CN108154521B - Moving target detection method based on target block fusion - Google Patents
Moving target detection method based on target block fusion Download PDFInfo
- Publication number
- CN108154521B CN108154521B CN201711288760.7A CN201711288760A CN108154521B CN 108154521 B CN108154521 B CN 108154521B CN 201711288760 A CN201711288760 A CN 201711288760A CN 108154521 B CN108154521 B CN 108154521B
- Authority
- CN
- China
- Prior art keywords
- target
- fusion
- blocks
- small
- block
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
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
- G06T7/223—Analysis of motion using block-matching
-
- 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 invention provides a moving target detection method based on target block fusion, which comprises the following steps: calculating the side length of each target block in the video image, and recording the maximum side length; selecting a current target block for fusion, calculating the distance between the current target block and all targets in a fused target set, recording the minimum distance, and adding the current target block into the target set if the minimum distance is smaller than the maximum side length; and fusing the rest target blocks in the video image until all the target blocks are fused to obtain a complete moving target image. The method continuously fuses all target blocks in the video image, can obtain complete and continuous moving targets, is convenient to track and identify, and is simple and effective; the influence of differential noise can be eliminated, and a new moving target can be detected.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a moving target detection method based on target block fusion.
Background
Target detection is carried out in video, and a common method is to use a difference method, namely image subtraction is directly carried out on two or more adjacent frames of images after registration. After registration, the relative position of a static object is unchanged, and the position of a moving object is changed, so that the moving object has traces in a difference image, and then a moving target region can be extracted after threshold segmentation.
Although the difference method is widely used for moving objects, the moving objects obtained by the difference method are not continuous, and the obtained moving objects are not complete. Although the difference method is an effective method for extracting a moving target, the obtained moving target area is usually discontinuous, i.e. a single target is easily detected into a plurality of discontinuous small target blocks; in addition, differential noise also exists in the moving target extracted by adopting a differential method. Although registration is performed first and then difference is performed, complete registration cannot be performed due to limited registration accuracy, and meanwhile, registration accuracy is affected by changes in conditions such as illumination. In addition, the change of the shooting angle can also increase the registration difficulty, thereby influencing the registration precision.
Disclosure of Invention
The invention aims to provide a moving target detection method based on target block fusion, which is used for solving the problems that the prior art method has differential noise and cannot acquire a complete moving target.
In order to achieve the purpose, the technical scheme of the invention is as follows:
a moving target detection method based on target block fusion comprises the following steps:
1) calculating the side length of each target block in the video image, and recording the maximum side length;
2) selecting a current target block for fusion, calculating the distance between the current target block and all targets in a fused target set, recording the minimum distance, and adding the current target block into the target set if the minimum distance is smaller than the maximum side length;
3) and fusing the rest target blocks in the video image until all the target blocks are fused to obtain a complete moving target image.
Furthermore, after the fusion of each target block is completed, a fusion flag bit is set for each target block.
Further, when the maximum area and the minimum area of the moving object in the video image are calculated, the video image needs to be subjected to image registration, image difference and threshold segmentation to obtain a binary image.
Further, a first aspect ratio and an area of a target block in the video image are calculated, and a maximum aspect ratio and a minimum aspect ratio, and a maximum area and a minimum area of the target block are recorded.
Further, corresponding circumscribed rectangles are set for all target blocks in the target set, the area of each circumscribed rectangle and a second aspect ratio are calculated, and if the area of each circumscribed rectangle is smaller than or equal to the maximum area and larger than or equal to the minimum area, and the second aspect ratio is smaller than or equal to the maximum aspect ratio and larger than or equal to the minimum aspect ratio, the target blocks corresponding to the circumscribed rectangles are added into the first target set.
The invention has the beneficial effects that:
the moving target detection method based on target block fusion provided by the invention continuously fuses all target blocks in the video image to obtain a complete and continuous moving target, is convenient to track and identify, and is simple and effective.
And if the area of the external rectangle is less than or equal to the maximum area of the target blocks in the target set and is greater than or equal to the minimum area of the target blocks in the target set, and the aspect ratio is less than or equal to the maximum aspect ratio of the target blocks in the target set and is greater than or equal to the minimum aspect ratio of the target blocks in the target set, adding the target blocks corresponding to the external rectangle into the final target set, eliminating the influence of differential noise through the steps, extracting a relatively complete moving target, and detecting a newly added moving target.
Drawings
FIG. 1 is a schematic diagram of small target blocks in a video image;
FIG. 2 is a schematic diagram of the final detected target;
FIG. 3 is a flow chart of a moving object detection method based on fast fusion of small objects.
Detailed Description
The following further describes embodiments of the present invention with reference to the drawings:
a moving target detection method based on target block fusion, as shown in fig. 3, includes the following steps:
1. acquiring a video image of a moving target in a video sequence, and performing image registration, image difference and threshold segmentation on the video image to obtain a binary image. The moving object is usually marked in the image by a discontinuous white area, which is referred to as a small object block set in this embodiment, as shown in fig. 1 and denoted by biFor the small target blocks, before the small target blocks are fused, the area of each small target block needs to be calculated, the maximum area and the minimum area of the target blocks are judged and obtained, the square roots of the small target blocks are respectively recorded as objMaxea and objMinarea, the square roots of the small target blocks are calculated, and approximate side lengths are obtained and are respectively recorded as objMaxlen and objMinlen. Simultaneously calculating the aspect ratio of each target block and obtaining the maximum width of the target blockThe height ratio and the minimum aspect ratio are denoted as objMaxWHRatio and objMinWHRatio, respectively.
2. Before the new small target block fusion is performed (i.e. the new small target block fusion is not performed before or a complete target has been extracted by the previous operation), it is first determined whether a fused target block exists before. If the small target blocks do not exist, the current small target block is used as a new target; otherwise, entering step 3 to perform small target block fusion. In addition, in order to prevent repeated fusion, a fusion flag bit flag is set for each small target blocki(i is not less than 0 and not more than N, N is the number of small target blocks), flagiWhen the value is 0, the corresponding target block is not fused, and when the value is flagiAnd other values indicate that the corresponding target blocks are fused.
3. Calculating the current small target block biWith fused target sets sjBelongs to S (j is more than or equal to 0 and less than or equal to M, M is the fused target number, wherein SjBeing small target blocks b belonging to a moving targetiSet of) and records the minimum distance value dminAnd target sjIndex j of (d). Determining the minimum distance value dminThe side lengths objMaxlen of the small object corresponding to the maximum side length are compared, if dminIf the value is less than objMaxlen, the target block b is processediAdding to the target sjPerforming the following steps; otherwise, the target block biAdded as a new target to the target set S. While flag will be seti=1。
4. And repeating the step 3 to fuse the rest small target blocks until flag bits of all the small target blocks are markediAs shown in fig. 2, the white square frame in fig. 2 is a target, which is a valid target after fusion, and the number of valid targets in fig. 2 is 3.
5. For the target set SjEach small target block in the middle is provided with a circumscribed rectangle and the area of the circumscribed rectangle is calculatedAspect ratioIf it is notAnd isThe bounding rectangle is considered a valid target and added to the final target set, i.e. the first target set okE.g. in O.
The specific embodiments are given above, but the present invention is not limited to the above-described embodiments. The basic idea of the present invention lies in the above basic scheme, and it is obvious to those skilled in the art that no creative effort is needed to design various modified models, formulas and parameters according to the teaching of the present invention. Variations, modifications, substitutions and alterations may be made to the embodiments without departing from the principles and spirit of the invention, and still fall within the scope of the invention.
Claims (2)
1. A moving target detection method based on target block fusion is characterized by comprising the following steps:
1) acquiring a video image of a moving target in a video sequence, and performing image registration, image difference and threshold segmentation on the video image to obtain a binary image; non-continuous white areas in the binary image are called small target blocks; determining the maximum area of each targetobjMaxareaAnd minimum areaobjMinarea(ii) a According to the maximum area of each targetobjMaxareaAnd minimum areaobjMinareaCalculating the maximum side length of each objectobjMaxlenAnd minimum side lengthobjMinlenSimultaneously calculating the aspect ratio of each target to obtain the maximum aspect ratio of each targetobjMaxWHRatioAnd minimum aspect ratioobjMinWHRatio;
2) Selecting small target blocks for fusion, and calculating the selected small target blocks and the fused target setSAnd recording the minimum distance and the target s corresponding to the minimum distancejIf the minimum distance is less than the target sjMaximum side length ofobjMaxlenThen the selected smallTarget block is added to the target sjIn the set of (1); otherwise, adding the selected small target block as a new target into the target set S; wherein j is more than or equal to 0 and less than or equal to M, M is the fused target number, sjA set of small target blocks belonging to a certain target;
3) fusing the other small target blocks in the video image according to the step 2) until the fusion of all the small target blocks is completed;
4) for the target sjSetting a circumscribed rectangle for each small target block, and calculating the area and aspect ratio of the circumscribed rectanglewhratio sj If, ifobjMinarea≤area sj ≤objMaxareaAnd isobjMinWHRatio≤whratio sj ≤objMaxWHRatioThen, the target s corresponding to the circumscribed rectangle is determinedjAdd to the final target set.
2. The method of claim 1, wherein a fusion flag is set for each small target block after completion of fusion of each small target block.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711288760.7A CN108154521B (en) | 2017-12-07 | 2017-12-07 | Moving target detection method based on target block fusion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711288760.7A CN108154521B (en) | 2017-12-07 | 2017-12-07 | Moving target detection method based on target block fusion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108154521A CN108154521A (en) | 2018-06-12 |
CN108154521B true CN108154521B (en) | 2021-05-04 |
Family
ID=62466313
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711288760.7A Active CN108154521B (en) | 2017-12-07 | 2017-12-07 | Moving target detection method based on target block fusion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108154521B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110796698B (en) * | 2019-11-07 | 2022-11-29 | 厦门市美亚柏科信息股份有限公司 | Vehicle weight removing method and device with maximum area and minimum length-width ratio |
CN111191730B (en) * | 2020-01-02 | 2023-05-12 | 中国航空工业集团公司西安航空计算技术研究所 | Method and system for detecting oversized image target oriented to embedded deep learning |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020097245A1 (en) * | 2000-12-27 | 2002-07-25 | Il-Kwon Jeong | Sensor fusion apparatus and method for optical and magnetic motion capture systems |
CN101621615A (en) * | 2009-07-24 | 2010-01-06 | 南京邮电大学 | Self-adaptive background modeling and moving target detecting method |
CN103514610A (en) * | 2013-09-17 | 2014-01-15 | 四川虹微技术有限公司 | Method for parting moving target with static background |
CN103559498A (en) * | 2013-09-24 | 2014-02-05 | 北京环境特性研究所 | Rapid man and vehicle target classification method based on multi-feature fusion |
CN103617410A (en) * | 2013-08-30 | 2014-03-05 | 重庆大学 | Highway tunnel parking detection method based on video detection technology |
CN105376505A (en) * | 2015-11-04 | 2016-03-02 | 上海兆芯集成电路有限公司 | Domain block fusion method and device utilizing method |
CN105405150A (en) * | 2015-10-21 | 2016-03-16 | 东方网力科技股份有限公司 | Abnormal behavior detection method and abnormal behavior detection device based fused characteristics |
CN105518702A (en) * | 2014-11-12 | 2016-04-20 | 深圳市大疆创新科技有限公司 | Method, device and robot for detecting target object |
CN106203274A (en) * | 2016-06-29 | 2016-12-07 | 长沙慧联智能科技有限公司 | Pedestrian's real-time detecting system and method in a kind of video monitoring |
-
2017
- 2017-12-07 CN CN201711288760.7A patent/CN108154521B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020097245A1 (en) * | 2000-12-27 | 2002-07-25 | Il-Kwon Jeong | Sensor fusion apparatus and method for optical and magnetic motion capture systems |
CN101621615A (en) * | 2009-07-24 | 2010-01-06 | 南京邮电大学 | Self-adaptive background modeling and moving target detecting method |
CN103617410A (en) * | 2013-08-30 | 2014-03-05 | 重庆大学 | Highway tunnel parking detection method based on video detection technology |
CN103514610A (en) * | 2013-09-17 | 2014-01-15 | 四川虹微技术有限公司 | Method for parting moving target with static background |
CN103559498A (en) * | 2013-09-24 | 2014-02-05 | 北京环境特性研究所 | Rapid man and vehicle target classification method based on multi-feature fusion |
CN105518702A (en) * | 2014-11-12 | 2016-04-20 | 深圳市大疆创新科技有限公司 | Method, device and robot for detecting target object |
CN105405150A (en) * | 2015-10-21 | 2016-03-16 | 东方网力科技股份有限公司 | Abnormal behavior detection method and abnormal behavior detection device based fused characteristics |
CN105376505A (en) * | 2015-11-04 | 2016-03-02 | 上海兆芯集成电路有限公司 | Domain block fusion method and device utilizing method |
CN106203274A (en) * | 2016-06-29 | 2016-12-07 | 长沙慧联智能科技有限公司 | Pedestrian's real-time detecting system and method in a kind of video monitoring |
Also Published As
Publication number | Publication date |
---|---|
CN108154521A (en) | 2018-06-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20130004079A1 (en) | Image processing apparatus, image processing method, and program thereof | |
KR102094506B1 (en) | Method for measuring changes of distance between the camera and the object using object tracking , Computer readable storage medium of recording the method and a device measuring changes of distance | |
CN109478329B (en) | Image processing method and device | |
TW201702989A (en) | A method for detecting and tracing a moving target and a target detection device | |
EP2993621B1 (en) | Method and apparatus for detecting shielding against object | |
TW201327474A (en) | Method for generating depth by disparity estimation | |
WO2012137621A1 (en) | Image processing method and device | |
JP2009512246A (en) | Method and apparatus for determining shot type of an image | |
CN108154521B (en) | Moving target detection method based on target block fusion | |
CN109697725B (en) | Background filtering method and device and computer readable storage medium | |
CN110766711A (en) | Video shot segmentation method, system, device and storage medium | |
WO2017081839A1 (en) | Moving body tracking method, moving body tracking device, and program | |
CN110114801B (en) | Image foreground detection device and method and electronic equipment | |
CN108010044A (en) | A kind of method of video boundaries detection | |
JP2016052013A (en) | Image processing device, image processing method, program and recording medium | |
KR101937436B1 (en) | Apparatus and method for separating background and foreground in 3d video | |
CN108764343B (en) | Method for positioning tracking target frame in tracking algorithm | |
CN111626179B (en) | Micro-expression detection method based on optical flow superposition | |
CN115880683B (en) | Urban waterlogging ponding intelligent water level detection method based on deep learning | |
US9525859B2 (en) | Refinement of user interaction | |
CN111160340B (en) | Moving object detection method and device, storage medium and terminal equipment | |
KR101667011B1 (en) | Apparatus and Method for detecting scene change of stereo-scopic image | |
CN104504700B (en) | A kind of method and system for obtaining image sensor noise level curve | |
JP2015207090A (en) | Image processor, and control method thereof | |
CN113096090B (en) | End face gap visual measurement method with chamfer, device, equipment and storage medium |
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 |