CN101056411B - Method for detecting the image displacement - Google Patents
Method for detecting the image displacement Download PDFInfo
- Publication number
- CN101056411B CN101056411B CN 200610075393 CN200610075393A CN101056411B CN 101056411 B CN101056411 B CN 101056411B CN 200610075393 CN200610075393 CN 200610075393 CN 200610075393 A CN200610075393 A CN 200610075393A CN 101056411 B CN101056411 B CN 101056411B
- Authority
- CN
- China
- Prior art keywords
- block
- view data
- benchmark
- image
- draw
- 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
- Image Analysis (AREA)
Abstract
An image shift detection method, comprises following steps: (a) acquiring a first image data in a first time point; (b) acquiring a second image data in a second time point; (c) comparing the image characteristics of a first region block in the first image data with image characteristics of a second region block in the second image data; (d) setting the first region block or the second region block as a reference region block according to a compared result in the step (c); (e) comparing image data, which is not corresponding to the reference region block, in the first image data or the second image data with that in the reference region block; and (f) calculating a shifted vector for the first image data moving to the second image data based on a compared result of the step (e).
Description
Technical field
The present invention provides a kind of image processing techniques, refers to a kind of method of detected image displacement especially.
Background technology
The process of many Video processing, for example the displacement pre-estimating technology (motionestimation) among the MPEG2/MPEG4 all needs to utilize the operation result of image block comparison (block matching), and tradition is to adopt typical block alignment algorithm (Block Matching Algorithm) to be used as the account form that displacement is estimated.See also Fig. 1, the schematic diagram of second view data 12 that Fig. 1 is obtained for first view data 10 before obtained in very first time point and second time point greater than this very first time point, it is benchmark block (Reference Block) 14 that a block less than first view data 10 is chosen in the center that at first can be fixed on first view data 10, again 14 pairs second view data 12 of benchmark block search that compares according to this.See also Fig. 2, the schematic diagram of Fig. 2 for before utilizing 14 pairs second view data 12 of benchmark block to compare and search, its comparison mode is the block that second view data 12 is divided into a piece and benchmark block 14 identical sizes, and the block group 16 that is referred to as to take a sample (Sample Blocks), through in sampling block group 16, finding out the sampling block the most similar after the certain operations, use the motion-vector that calculates image again to benchmark block 14.For instance, as shown in Figure 2, one 3 * 3 square blocks are got in center in 9 * 9 square first view data 10, it is meaning benchmark block 14, just begin by the upper left corner of second view data 12 afterwards, take out one 3 * 3 block in the same way, as sample block 161, the pixel that then moves right takes out another block of 3 * 3 again, as sample block 162, so in the whole second complete view data 12 from left to right, from top to bottom, can take out (9-3+1) * (9-3+1) totally 49 sample blocks, this is region-wide search method (full search).Through drawing a sample block the most similar to benchmark block 14 after these the 49 times comparisons, this is the block alignment algorithm, is sample block 1615 to the most similar sample block of benchmark block 14 in this example.See also Fig. 3, Fig. 3 is that previous Fig. 2 is through comparing the schematic diagram of the motion vector that is drawn after the search, can calculate the difference of the coordinate position of the coordinate position of sample block 1615 and benchmark block 14, and then draw the motion vector that first view data 10 moves to second view data 12.
In sum, before be that mode with region-wide search method collocation block alignment algorithm reaches the purpose that displacement is estimated, yet when the benchmark block 14 of the selected taking-up in center of first view data 10 does not have enough available or effective characteristics of image (image feature), make comparison result not exclusive easily, and increase the probability of estimating the motion-vector that makes mistake, and reduce image identification rate and correctness.
Summary of the invention
The present invention provides a kind of method of detected image displacement, to solve the above problems.
One aspect of the present invention is the method that discloses a kind of detected image displacement, and it includes the following step: (a) obtain first view data in very first time point; (b) obtain second view data in second time point; (c) compare in effective characteristics of image quantity of first block in this first view data and this second view data effective characteristics of image quantity with respect to second block of this first block position; (d) according to the comparison result of step (c), when effective characteristics of image quantity of this first block in this first view data greater than this second view data in during with respect to the effective characteristics of image quantity of this second block of this first block position, setting this first block is the benchmark block; When effective characteristics of image quantity of this first block in this first view data less than this second view data in during with respect to the effective characteristics of image quantity of this second block of this first block position, setting this second block is the benchmark block; (e) when this first block of setting is the benchmark block, compare this benchmark block and this second view data, to draw the image block that corresponds to this benchmark block in this second view data; (f) comparison result of foundation (e), calculate the difference of the coordinate position of the coordinate position of the image block that corresponds to this benchmark block in this second view data and this benchmark block, to draw the motion vector that this first view data moves to this second view data; (g) when this second block of setting is the benchmark block, compare this benchmark block and this first view data, to draw the image block that corresponds to this benchmark block in this first view data; (h) comparison result of foundation (g), calculate the difference of the coordinate position of the image block that corresponds to this benchmark block in the coordinate position of this benchmark block and this first view data, to draw the motion vector that this first view data moves to this second view data.
Description of drawings
The schematic diagram of second view data that Fig. 1 is obtained for first view data before obtained in very first time point and second time point.
The schematic diagram of Fig. 2 for before utilizing the benchmark block that second view data is compared and searches.
Fig. 3 is that previous Fig. 2 is through comparing the schematic diagram of the motion vector that is drawn after the search.
Fig. 4 is the flow chart of detected image displacement of the present invention.
The schematic diagram of second view data that first view data that Fig. 5 is obtained in very first time point for first embodiment of the invention and second time point are obtained.
The schematic diagram that Fig. 6 utilizes the benchmark block that second view data is compared and searches for first embodiment of the invention.
Fig. 7 is the schematic diagram of the motion vector that Fig. 6 is drawn after comparison is searched in the first embodiment of the invention.
The schematic diagram of second view data that first view data that Fig. 8 is obtained in very first time point for second embodiment of the invention and second time point are obtained.
The schematic diagram that Fig. 9 utilizes the benchmark block that first view data is compared and searches for second embodiment of the invention.
Figure 10 is the schematic diagram of the motion vector that Fig. 9 is drawn after comparison is searched in the second embodiment of the invention.
[main element label declaration]
10 first view data, 12 second view data
14 benchmark blocks, 16 sampling block groups
161-1649 sample block
50 first view data, 52 second view data
54 first blocks, 56 second blocks
58 sampling block group 581-5849 sample blocks
70 first view data, 72 second view data
74 first blocks, 76 second blocks
78 sampling block group 781-7849 sample blocks
Step 100,102,104,106,108,110,112,114,116,118,120,122,124
Embodiment
See also Fig. 4, Fig. 4 is the flow chart of detected image displacement of the present invention, and the present invention comprises the following step:
Step 100: beginning.
Step 102: obtain first view data in very first time point.
Step 104: obtain second view data in second time point.
Step 106: choosing a block less than this first view data in the center of this first view data is first block.
Step 108: choosing a block less than this second view data in the center (with respect to this first block position) of this second view data is second block.
Step 110: the characteristics of image of comparing this second block in the characteristics of image (image feature) of this first block in this first view data and this second view data, when effective characteristics of image quantity of this first block during greater than the effective characteristics of image quantity of this second block, execution in step 112; When effective characteristics of image quantity of this first block during less than the effective characteristics of image quantity of this second block, execution in step 118.
Step 112: setting this first block is benchmark block (Reference Block).
Step 114: compare this benchmark block and this second view data, to draw the image block that corresponds to this benchmark block in this second view data.
Step 116: calculate the difference of the coordinate position of the coordinate position of the image block that corresponds to this benchmark block in this second view data and this benchmark block, to draw the motion vector that this first view data moves to this second view data.
Step 118: setting this second block is the benchmark block.
Step 120: compare this benchmark block and this first view data, to draw the image block that corresponds to this benchmark block in this first view data.
Step 122: calculate the difference of the coordinate position of the image block that corresponds to this benchmark block in the coordinate position of this benchmark block and this first view data, to draw the motion vector that this first view data moves to this second view data.
Step 124: finish.
In this above-mentioned steps is described in detail, see also Fig. 5, the schematic diagram of second view data 52 that first view data 50 that Fig. 5 is obtained in very first time point for first embodiment of the invention and second time point greater than this very first time point are obtained, at first the block that can choose in first view data 50 less than first view data 50 is first block 54, and wherein first block 54 can be positioned at the center of first view data 50; In like manner, the block that can choose in second view data 52 less than second view data 52 is second block 56, and wherein second block 56 can be positioned at the center of second view data 52.Afterwards, can compare the characteristics of image of first block 54 in first view data 50 and the characteristics of image of second block 56 in second view data 52, for example can compare available or effective characteristics of image quantity of first block 54 and available or effective characteristics of image quantity of second block 56, in first embodiment of Fig. 5, available or effective characteristics of image quantity of first block 54 is the available or effective characteristics of image quantity greater than second block 56, so in time, can be set first block 54 and is the benchmark block, again according to this benchmark block to the search that compares of second view data 52, wherein this comparison method for searching is to use region-wide search method (full search).
See also Fig. 6, the schematic diagram that Fig. 6 utilizes this benchmark block that second view data 52 is compared and searches for first embodiment of the invention, its comparison mode is the block that second view data 52 is divided into a piece and the identical size of this benchmark block, and the block group 58 that is referred to as to take a sample (Sample Blocks), through in sampling block group 58, finding out a sampling block the most similar after the certain operations, use the motion-vector that calculates image again to this benchmark block.For instance, as shown in Figure 6, one 3 * 3 first square blocks 54 are got in center in 9 * 9 square first view data 50, and set first block 54 and be this benchmark block, just begin by the upper left corner of second view data 52 afterwards, take out one 3 * 3 block in the same way, as sample block 581, the pixel that then moves right takes out another block of 3 * 3 again, as sample block 582, so in the whole second complete view data 52 from left to right, from top to bottom, can take out (9-3+1) * (9-3+1) totally 49 sample blocks, this is region-wide search method.Through drawing a sample block the most similar after these the 49 times comparisons to this benchmark block, this is block alignment algorithm (Block Matching Algorithm), is sample block 5815 to the most similar sample block of this benchmark block in first embodiment of the invention.See also Fig. 7, Fig. 7 is the schematic diagram of the motion vector that Fig. 6 is drawn after comparison is searched in the first embodiment of the invention, can calculate the difference of coordinate position with the coordinate position of this benchmark block that corresponds to first block 54 of sample block 5815, and then draw the motion vector that first view data 50 moves to second view data 52.
See also Fig. 8, the schematic diagram of second view data 72 that first view data 70 that Fig. 8 is obtained in very first time point for second embodiment of the invention and second time point greater than this very first time point are obtained, at first the block that can choose in first view data 70 less than first view data 70 is first block 74, and wherein first block 74 can be positioned at the center of first view data 70; In like manner, the block that can choose in second view data 72 less than second view data 72 is second block 76, and wherein second block 76 can be positioned at the center of second view data 72.Afterwards, can compare the characteristics of image of first block 74 in first view data 70 and the characteristics of image of second block 76 in second view data 72, for example can compare available or effective characteristics of image quantity of first block 74 and available or effective characteristics of image quantity of second block 76, in second embodiment of Fig. 8, available or effective characteristics of image quantity of first block 74 is the available or effective characteristics of image quantity less than second block 76, if when still setting first block 74 this moment for the benchmark block, because first block 74 does not have enough available or effective characteristics of image, so make comparison result not exclusive easily, and increase the probability of estimating the motion-vector that makes mistake, therefore can set second block 76 this moment and be the benchmark block, again according to this benchmark block to the search that compares of first view data 70, wherein this comparison method for searching is to use region-wide search method.
See also Fig. 9, the schematic diagram that Fig. 9 utilizes this benchmark block that first view data 70 is compared and searches for second embodiment of the invention, its comparison mode is the block that first view data 70 is divided into a piece and the identical size of this benchmark block, and the block group 78 that is referred to as to take a sample, through in sampling block group 78, finding out a sampling block the most similar after the certain operations, use the motion-vector that calculates image again to this benchmark block.For instance, as shown in Figure 9, one 3 * 3 second square blocks 76 are got in center in 9 * 9 square second view data 72, and set second block 76 and be this benchmark block, just begin by the upper left corner of first view data 70 afterwards, take out one 3 * 3 block in the same way, as sample block 781, the pixel that then moves right takes out another block of 3 * 3 again, as sample block 782, so in the whole first complete view data 70 from left to right, from top to bottom, can take out (9-3+1) * (9-3+1) totally 49 sample blocks, this is region-wide search method.Through drawing a sample block the most similar to this benchmark block after these the 49 times comparisons, this is the block alignment algorithm, is sample block 7815 to the most similar sample block of this benchmark block in second embodiment of the invention.See also Figure 10, Figure 10 is the schematic diagram of the motion vector that Fig. 9 is drawn after comparison is searched in the second embodiment of the invention, can calculate the difference of coordinate position with the coordinate position of this benchmark block that corresponds to second block 74 of sample block 7815 earlier, and draw the motion vector that second view data 72 moves to first view data 70, the motion vector that again second view data 72 is moved to first view data 70 afterwards is reverse, and can draw the motion vector that first view data 70 moves to second view data 72; Or directly calculate the difference of the coordinate position of the coordinate position of this benchmark block that corresponds to second block 76 and sample block 7815, and then draw the motion vector that first view data 70 moves to second view data 72.
Method compared to known detected image displacement, what person of block who is obtained out in two view data before and after method of the present invention focuses on differentiating earlier has more available or effective characteristics of image quantity, be that the benchmark block is compared another view data of search with block again with more available or effective characteristics of image quantity, to find out the sampling block the most similar to this benchmark block, use the motion-vector of image before and after calculating, therefore choose the tradition comparison mode of benchmark block compared to the picture centre that is fixed on last time point, the present invention can reduce because of the characteristics of image at the picture centre place of last time point or the not enough situation generation that causes judging by accident displacement of available information, and then can effectively increase the probability of estimating correct motion-vector, use and improve image identification rate and correctness.
The above only is preferred embodiment of the present invention, and all equalizations of being done according to claim scope of the present invention change and modify, and all should belong to the covering scope of claim of the present invention.
Claims (4)
1. the method for a detected image displacement, it includes the following step:
(a) obtain first view data in very first time point;
(b) obtain second view data in second time point;
(c) compare in effective characteristics of image quantity of first block in this first view data and this second view data effective characteristics of image quantity with respect to second block of this first block position;
(d) according to the comparison result of step (c), when effective characteristics of image quantity of this first block in this first view data greater than this second view data in during with respect to the effective characteristics of image quantity of this second block of this first block position, setting this first block is the benchmark block; When effective characteristics of image quantity of this first block in this first view data less than this second view data in during with respect to the effective characteristics of image quantity of this second block of this first block position, setting this second block is the benchmark block;
(e) when this first block of setting is the benchmark block, compare this benchmark block and this second view data, to draw the image block that corresponds to this benchmark block in this second view data;
(f) comparison result of foundation (e), calculate the difference of the coordinate position of the coordinate position of the image block that corresponds to this benchmark block in this second view data and this benchmark block, to draw the motion vector that this first view data moves to this second view data;
(g) when this second block of setting is the benchmark block, compare this benchmark block and this first view data, to draw the image block that corresponds to this benchmark block in this first view data;
(h) comparison result of foundation (g), calculate the difference of the coordinate position of the image block that corresponds to this benchmark block in the coordinate position of this benchmark block and this first view data, to draw the motion vector that this first view data moves to this second view data.
2. method according to claim 1, wherein this first block is the center that is positioned at this first view data, and this second block is the center that is positioned at this second view data.
3. method according to claim 1, wherein step (e) also comprises region-wide search and compares this benchmark block and this second view data, to draw the image block that corresponds to this benchmark block in this second view data.
4. method according to claim 1, wherein step (g) also comprises region-wide search and compares this benchmark block and this first view data, to draw the image block that corresponds to this benchmark block in this first view data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610075393 CN101056411B (en) | 2006-04-11 | 2006-04-11 | Method for detecting the image displacement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610075393 CN101056411B (en) | 2006-04-11 | 2006-04-11 | Method for detecting the image displacement |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101056411A CN101056411A (en) | 2007-10-17 |
| CN101056411B true CN101056411B (en) | 2010-09-01 |
Family
ID=38795978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200610075393 Active CN101056411B (en) | 2006-04-11 | 2006-04-11 | Method for detecting the image displacement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101056411B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100585328C (en) * | 2008-02-22 | 2010-01-27 | 济南大学 | Laser image and corresponding pixel distance measurement based displacement measuring device and method |
| CN101799291B (en) * | 2009-02-05 | 2012-06-27 | 义隆电子股份有限公司 | Method for detecting imaging quality of optical tracking inductor |
| CN110675517A (en) * | 2019-09-20 | 2020-01-10 | 青岛海信商用显示股份有限公司 | Express item detection method, device, equipment and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1314051A (en) * | 1999-04-06 | 2001-09-19 | 皇家菲利浦电子有限公司 | Motion estimation |
-
2006
- 2006-04-11 CN CN 200610075393 patent/CN101056411B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1314051A (en) * | 1999-04-06 | 2001-09-19 | 皇家菲利浦电子有限公司 | Motion estimation |
Non-Patent Citations (2)
| Title |
|---|
| 何小海 腾奇志.图像通信 2005年5月第1版，2005年5月第1次印刷.西安电子科技大学出版社 |
| 何小海,腾奇志.图像通信 2005年5月第1版,2005年5月第1次印刷.西安电子科技大学出版社,2005,149-155. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101056411A (en) | 2007-10-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108960211B (en) | Multi-target human body posture detection method and system | |
| US8744195B2 (en) | Object detection metadata | |
| US8433109B2 (en) | Direction controlling system and method of an electronic device | |
| CN101448100B (en) | Method for extracting video captions quickly and accurately | |
| CN103699869B (en) | Method and device for recognizing two-dimension codes | |
| US20130243249A1 (en) | Electronic device and method for recognizing image and searching for concerning information | |
| EP2262221A1 (en) | Portable terminal and user interface control method thereof based on pattern recognition and analysis of image captured by camera | |
| JP5445467B2 (en) | Credit information section detection method, credit information section detection device, and credit information section detection program | |
| CN106375771B (en) | Image Feature Matching method and apparatus | |
| CN103106388A (en) | Method and system of image recognition | |
| CN101056411B (en) | Method for detecting the image displacement | |
| US9117135B2 (en) | Corresponding point searching apparatus | |
| EP3048558A1 (en) | Object detecting method and object detecting apparatus | |
| CN103400130B (en) | File image inclination degree based on energy minimization framework detects and correcting method | |
| CN103324906B (en) | A kind of method and apparatus of legacy detection | |
| CN102737383A (en) | Camera movement analyzing method and device in video | |
| US8837595B2 (en) | Motion estimation method | |
| JP5788299B2 (en) | Image search apparatus, image search method, and program | |
| Huang et al. | Detecting moving text in video using temporal information | |
| CN103646401B (en) | The method that video finger print extracts is realized based on time gradient and spatial gradient | |
| CN108428241A (en) | The movement locus catching method of mobile target in HD video | |
| CN105930813B (en) | A method of detection composes a piece of writing this under any natural scene | |
| TW200937349A (en) | Image displacement detection method | |
| CN103778398B (en) | Image blur method of estimation | |
| CN101986242A (en) | Method for tracking target track in video compression coding process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |