CN108428241A - The movement locus catching method of mobile target in HD video - Google Patents
The movement locus catching method of mobile target in HD video Download PDFInfo
- Publication number
- CN108428241A CN108428241A CN201810427657.4A CN201810427657A CN108428241A CN 108428241 A CN108428241 A CN 108428241A CN 201810427657 A CN201810427657 A CN 201810427657A CN 108428241 A CN108428241 A CN 108428241A
- Authority
- CN
- China
- Prior art keywords
- image
- image block
- mobile target
- block
- frame image
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 230000033001 locomotion Effects 0.000 title claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 21
- 230000011218 segmentation Effects 0.000 claims description 10
- 238000003709 image segmentation Methods 0.000 claims description 5
- 230000000717 retained effect Effects 0.000 claims description 4
- 238000005352 clarification Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims 1
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/215—Motion-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses the movement locus catching method of mobile target in HD video, utilize video cutting technique, high-definition image is divided into polylith small pixel area, it is the small pixel area split to participate in computer subsequent arithmetic, useless region part, which is lost, does not participate in operation, and image data amount is greatly reduced, to improve the processing speed of useful data amount, guarantee is provided to the identification of high-speed moving object, the solution proposed thus.
Description
Technical field
The present invention relates to technical field of image processing, and in particular to the movement locus capture of mobile target in HD video
Method.
Background technology
Image recognition is the chief component of current intelligent development, wherein the image for image recognition relies on more
In camera shooting camera.Popularizing for high-definition camera camera, more accelerates the development of intelligent identifying system, the clarity of image constantly carries
Height is promoted by 1080P quickly to 4K, and high-definition image is conducive to improve the visual field, to realize the identification to a wide range of Small object.So
And the complexity for considerably increasing target and especially moving target identification processing of the quality of image, the thing followed are exactly pair
More stringent requirements are proposed for the performance of image processing hardware, and then causes the cost of hardware product is also continuous therewith to increase.So
And for the development of current technology, the promotion of hardware performance is and the promotion of software Processing Algorithm relatively difficultly
Feasibility it is relatively high.Therefore, how to allow product under the premise of ensureing high-definition image processing capacity, by image
The improvement of Processing Algorithm makes hardware product accomplish inexpensive miniaturization, is the target that people pursue.
Invention content
The target identification method to be solved by this invention for being the non-HD video of tradition is in the mesh suitable for HD video
When identifying other, to the higher problem of hardware product performance requirement, the movement locus for providing mobile target in a kind of HD video is caught
Method is obtained, performance requirement of the high-definition image processing procedure to hardware can be effectively reduced.
To solve the above problems, the present invention is achieved by the following technical solutions:
The movement locus catching method of mobile target in HD video, specifically includes that steps are as follows:
Step 1 determines in video per the segmentation block number N of frame image;
Step 2, the frame image being successively read in video, and the frame image segmentation is obtained into N number of image at N deciles
Block;
Step 3 is captured successively according to N number of image block of ordered pair current frame image when scheduled crawl;
The image block currently grabbed and given mobile clarification of objective image to be detected are carried out image by step 4
Match, to judge to whether there is mobile target to be detected in the image block currently grabbed;If in the image block currently grabbed
There are mobile targets to be detected, then retain current crawl image block;Otherwise, give up the image block currently grabbed;
Step 5, according to retained image block, present position and mobile target to be detected are being protected in current frame image
Stay image position in the block, determine mobile target to be detected current frame image accurate location coordinate;
Step 6 repeats step 2-5, is handled to obtain mobile target to be detected in every frame to every frame image in video
The position coordinates of image, you can obtain the movement locus of mobile target to be detected.
In above-mentioned steps 1, determine that it can handle the size s of image and pending regards according to image processing hardware performance
The size S of every frame image of frequency, to determine segmentation block number N, wherein N >=S/s.
In above-mentioned steps 1, in step 3, the crawl sequential of N number of image block is:With the image block in the upper left corner of every frame image
As start image block, first the image block of the first row is captured from left to right, then to the image block of the second row carry out from
It captures left-to-rightly, and so on, the image block of last column is captured from left to right.
In above-mentioned steps 4, once finding that current crawl image block, then directly will be current there are when mobile target to be detected
The image block of sequential is given up after the image block grabbed.
In above-mentioned steps 4, if complete mobile target to be detected is not present in N number of image block of current frame image,
Current frame image is abandoned, and directly handles the next frame image of video.
Compared with prior art, the present invention utilizes video cutting technique, and high-definition image is divided into polylith small pixel area
Domain, it is the small pixel area split to participate in computer subsequent arithmetic, and useless region part, which is lost, does not participate in operation,
Image data amount is greatly reduced, to improve the processing speed of useful data amount, the identification of high-speed moving object is carried
It has supplied to ensure, the solution proposed thus.
Description of the drawings
Fig. 1 is the flow chart of the movement locus catching method of mobile target in HD video.
Fig. 2 is that single-frame images divides schematic diagram;Wherein (a) is before dividing, after (b) dividing.
Specific implementation mode
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with specific example, and reference
Attached drawing, the present invention is described in more detail.
Referring to Fig. 1, the movement locus catching method of mobile target in HD video specifically includes that steps are as follows:
Step 1 determines in video per the segmentation block number N of frame image.
Dividing block number N rule of thumb can directly give concrete numerical value, such as according to one vertical frame dimension of system configuration computer disposal
Time and design requirement needed for clear image complete to detect time ratios to target to determine the multiple of segmentation image, generally
4K or 1080P image segmentations is the most suitable to 640 × 480 resolution ratio, common computer to the processing speed of this resolution ratio
Enough, corresponding multiple segmentation is carried out to a frame image in accordance with this principle;Corresponding numerical value can also be obtained by calculating, i.e.,
Determine that it can handle the size S of every frame image of the size s of image and pending video according to image processing hardware performance,
To determine segmentation block number N, wherein N >=S/s.
In addition, segmentation block number N also needs to the size in view of mobile target to be detected, ensure be partitioned into image block
Size be greater than mobile target to be detected, i.e., mobile target to be detected can completely occur in wherein 1 image block.
Step 2, the frame image being successively read in video, and by the frame image segmentation at N deciles to get to N number of image
Block.
Step 3 is captured successively according to N number of image block of ordered pair current frame image when scheduled crawl.
The sequential captured to N number of image block of current frame image can be according to requiring to be determined.Such as starting
In the selection of image block, the image block in the upper left corner of every frame image can be selected, the lower left corner of every frame image can also be selected
Image block.Such as in the selection of scan mode:(1) when the image block in the upper left corner for selecting every frame image is as start image
When block:From left to right scanning first can be carried out to first row and captured, then next column image is carried out from a left side first with behavior unit
It is captured to right scanning ... ...;Scanning from top to bottom first can also be carried out to the first row and captured, then to next line to arrange as unit
Carry out scanning crawl ... ... from top to bottom;(2) when the image block in the lower left corner for selecting every frame image is as start image block
When:From left to right scanning first can be carried out to last row and captured, then previous column image is carried out from a left side first with behavior unit
It is captured to right scanning ... ...;Scanning from top to bottom first can also be carried out to the first row and captured, then to next line to arrange as unit
Carry out scanning crawl ... ... from top to bottom.Certainly the selection of crawl sequential is not limited only to aforesaid way, can also use other
Mode, the mode even captured at random.
In the present embodiment, it is to the crawl sequential of N number of image block of current frame image:With the upper left corner of every frame image
Image block as start image block, first the image block of the first row is captured from left to right, then to the image of the second row
Block is captured from left to right, and so on, the image block of last column is captured from left to right.
Current crawl image block and given mobile clarification of objective image to be detected are carried out images match by step 4,
To judge that the image block currently grabbed whether there is mobile target to be detected.When currently, there are shiftings to be detected for crawl image block
When moving-target, then retain current crawl image block.Otherwise, give up current crawl image block.
When it is 1 to determine mobile target to be detected, then conservative manner may be used, i.e., to all N of current frame image
A image block carries out mobile target identification to be detected, at this time only after N number of image block has carried out images match and judged,
It is transferred to subsequent step 5 again;The processing mode of more high speed can also be used, i.e., once finding wherein 1 figure of current frame image
As block is there are when mobile target to be detected, just directly subsequent image block is given up, and enter step 5.
In view of the same target is located exactly at the edge of segmentation image block, and partly appeared in 2 or more images
Situation in the block, the present invention are needed to using image complete match mode.I.e. when mobile target to be detected completely occurs wherein
When in 1 image block, which can just be retained;If N number of image block of current frame image is complete to be checked without occurring
Mobile target is surveyed, then is given up full frame image.
Step 5, according to retained image block, present position and mobile target to be detected are being protected in current frame image
Stay image position in the block, determine mobile target to be detected current frame image accurate location coordinate.
Step 6, the processing for every frame image in video repeat step 3-5, are existed with obtaining mobile target to be detected
Per the position coordinates of frame image, the movement locus of mobile target to be detected thus can be obtained.
1/N part image informations that the present invention only goes out cut and extract when carrying out data processing participate in calculation process, make ginseng
The data volume of calculation process is added to reduce N times, to make the speed of Image Information Processing also improve N times, the calculating after piecemeal
Speed is many soon, effectively obtains real-time processing data, improves the success rate arrested, and realizes that high speed obtains high-speed mobile target
Location coordinate information;Judge in this 1/N block message whether information object in need, if not then by first frame figure
All data of picture are lost, then capture the second frame image, and the second frame image is divided into N parts of regions identical as first frame image
And the N block image zooming-outs for being cut, but being cut into, continue to continue the sequence from first frame image segmentation N BOB(beginning of block)s, extract
The next piece of bands of position 1/N image data, is analyzed and determined, such as without finding target, is continued cycling through under next frame image
One 1/N block diagrams picture carries out analysis until finding target, the target information got is by operational analysis, constantly by crawl
Coordinates of targets information is sent to slave computer, on the other hand performs image display;Thus the processing speed of a frame image is improved
At the equivalent multiple of dividing number, the operational data amount of computer consequently reduces equivalent multiple, greatly improves computer
The processing speed of image is realized to high-speed target processing speed, and the requirement to allocation of computer is reduced, and realizes that low cost is high
Fast interpretative version.
Compliance test result:
The method of the present invention is carried out in OPENCV softwares, is obtained by function cvSetImageROI in OPENCV softwares
Rectangle image is set, by the ROI planned in one frame image of extracted region, pass through function cvCreateImage and create target
Image, then image is copied to the information processing of motion tracking and whether finds useful target, it such as finds target, this 1/N is cut
Image is taken away processing;If do not found target, this frame image is lost, the 1/N images of crawl next frame image subsequent region is gone to carry out
Judge, recycles into line search from each frame N parts of images of cutting and handled until finding useful target.
When whether there is small pentagonal target in the picture frame of HD video shown in Fig. 2 (a) to be identified:First,
The picture frame is divided into 4 deciles as required, as shown in Fig. 2 (b).Then, it for dividing obtained 4 image blocks, uses
From left to right, from top to bottom when ordered pair its extract.Then, 1 region (i.e. the first image block) is read, does not find mesh
Mark, this image block is lost;2 regions (i.e. the second image block) are read, target is not found, this image block is lost;It reads
3 regions (i.e. third image block) are taken, find target, retain the image block, and carry out the identification of dbjective state to the image block
Processing;4 regions (i.e. the 4th image block) are read, target is not found, this image block is lost.
The method of the present invention is the low cost solution for being suitable for obtaining high definition high-speed target information, first according to system
General layout Plan distributes to the time of camera detecting target, and computer must calculate target position before the deadline
Confidence is ceased and is exported, and is generally all completed within the scope of several Milliseconds.
It should be noted that although the above embodiment of the present invention is illustrative, this is not to the present invention
Limitation, therefore the invention is not limited in above-mentioned specific implementation mode.Without departing from the principles of the present invention, all
It is the other embodiment that those skilled in the art obtain under the inspiration of the present invention, is accordingly to be regarded as within the protection of the present invention.
Claims (5)
1. the movement locus catching method of mobile target in HD video, characterized in that specifically include that steps are as follows:
Step 1 determines in video per the segmentation block number N of frame image;
Step 2, the frame image being successively read in video, and the frame image segmentation is obtained into N number of image block at N deciles;
Step 3 is captured successively according to N number of image block of ordered pair current frame image when scheduled crawl;
The image block currently grabbed and given mobile clarification of objective image to be detected are carried out images match by step 4, with
Judge to whether there is mobile target to be detected in the image block currently grabbed;It is waited for if existed in the image block currently grabbed
The mobile target of detection then retains current crawl image block;Otherwise, give up the image block currently grabbed;
Step 5, according to retained image block, present position and mobile target to be detected are retaining figure in current frame image
As position in the block, determine mobile target to be detected current frame image accurate location coordinate;
Step 6 repeats step 2-5, is handled to obtain mobile target to be detected in every frame image to every frame image in video
Position coordinates, you can obtain the movement locus of mobile target to be detected.
2. the movement locus catching method of mobile target in HD video according to claim 1, characterized in that step 1
In, determine that it can handle the size of every frame image of the size s of image and pending video according to image processing hardware performance
S, to determine segmentation block number N, wherein N >=S/s.
3. the movement locus catching method of mobile target in HD video according to claim 1, characterized in that step 1
In, in step 3, the crawl sequential of N number of image block is:Using the image block in the upper left corner of every frame image as start image block, first
The image block of the first row is captured from left to right, then the image block of the second row is captured from left to right, with this
Analogize, the image block of last column is captured from left to right.
4. the movement locus catching method of mobile target in HD video according to claim 1, characterized in that step 4
In, once finding that current crawl image block, then directly will be after the image block currently grabbed there are when mobile target to be detected
The image block of sequential is given up.
5. the movement locus catching method of mobile target in HD video according to claim 1, characterized in that step 4
In, if complete mobile target to be detected is not present in N number of image block of current frame image, current frame image is abandoned,
And directly handle the next frame image of video.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810427657.4A CN108428241B (en) | 2018-05-07 | 2018-05-07 | Method for capturing motion trail of moving target in high-definition video |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810427657.4A CN108428241B (en) | 2018-05-07 | 2018-05-07 | Method for capturing motion trail of moving target in high-definition video |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108428241A true CN108428241A (en) | 2018-08-21 |
CN108428241B CN108428241B (en) | 2022-05-24 |
Family
ID=63162510
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810427657.4A Expired - Fee Related CN108428241B (en) | 2018-05-07 | 2018-05-07 | Method for capturing motion trail of moving target in high-definition video |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108428241B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111275743A (en) * | 2020-01-20 | 2020-06-12 | 深圳奇迹智慧网络有限公司 | Target tracking method, device, computer readable storage medium and computer equipment |
CN116665256A (en) * | 2023-06-05 | 2023-08-29 | 长春理工大学 | Fingerprint matching method based on fingerprint image local area quality |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101957993A (en) * | 2010-10-11 | 2011-01-26 | 上海交通大学 | Adaptive infrared small object detection method |
US20140098999A1 (en) * | 2012-10-04 | 2014-04-10 | Honeywell International Inc. | Detecting motion in a high resolution video |
CN105678807A (en) * | 2016-01-04 | 2016-06-15 | 广东威创视讯科技股份有限公司 | Joined screen-based motion trajectory extraction method and system |
CN106127802A (en) * | 2016-06-16 | 2016-11-16 | 南京邮电大学盐城大数据研究院有限公司 | A kind of movement objective orbit method for tracing |
-
2018
- 2018-05-07 CN CN201810427657.4A patent/CN108428241B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101957993A (en) * | 2010-10-11 | 2011-01-26 | 上海交通大学 | Adaptive infrared small object detection method |
US20140098999A1 (en) * | 2012-10-04 | 2014-04-10 | Honeywell International Inc. | Detecting motion in a high resolution video |
CN105678807A (en) * | 2016-01-04 | 2016-06-15 | 广东威创视讯科技股份有限公司 | Joined screen-based motion trajectory extraction method and system |
CN106127802A (en) * | 2016-06-16 | 2016-11-16 | 南京邮电大学盐城大数据研究院有限公司 | A kind of movement objective orbit method for tracing |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111275743A (en) * | 2020-01-20 | 2020-06-12 | 深圳奇迹智慧网络有限公司 | Target tracking method, device, computer readable storage medium and computer equipment |
CN111275743B (en) * | 2020-01-20 | 2024-03-12 | 深圳奇迹智慧网络有限公司 | Target tracking method, device, computer readable storage medium and computer equipment |
CN116665256A (en) * | 2023-06-05 | 2023-08-29 | 长春理工大学 | Fingerprint matching method based on fingerprint image local area quality |
CN116665256B (en) * | 2023-06-05 | 2024-03-15 | 长春理工大学 | Fingerprint matching method based on fingerprint image local area quality |
Also Published As
Publication number | Publication date |
---|---|
CN108428241B (en) | 2022-05-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9690388B2 (en) | Identification of a gesture | |
US9589595B2 (en) | Selection and tracking of objects for display partitioning and clustering of video frames | |
JP6417702B2 (en) | Image processing apparatus, image processing method, and image processing program | |
JP4373840B2 (en) | Moving object tracking method, moving object tracking program and recording medium thereof, and moving object tracking apparatus | |
CN107633241A (en) | A kind of method and apparatus of panoramic video automatic marking and tracking object | |
CN107710280B (en) | Object visualization method | |
CN108257186B (en) | Method and device for determining calibration image, camera and storage medium | |
JP2011029737A (en) | Surveillance image retrieval apparatus and surveillance system | |
KR20130056309A (en) | Text-based 3d augmented reality | |
JP2013122763A (en) | Video processor and video processing method | |
CN105933665B (en) | A kind of method and device for having access to camera video | |
CN110297545B (en) | Gesture control method, gesture control device and system, and storage medium | |
US11003963B2 (en) | Method and system for identifying targets in scenes shot by a camera | |
CN107094234B (en) | Shooting area shielding method and device applied to mobile camera terminal | |
CN113362441B (en) | Three-dimensional reconstruction method, three-dimensional reconstruction device, computer equipment and storage medium | |
WO2017081839A1 (en) | Moving body tracking method, moving body tracking device, and program | |
CN115103120B (en) | Shooting scene detection method and device, electronic equipment and storage medium | |
US9947106B2 (en) | Method and electronic device for object tracking in a light-field capture | |
CN108428241A (en) | The movement locus catching method of mobile target in HD video | |
CN113194253A (en) | Shooting method and device for removing image reflection and electronic equipment | |
CN114422720B (en) | Video concentration method, system, device and storage medium | |
CN111986229A (en) | Video target detection method, device and computer system | |
WO2011141990A1 (en) | Image processing device and image processing program | |
CN105930813B (en) | A method of detection composes a piece of writing this under any natural scene | |
CN113361360B (en) | Multi-person tracking method and system based on deep learning |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220524 |