of strong limit, details and level and smooth three kinds of different qualities;
Described human perceptual model is to be the strong limit part of seismic section according to the main source that forms the stimulation of human vision;
Described strong limit refers to that larger brightness changes and the hatch region of less border width; Corresponding brightness change intensity is less, and the unconspicuous region of variational regularity is detail section; And smooth is the region that brightness slowly changes;
Therefore when setting up human perceptual model, need to take different vision weights according to the heterogeneity of seismic section;
The described part of strong limit, details and level and smooth three kinds of different qualities that geological data piecemeal is decomposed into is that geological data piecemeal is divided into strong limit piece, detailed block, smooth block 3 classes;
Its assorting process is: entropy and the variance of calculating each geological data piecemeal; The geological data piecemeal that entropy is less is smooth block; The geological data piecemeal that entropy is larger is detailed block or strong limit piece; The geological data piecemeal that variance is less is detailed block, and the geological data piecemeal that variance is larger is strong limit piece;
5, the part of three kinds of different qualities of earthquake deblocking is compressed to processing:
GPU graphic process unit is partly carried out three kinds of different qualities of geological data piecemeal parallelly compressed, obtains the compression result of geological data piecemeal;
Described parallelly compressed 3 parts that comprise, (1) carries out Fast Lifting Wavelet Transform to earthquake deblocking; (2), according to human-eye visual characteristic, to three of earthquake deblocking kinds of different qualities, partly give different vision weights; (3) utilize spiht algorithm to carry out compressed encoding, obtain the compression result of geological data piecemeal;
5.1 Fast Lifting Wavelet Transforms are by division, predict and upgrade three steps and realize the separated of HFS in signal and low frequency part;
5.2 vision weights are by give the method for different vision weights (in sensitizing range, wavelet coefficient is given large vision weights) to three kinds of corresponding wavelet coefficients of different qualities part, guarantee the visual most important coefficient of prioritised transmission, to further improve geological data recovering quality;
5.3 compression algorithm, spiht algorithm is by significant coefficient being sorted and generating significant bits, according to the significant bits generating, progressively approach each wavelet coefficient, spiht algorithm usage space orientation tree is organized wavelet coefficient, and direction in space tree refers to a kind of spatial relationship in hierarchy; Spiht algorithm adopts three order chained lists to carry out the coding of coefficient: inessential subset table LIS, inessential coefficient table LIP and significant coefficient table LSP;
6, by video memory, pass the compression result of geological data piecemeal back internal memory, and be stored in hard disc of computer;
7, repeat above step 4-6, all geological data piecemeals are carried out parallelly compressed, and compression result is stored in hard disc of computer;
8, the compression result of all geological data piecemeals is merged, obtain the compression result of whole geological data.
The present invention compared with prior art, by using CPU that 3D seismic data is converted to geological data volume data, and seismic data volume data are decomposed and secondary decomposition, obtain geological data sub-block and geological data piecemeal, again geological data piecemeal is passed and gives GPU, adopt human perceptual model to carry out parallelly compressed, can be when guaranteeing compression quality, greatly improve the compression speed of geological data, make also can carry out the parallelly compressed of geological data by common GPU in common computer, saved operating cost, easy to use, good in economic efficiency.
Accompanying drawing explanation:
Fig. 1 is the schematic flow sheet of the method for a kind of parallelly compressed mass seismic data of the present invention.
Fig. 2 is the schematic diagram that geological data of the present invention is transformed into geological data volume data.
Embodiment:
Below in conjunction with drawings and Examples, the invention will be further described.
Embodiment:
In recent years, GPU (Graphics Processing Unit, graphic process unit) technology has obtained development at full speed, and as a kind of dedicated graphics rendering hardware for PC, GPU has great meaning to the development of parallel processing technique.GPU is exactly speed with respect to the main advantage of CPU, and the speed advantage of GPU is mainly derived from its unique hardware systems design.The method of a kind of parallelly compressed mass seismic data that the present embodiment provides, can make CPU and GPU parallel processing 3D seismic data, thereby reaches the object of using common PC can carry out the parallelly compressed processing of geological data.
The method of a kind of parallelly compressed mass seismic data provided by the present invention, comprises the steps:
1,3D seismic data is converted to geological data volume data:
By cpu central processing unit, 3D seismic data is changed, obtain geological data volume data;
Parallelly compressed for mass seismic data is carried out, the 3D seismic data that each sampled point need to be obtained convert a voxel to, and each voxel has the amplitude of the 3D seismic data that a corresponding sampled point obtains;
When a region is explored, should select a plurality of places in this region to carry out data acquisition, each place is a seismic trace, by seismic prospecting instrument, in this place, carry out data acquisition, use the vibrations of recording the earth perpendicular to the prone a plurality of sampled points of level, and store on tape, each seismic trace is converted into a voxel road, and a plurality of voxels road is geological data volume data.
2, by seismic data volume data decomposition, be geological data sub-block:
According to the size of calculator memory, by seismic data volume data decomposition, be geological data sub-block, and read in calculator memory;
Described is geological data sub-block by seismic data volume data decomposition, that seismic data volume data acquisition is carried out to 2 * 2 * 2 decomposition by the structure of Octree, be divided into geological data sub-block, the corresponding corresponding geological data sub-block of each node that makes Octree, each earthquake data sub-block deposits in disk standby;
When earthquake data sub-block belongs to following 3 kinds of situations, earthquake data sub-block is decomposed again by octree structure;
1. in geological data sub-block, the consistance of voxel is less than the determined critical value of user;
2. the numerical value of geological data sub-block is greater than the numerical value of calculator memory; In the present embodiment, in computing machine, save as 4GB; Design parameter is: Samsung DDR3,1333MHz, 4GB;
3. the numerical value of geological data sub-block is greater than the numerical value of the determined minimum earthquake data sub-block of user;
In described geological data sub-block, the consistance of voxel refers to the similarity of voxel in geological data sub-block.
3, geological data sub-block is decomposed into geological data piecemeal:
According to the size of computing machine video memory, geological data sub-block is decomposed into geological data piecemeal, and is read in computing machine video memory;
Described is decomposed into geological data piecemeal by geological data sub-block, that geological data sub-block data acquisition is carried out to 2 * 2 * 2 decomposition by the structure of Octree, be divided into geological data piecemeal, make the corresponding corresponding geological data piecemeal of each node of Octree, to obtain less data block, and read in computing machine video memory;
When earthquake deblocking belongs to following 3 kinds of situations, earthquake deblocking is decomposed again by octree structure;
1. in geological data piecemeal, the consistance of voxel is less than the determined critical value of user;
2. the numerical value of geological data piecemeal is greater than the numerical value of computing machine video memory; In the present embodiment, the video memory of computer display card is 1.5GB; Design parameter is: Nvidia GeForce GTX560M, 1536MB;
3. the numerical value of geological data piecemeal is greater than the numerical value of the determined minimum earthquake deblocking of user;
In described geological data piecemeal, the consistance of voxel refers to the similarity of voxel in geological data piecemeal.
4, adopt human perceptual model to process geological data piecemeal:
In video memory, adopt human perceptual model, geological data piecemeal is decomposed into the part of strong limit, details and level and smooth three kinds of different qualities;
Described human perceptual model is to be the strong limit part of seismic section according to the main source that forms the stimulation of human vision;
Described strong limit refers to that larger brightness changes and the hatch region of less border width; Corresponding brightness change intensity is less, the unconspicuous region of variational regularity is detail section; And smooth is the region that brightness slowly changes;
Therefore when setting up human perceptual model, need to take different vision weights according to the heterogeneity of seismic section;
According to certain classifying rules, geological data piecemeal is decomposed into the part of strong limit, details and level and smooth three kinds of different qualities, be about to geological data piecemeal and be divided into strong limit piece, detailed block, smooth block 3 classes;
Its assorting process is: entropy and the variance of calculating each geological data piecemeal; The geological data piecemeal that entropy is less is smooth block; The geological data piecemeal that entropy is larger is detailed block or strong limit piece; The geological data piecemeal that variance is less is detailed block, and the geological data piecemeal that variance is larger is strong limit piece.
5, the part of three kinds of different qualities of earthquake deblocking is compressed to processing:
To the stream handle quantity in the GPU of the data based video card in video card, it is carried out to adaptive decomposition, in the video card GTX560M selecting in the present embodiment, comprise 192 stream handles, therefore the data in video memory can be divided into 192 parts, distribute to 192 stream handles and do parallel computation;
The function moving on GPU is called core (Kernel), it is characterized in that operating all elements on a plurality of stream and is not only and operates independently element, and executive routine is when calling core program with asynchronous system.Serial part in executive routine is carried out on CPU, core is only carried out as parallel section on GPU, now, the procedure division of moving on GPU is become grid (Grid), piece (Block), three thread ranks of thread (Thread) implement respectively; According to the framework of GPU, 3D seismic data is done to adaptive decomposition, wherein, earthquake block data is corresponding with the grid (Grid) in GPU, cross-sectional data in piecemeal geological data is corresponding with the piece (Block) in GPU, and the track data in cross-sectional data is corresponding with the thread (Thread) in GPU;
GPU graphic process unit is partly carried out three kinds of different qualities of geological data piecemeal parallelly compressed, obtains the compression result of geological data piecemeal;
Described parallelly compressed 3 parts that comprise, (1) carries out Fast Lifting Wavelet Transform to earthquake deblocking; (2), according to human-eye visual characteristic, to three of earthquake deblocking kinds of different qualities, partly give different vision weights; (3) utilize spiht algorithm to carry out compressed encoding, obtain the compression result of geological data piecemeal;
5.1 Fast Lifting Wavelet Transforms:
Be mainly by division, predict and upgrade three steps and realize the separated of HFS in signal and low frequency part;
(1) division: by original signal s[n] be split into two less subsets, the simplest method is to adopt odd even division, obtains even number part s
even[n] comprises s[n] in the value of all even items, odd number part s
odd[n] comprises s[n] in all odd terms,
s
even[n]=s[2n]s
odd[n]=s[2n+1]
(2) prediction: on the basis of raw data correlativity, with the value s of even number part
even[n] predicts odd number part s
odd[n], s[n] odd and even number partly there is certain correlativity, so odd number part can obtain according near even number partial data prediction, actual value and the predicted value of odd number part are subtracted each other the radio-frequency component that obtains signal; Press following formula definition predictive operator:
d[n]=s
odd[n]-P(s
even[n])
D[n] be the radio-frequency component of signal wavelet transformation, i.e. the detail section of signal;
(3) upgrade: in order to make some global characteristics of original signal collection at subset s
eventhe continuation of insurance of [n] relaying is held, and need to upgrade.Introduce and upgrade operator U, acted on d[n] upper, generate a better subset x[n], renewal process is as follows:
x[n]=s
even[n]+U(d[n])
X[n] be the low frequency part after signal wavelet transformation, the i.e. profile of signal.For x[n] repeat above-mentioned steps and can realize lifting wavelet transform;
5.2 vision weights:
By give the method for different vision weights (in sensitizing range, wavelet coefficient is given large vision weights) to three kinds of corresponding wavelet coefficients of different qualities part, guarantee the visual most important coefficient of prioritised transmission, to further improve geological data recovering quality.Because choosing with wavelet decomposition yardstick of the vision weights of wavelet coefficient in high-frequency sub-band under the thickest yardstick is closely related, thus the feature of successively decreasing according to wavelet coefficient energy, by thick yardstick to the thin yardstick vision weights that successively decrease.The present embodiment is chosen for 4/5 by the decrement factor of vision weights, and the initial visual weights of detailed block, smooth block, strong limit piece are made as 0.8,1.0,1.2;
5.3 compression algorithms:
The basic thought of spiht algorithm is by significant coefficient being sorted and generating significant bits, according to the significant bits generating, progressively approach each wavelet coefficient, spiht algorithm usage space orientation tree is organized wavelet coefficient, and direction in space tree refers to a kind of spatial relationship in hierarchy.Spiht algorithm adopts three order chained lists to carry out the coding of coefficient: inessential subset table LIS, inessential coefficient table LIP and significant coefficient table LSP; The key step of SPITH coding:
(1) initialization of threshold value and ordered list;
(2) sequential scanning: the object of sequential scanning is the significant coefficient of coding present bit plane.By following two large steps, formed: 1) all wavelet coefficients in sequential scanning LIP; 2) to each the list item processed in sequence in LIS;
(3) fine scanning: for each list item (i, j) in LSP, if (i, j) is new interpolation, not the scanning process of carrying out just, the n of output (i, j) most important;
(4) carry out sequential scanning next time and fine scanning.
6, by video memory, pass the compression result of geological data piecemeal back internal memory, and be stored in hard disc of computer.
7, repeat above step 4-6, all geological data piecemeals are carried out parallelly compressed, and compression result is stored in hard disc of computer.
8, the compression result of all geological data piecemeals is merged, obtain the compression result of whole geological data.
The method of a kind of parallelly compressed mass seismic data provided by the present invention, by using CPU that 3D seismic data is converted to geological data volume data, and seismic data volume data are decomposed and secondary decomposition, obtain geological data sub-block and geological data piecemeal, again geological data piecemeal is passed and gives GPU, in conjunction with human perceptual model, carry out parallelly compressed, can guarantee under the prerequisite of compression quality, greatly improve the compression speed of mass seismic data, make also can carry out by common GPU the Fast Compression of mass seismic data in common computer, easy to use, and cost-saving.
Claims (6)
1. a method for parallelly compressed mass seismic data, is characterized in that comprising the steps:
(1), 3D seismic data is converted to geological data volume data:
By cpu central processing unit, 3D seismic data is changed, obtain geological data volume data;
(2) by seismic data volume data decomposition, be, geological data sub-block:
According to the size of calculator memory, by seismic data volume data decomposition, be geological data sub-block, and read in calculator memory;
(3), geological data sub-block is decomposed into geological data piecemeal:
According to the size of computing machine video memory, geological data sub-block is decomposed into geological data piecemeal, and is read in computing machine video memory;
(4), adopt human perceptual model to process geological data piecemeal:
In video memory, adopt human perceptual model, geological data piecemeal is decomposed into the part of strong limit, details and level and smooth three kinds of different qualities; Described human perceptual model is to be the strong limit part of seismic section according to the main source that forms the stimulation of human vision; Described strong limit refers to that larger brightness changes and the hatch region of less border width; Corresponding brightness change intensity is less, and the unconspicuous region of variational regularity is detail section; And smooth is the region that brightness slowly changes;
(5), the part of three kinds of different qualities of earthquake deblocking is compressed to processing:
GPU graphic process unit is partly carried out three kinds of different qualities of geological data piecemeal parallelly compressed, obtains the compression result of geological data piecemeal;
Described parallelly compressed 3 parts that comprise, (1) carries out Fast Lifting Wavelet Transform to earthquake deblocking; (2), according to human-eye visual characteristic, to three of earthquake deblocking kinds of different qualities, partly give different vision weights; (3) utilize spiht algorithm to carry out compressed encoding, obtain the compression result of geological data piecemeal;
(6), by video memory, pass the compression result of geological data piecemeal back internal memory, and be stored in hard disc of computer;
(7), repeat above step 4-6, all geological data piecemeals are carried out parallelly compressed, and compression result is stored in hard disc of computer;
(8), the compression result of all geological data piecemeals is merged, obtain the compression result of whole geological data.
2. the method for a kind of parallelly compressed mass seismic data according to claim 1, it is characterized in that described 3D seismic data is changed, be that the 3D seismic data that each sampled point is obtained converts a voxel to, each voxel has the amplitude of the 3D seismic data that a corresponding sampled point obtains; When a region is explored, select a plurality of places in this region to carry out data acquisition, each place is a seismic trace, by seismic prospecting instrument, in this place, carry out data acquisition, use the vibrations of recording the earth perpendicular to the prone a plurality of sampled points of level, and store on tape, each seismic trace is converted into a voxel road, and a plurality of voxels road is geological data volume data.
3. the method for a kind of parallelly compressed mass seismic data according to claim 1, described in it is characterized in that is geological data sub-block by seismic data volume data decomposition, that seismic data volume data acquisition is carried out to 2 * 2 * 2 decomposition by the structure of Octree, be divided into geological data sub-block, the corresponding corresponding geological data sub-block of each node that makes Octree, each earthquake data sub-block deposits in disk standby;
When earthquake data sub-block belongs to following 3 kinds of situations, earthquake data sub-block is decomposed again by octree structure;
1. in geological data sub-block, the consistance of voxel is less than the determined critical value of user;
2. the numerical value of geological data sub-block is greater than the numerical value of calculator memory;
3. the numerical value of geological data sub-block is greater than the numerical value of the determined minimum earthquake data sub-block of user;
In described geological data sub-block, the consistance of voxel refers to the similarity of voxel in geological data sub-block.
4. the method for a kind of parallelly compressed mass seismic data according to claim 1, it is characterized in that described geological data sub-block being decomposed into geological data piecemeal, that geological data sub-block data acquisition is carried out to 2 * 2 * 2 decomposition by the structure of Octree, be divided into geological data piecemeal, make the corresponding corresponding geological data piecemeal of each node of Octree, and read in computing machine video memory;
When earthquake deblocking belongs to following 3 kinds of situations, earthquake deblocking is decomposed again by octree structure;
1. in geological data piecemeal, the consistance of voxel is less than the determined critical value of user;
2. the numerical value of geological data piecemeal is greater than the numerical value of computing machine video memory;
3. the numerical value of geological data piecemeal is greater than the numerical value of the determined minimum earthquake deblocking of user;
In described geological data piecemeal, the consistance of voxel refers to the similarity of voxel in geological data piecemeal.
5. the method for a kind of parallelly compressed mass seismic data according to claim 1, it is characterized in that the described part of strong limit, details and level and smooth three kinds of different qualities that geological data piecemeal is decomposed into, is that geological data piecemeal is divided into strong limit piece, detailed block, smooth block 3 classes; Its assorting process is: entropy and the variance of calculating each geological data piecemeal; The geological data piecemeal that entropy is less is smooth block; The geological data piecemeal that entropy is larger is detailed block or strong limit piece; The geological data piecemeal that variance is less is detailed block, and the geological data piecemeal that variance is larger is strong limit piece.
6. the method for a kind of parallelly compressed mass seismic data according to claim 1, is characterized in that described Fast Lifting Wavelet Transform is by division, predicts and upgrade three steps and realize the separated of HFS in signal and low frequency part; Described vision weights are by three kinds of methods that the corresponding wavelet coefficient of different qualities part is given different vision weights, guarantee the visual most important coefficient of prioritised transmission, to further improve geological data recovering quality; Described spiht algorithm is by significant coefficient being sorted and generating significant bits, according to the significant bits generating, progressively approach each wavelet coefficient, spiht algorithm usage space orientation tree is organized wavelet coefficient, and direction in space tree refers to a kind of spatial relationship in hierarchy; Spiht algorithm adopts three order chained lists to carry out the coding of coefficient: inessential subset table LIS, inessential coefficient table LIP and significant coefficient table LSP.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310570935.9A CN103605159A (en) | 2013-11-13 | 2013-11-13 | A method for parallel compression of massive seismic data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310570935.9A CN103605159A (en) | 2013-11-13 | 2013-11-13 | A method for parallel compression of massive seismic data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103605159A true CN103605159A (en) | 2014-02-26 |
Family
ID=50123399
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310570935.9A Pending CN103605159A (en) | 2013-11-13 | 2013-11-13 | A method for parallel compression of massive seismic data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103605159A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104166160A (en) * | 2014-07-25 | 2014-11-26 | 中国石油天然气集团公司 | Earthquake data combination method, apparatus and system |
| CN104202124A (en) * | 2014-09-12 | 2014-12-10 | 成都博智维讯信息技术有限公司 | Erp (enterprise resource planning) data package communication method |
| CN104816742A (en) * | 2015-05-13 | 2015-08-05 | 张凤亮 | Train operation disaster early warning and monitoring system |
| CN105093283A (en) * | 2014-05-16 | 2015-11-25 | 中国石油化工股份有限公司 | Three-dimensional observation system surface element attribute multi-thread rapid display method |
| CN105512120A (en) * | 2014-09-22 | 2016-04-20 | 中国石油化工股份有限公司 | Earthquake data compression method, compression storage structure and random access method |
| CN107333136A (en) * | 2017-06-26 | 2017-11-07 | 西安万像电子科技有限公司 | Method for encoding images and device |
| CN110555043A (en) * | 2018-03-30 | 2019-12-10 | 中国石油化工股份有限公司 | Seismic data multi-level caching method based on octree |
| CN110971340A (en) * | 2019-10-17 | 2020-04-07 | 北京邮电大学 | Disaster information coding and decoding method and device based on tree structure |
| CN111290697A (en) * | 2018-12-07 | 2020-06-16 | 上海寒武纪信息科技有限公司 | Data compression method, encoding circuit and arithmetic device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5754793A (en) * | 1995-04-19 | 1998-05-19 | Samsung Electronics Co., Ltd. | Wavelet image compression/recovery apparatus and method using human visual system modeling |
| CN101783939A (en) * | 2009-01-16 | 2010-07-21 | 复旦大学 | Picture coding method based on human eye visual characteristic |
| CN103308942A (en) * | 2012-03-12 | 2013-09-18 | 中国石油天然气股份有限公司 | Method and system for visualizing seismic data |
-
2013
- 2013-11-13 CN CN201310570935.9A patent/CN103605159A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5754793A (en) * | 1995-04-19 | 1998-05-19 | Samsung Electronics Co., Ltd. | Wavelet image compression/recovery apparatus and method using human visual system modeling |
| CN101783939A (en) * | 2009-01-16 | 2010-07-21 | 复旦大学 | Picture coding method based on human eye visual characteristic |
| CN103308942A (en) * | 2012-03-12 | 2013-09-18 | 中国石油天然气股份有限公司 | Method and system for visualizing seismic data |
Non-Patent Citations (3)
| Title |
|---|
| ANTHONY VASSILIOU ET AL.: "《Comparison of Wavelet Image Coding Schemes for Seismic Data Compression》", 《SPIE PROCEEDINGS》 * |
| 余平等: "《基于提升小波的地震数据压缩编码算法》", 《苏州大学学报（工科版）》 * |
| 谢凯等: "《一种基于人类视觉特性的医学图像压缩算法》", 《上海交通大学学报》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105093283A (en) * | 2014-05-16 | 2015-11-25 | 中国石油化工股份有限公司 | Three-dimensional observation system surface element attribute multi-thread rapid display method |
| CN105093283B (en) * | 2014-05-16 | 2017-12-01 | 中国石油化工股份有限公司 | A kind of stereo observing system bin attributes multithreading fast display method |
| CN104166160A (en) * | 2014-07-25 | 2014-11-26 | 中国石油天然气集团公司 | Earthquake data combination method, apparatus and system |
| CN104202124A (en) * | 2014-09-12 | 2014-12-10 | 成都博智维讯信息技术有限公司 | Erp (enterprise resource planning) data package communication method |
| CN104202124B (en) * | 2014-09-12 | 2017-10-10 | 成都博智维讯信息技术有限公司 | A kind of erp packet communications method |
| CN105512120B (en) * | 2014-09-22 | 2019-06-04 | 中国石油化工股份有限公司 | A kind of seismic data compression method, compression and storage method and method of random access |
| CN105512120A (en) * | 2014-09-22 | 2016-04-20 | 中国石油化工股份有限公司 | Earthquake data compression method, compression storage structure and random access method |
| CN104816742A (en) * | 2015-05-13 | 2015-08-05 | 张凤亮 | Train operation disaster early warning and monitoring system |
| CN107333136A (en) * | 2017-06-26 | 2017-11-07 | 西安万像电子科技有限公司 | Method for encoding images and device |
| CN110555043A (en) * | 2018-03-30 | 2019-12-10 | 中国石油化工股份有限公司 | Seismic data multi-level caching method based on octree |
| CN111290697A (en) * | 2018-12-07 | 2020-06-16 | 上海寒武纪信息科技有限公司 | Data compression method, encoding circuit and arithmetic device |
| CN111290697B (en) * | 2018-12-07 | 2022-01-28 | 上海寒武纪信息科技有限公司 | Data compression method, encoding circuit and arithmetic device |
| CN110971340A (en) * | 2019-10-17 | 2020-04-07 | 北京邮电大学 | Disaster information coding and decoding method and device based on tree structure |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103605159A (en) | A method for parallel compression of massive seismic data | |
| Jafarpour | Wavelet reconstruction of geologic facies from nonlinear dynamic flow measurements | |
| EP2869096A2 (en) | Systems and methods of multi-scale meshing for geologic time modeling | |
| US11009617B2 (en) | Method for fast calculation of seismic attributes using artificial intelligence | |
| CN106249299A (en) | The lower thin layer weak reflected seismic energy restoration methods of strong reflection shielding and device | |
| CN103308942A (en) | Method and system for visualizing seismic data | |
| Ye et al. | In situ generated probability distribution functions for interactive post hoc visualization and analysis | |
| CN102576370A (en) | System and method for lacunarity analysis | |
| Usher et al. | Interactive visualization of terascale data in the browser: Fact or fiction? | |
| CN102855644B (en) | Large-scale topographic data compression and decompression method based on wavelet self-feedback | |
| CN101876714A (en) | Method and device for acquiring three-dimensional imaging data of geologic structure | |
| CN104656132B (en) | A kind of method for improving multi-faceted stratigraphic dip information storage efficiency | |
| Franklin et al. | Smugglers and border guards: the geostar project at RPI | |
| Peyrot et al. | HexaShrink, an exact scalable framework for hexahedral meshes with attributes and discontinuities: multiresolution rendering and storage of geoscience models | |
| CN1956009A (en) | Presenting method of large scale three-D data set | |
| Liu et al. | Simulation of complex geological architectures based on multi-stage generative adversarial networks integrating with attention mechanism and spectral normalization | |
| Feng et al. | Connect the Dots: In Situ 4-D Seismic Monitoring of CO 2 Storage With Spatio-Temporal CNNs | |
| CN109782344A (en) | Depositional sequence Boundary Recognition method and device | |
| Dubois et al. | Highly Efficient Controlled Hierarchical Data Reduction techniques for Interactive Visualization of Massive Simulation Data. | |
| CN115880455A (en) | Three-dimensional intelligent interpolation method based on deep learning | |
| Tian et al. | Dynam: Dynamic multiresolution data representation for large-scale scientific analysis | |
| Shi et al. | Real-time switching and visualization of logging attributes based on subspace learning | |
| Sebai et al. | Seismic data compression: an overview | |
| Clyne | The multiresolution toolkit: Progressive access for regular gridded data | |
| Soler | Large data reduction and structure comparison with topological data analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140226 |