CN109949337A - Moving target detecting method and device based on Gaussian mixture model-universal background model - Google Patents
Moving target detecting method and device based on Gaussian mixture model-universal background model Download PDFInfo
- Publication number
- CN109949337A CN109949337A CN201910287686.XA CN201910287686A CN109949337A CN 109949337 A CN109949337 A CN 109949337A CN 201910287686 A CN201910287686 A CN 201910287686A CN 109949337 A CN109949337 A CN 109949337A
- Authority
- CN
- China
- Prior art keywords
- pixel
- model
- gaussian mixture
- universal background
- moving target
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The embodiment of the present invention is the moving target detecting method and device about a kind of based on Gaussian mixture model-universal background model, is related to image identification technical field mainly solving the technical problems that Gaussian mixture model-universal background model method is not high to the accuracy of moving object detection.The technical solution mainly used are as follows: the moving target detecting method based on Gaussian mixture model-universal background model includes: to establish Gaussian mixture model-universal background model based on video image pixel；The parameter of Gaussian mixture model-universal background model is updated based on current frame image pixel, and foreground pixel is obtained to current frame image pixel analysis；Denoising is carried out to the foreground pixel using wavelet threshold denoising method and generates the first filtering image；The moving target that denoising generates current frame image is carried out to first filtering image using the closed operation in Mathematical Morphology.Compared with the existing technology, the noise jamming generated under dynamic background can be eliminated, so that the moving target detected is more complete, to improve moving object detection accuracy.
Description
Technical field
The present embodiments relate to image identification technical field, more particularly to a kind of based on Gaussian mixture model-universal background model
Moving target detecting method and device.
Background technique
With the development of digitized video technology, people can detect the moving object in monitor video, track,
The operation such as identification and analysis.Using these technologies, people can be quickly obtained the position of moving target for needing to detect, track with
And the effective informations such as behavior.Moving object detection is the basis of the technologies such as motion target tracking, Activity recognition and scene description, inspection
The result of survey directly affects the accuracy of subsequent algorithm.Therefore, the accuracy and robustness for how improving target detection, become meter
One of the main direction of studying of calculation machine visual field.Currently, moving target detecting method mainly has: frame differential method, background subtract
Division and optical flow method.
Optical flow method is the method detected according to the luminance information of the target image detected, this method computation complexity
Height, anti-interference ability is weak, so do not use generally.
Frame differential method is to carry out calculus of differences with successive video frames image, realizes the extraction of moving target, is become to background
The adaptability of change is stronger, but there are cavitations for the target detected, and there are missing inspections for moving slow target.
Background subtraction method is most popular method, and background subtraction is first to establish background model, then by present frame figure
Moving target is extracted as making the difference with background model, and this method relies primarily on stable background model to obtain than before more completely
Scape feature obtains moving target by comparing present frame and background model.Gaussian mixture model-universal background model method GMM belongs to background subtraction
One kind, be most popular background subtraction algorithm, but there is also under dynamic background moving object detection it is accurate
The not high problem of property.
Summary of the invention
In view of this, the embodiment of the present invention provide a kind of moving target detecting method based on Gaussian mixture model-universal background model and
Device, mainly solving the technical problems that Gaussian mixture model-universal background model method is not high to the accuracy of moving object detection.
In order to achieve the above objectives, the embodiment of the present invention mainly provides the following technical solutions:
On the one hand, the embodiment of the present invention provides a kind of moving target detecting method based on Gaussian mixture model-universal background model,
Include:
Gaussian mixture model-universal background model is established based on video image pixel；
The parameter of Gaussian mixture model-universal background model is updated based on current frame image pixel, and current frame image pixel analysis is obtained
Obtain foreground pixel；
Denoising is carried out to the foreground pixel using wavelet threshold denoising method and generates the first filtering image；
The movement that denoising generates current frame image is carried out to first filtering image using the closed operation in Mathematical Morphology
Target.
It the purpose of the embodiment of the present invention and solves its technical problem also following technical measures can be used to further realize.
Optionally, the moving target detecting method above-mentioned based on Gaussian mixture model-universal background model, wherein the wavelet threshold
Denoising method is half Threshold Denoising Method of small echo, Wavelet Soft-threshold Denoising Method, any one in small echo hard-threshold denoising method
Kind.
Optionally, the moving target detecting method above-mentioned based on Gaussian mixture model-universal background model, wherein being based on video image
Pixel establishes Gaussian mixture model-universal background model, comprising:
M sub-regions are divided into the video image, M is the positive integer more than or equal to 2；
The mean value of each subregion pixel is sought respectively；
The mixing for constructing each subregion different pixels according to the multiple Gaussian Profiles of the mean value of each subregion pixel is high
This model.
Optionally, the moving target detecting method above-mentioned based on Gaussian mixture model-universal background model, wherein being based on present frame figure
As pixel update Gaussian mixture model-universal background model parameter, specifically:
The parameter of the Gaussian mixture model-universal background model of each pixel is updated based on each pixel of current frame image.
Optionally, the moving target detecting method above-mentioned based on Gaussian mixture model-universal background model, wherein to the video figure
As being divided into M sub-regions, include: before
Obtain the hardware configuration information for executing the device of the moving target detecting method based on Gaussian mixture model-universal background model, root
M value is determined according to the hardware configuration information.
On the other hand, the embodiment of the present invention provides a kind of moving object detection dress based on Gaussian mixture model-universal background model
It sets, comprising:
Unit is established, for establishing Gaussian mixture model-universal background model based on video image pixel；
Analytical unit, for updating the parameter of Gaussian mixture model-universal background model based on current frame image pixel, and to present frame
Image pixel analysis obtains foreground pixel；
First removes dryness unit, generates the first filter for carrying out denoising to the foreground pixel using wavelet threshold denoising method
Wave image；
Second removes dryness unit, for carrying out denoising generation to first filtering image using the closed operation in Mathematical Morphology
The moving target of current frame image.
It the purpose of the embodiment of the present invention and solves its technical problem also following technical measures can be used to further realize.
Optionally, the moving object detection device above-mentioned based on Gaussian mixture model-universal background model, wherein establishing unit and including:
Division module, for being divided into M sub-regions to the video image, M is the positive integer more than or equal to 2；
Module is sought, for seeking the mean value of each subregion pixel respectively；
Module is constructed, for constructing each subregion difference according to the multiple Gaussian Profiles of mean value of each subregion pixel
The mixed Gauss model of pixel.
Optionally, the moving object detection device above-mentioned based on Gaussian mixture model-universal background model, wherein establishing unit and including:
Determining module, for obtain execute the moving target detecting method based on Gaussian mixture model-universal background model device it is hard
Part configuration information determines M value according to the hardware configuration information.
On the other hand, the embodiment of the present invention provides a kind of storage medium, and the storage medium includes the program of storage,
In, equipment where controlling the storage medium in described program operation executes above-mentioned moving target detecting method.
On the other hand, the embodiment of the present invention provides a kind of moving object detection dress based on Gaussian mixture model-universal background model
It sets, described device includes storage medium；And one or more processor, the storage medium are coupled with the processor, institute
Processor is stated to be configured as executing the program instruction stored in the storage medium；It is executed when described program instruction operation above-mentioned
Moving target detecting method.
By above-mentioned technical proposal, the moving target based on Gaussian mixture model-universal background model that technical solution of the present invention provides is examined
Method and device is surveyed at least to have the advantage that
In technical solution provided in an embodiment of the present invention, the Gaussian mixture model-universal background model pair based on the foundation of video image pixel
After current frame image pixel analysis obtains foreground pixel, first the foreground pixel is denoised using wavelet threshold denoising method
The first filtering image is generated, then denoising is carried out to first filtering image using the closed operation in Mathematical Morphology and generates present frame
The moving target of image.Compared with the existing technology, the noise jamming generated under dynamic background can be eliminated, so that the fortune detected
Moving-target is more complete, to improve moving object detection accuracy.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technology of the embodiment of the present invention
Means, and being implemented in accordance with the contents of the specification with presently preferred embodiments of the present invention and cooperate attached drawing to be described in detail below
As after.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 is a kind of moving target detecting method based on Gaussian mixture model-universal background model that the embodiment of the present invention provides
Flow diagram；
Fig. 2 is another moving target detecting method based on Gaussian mixture model-universal background model that the embodiment of the present invention provides
Flow diagram；
Fig. 3 is a kind of moving object detection device based on Gaussian mixture model-universal background model that the embodiment of the present invention provides
Cellular construction schematic diagram.
Specific embodiment
It is of the invention to reach the technical means and efficacy that predetermined inventive embodiments purpose is taken further to illustrate, with
Lower combination attached drawing and preferred embodiment, to the moving target based on Gaussian mixture model-universal background model proposed according to an embodiment of the present invention
Detection method and device its specific embodiments, structure, feature and its effect, detailed description is as follows.In the following description, different
" embodiment " or " embodiment " refer to be not necessarily the same embodiment.In addition, special characteristic in one or more embodiments,
Structure or feature can be combined by any suitable form.
In a first aspect, Fig. 1 is that the moving target detecting method one provided by the invention based on Gaussian mixture model-universal background model is real
Example is applied, referring to Fig. 1, the moving object detection side based on Gaussian mixture model-universal background model that one embodiment of the present of invention proposes
Method, comprising:
101, Gaussian mixture model-universal background model is established based on video image pixel；
Gaussian mixture model-universal background model makes to be constructed to be formed with multiple Gaussian Profiles to each pixel in video image.Example
Such as, in some embodiments, multiple Gaussian Profiles of each pixel can be successively calculated and constructed respectively, but are not limited to
This.
102, the parameter of Gaussian mixture model-universal background model is updated based on current frame image pixel, and to current frame image pixel point
Analysis obtains foreground pixel；
The analysis of foreground pixel can use updated Gaussian mixture model-universal background model, can also be using the Gauss before updating
Pixel value is matched with the Gaussian Profile of Gaussian mixture model-universal background model, is matched if it exists, then the pixel by mixture model-universal background model
Point is background pixel, and otherwise the pixel is detected as foreground pixel, and specific matching process can be found in existing Gaussian Mixture
Background model method is not specially limited in the present embodiment.
103, denoising is carried out to the foreground pixel using wavelet threshold denoising method and generates the first filtering image；
Wherein, the wavelet threshold denoising method can for half Threshold Denoising Method of small echo, Wavelet Soft-threshold Denoising Method,
Any one in small echo hard-threshold denoising method.Wherein, select half Threshold Denoising Method of small echo can be to the high frequency division of image
Amount generates preferably denoising effect.In the moving object detection stage, in order to effectively remove the interference of noise, using half threshold of small echo
The influence that value function denoising method is combined with mathematical morphology denoising method to remove noise to detection effect.
104, denoising is carried out to first filtering image using the closed operation in Mathematical Morphology and generates current frame image
Moving target.
More cavity is remained in first filtering image, is handled by the closed operation in Mathematical Morphology, it will be empty
Filling can eliminate the noise jamming of dynamic background generation.
In technical solution provided in an embodiment of the present invention, the Gaussian mixture model-universal background model pair based on the foundation of video image pixel
After current frame image pixel analysis obtains foreground pixel, first the foreground pixel is denoised using wavelet threshold denoising method
The first filtering image is generated, then denoising is carried out to first filtering image using the closed operation in Mathematical Morphology and generates present frame
The moving target of image.Compared with the existing technology, the noise jamming generated under dynamic background can be eliminated, so that the fortune detected
Moving-target is more complete, to improve moving object detection accuracy.
Second aspect, wherein stepped up for the quality of video pictures, the number of pixel is also provided with big data
The characteristics of, the problem of thus bringing is that the calculation amount of the building of Gaussian mixture model-universal background model is exponentially increased, and leads to Gaussian Mixture
Background model method is difficult to complete the rapid modeling of the video pictures of new high quality.For this purpose, being drawn to be applicable in the video of high quality
The rapid modeling in face, Fig. 2 are moving target detecting method one embodiment provided by the invention based on Gaussian mixture model-universal background model,
Referring to Fig. 2, the moving target detecting method based on Gaussian mixture model-universal background model that one embodiment of the present of invention proposes, packet
It includes:
Gaussian mixture model-universal background model is established based on video image pixel, comprising:
201, the hardware configuration letter for executing the device of the moving target detecting method based on Gaussian mixture model-universal background model is obtained
Breath, determines M value according to the hardware configuration information.
In some embodiments, hardware configuration information includes the information such as processor processing speed, memory size, comprehensive hardware
The processing capacity of configuration information determines M value, and in implementation, M value is positively correlated with processor processing speed, memory size.Wherein, it needs
It is noted that processor processing speed, memory size can also be other information, in above-mentioned step 201 for determining M
Value, for example, in some embodiments, hardware configuration information includes the parameters such as size or the resolution ratio of display, in implementation, M value
It is negatively correlated with the size or resolution ratio of display.Or the determination method of M value can also be other methods, example in step 201
Such as, M value is determined according to the pixel quantity of frame image every in video image, in implementation, the picture of every frame image in M value and video image
Prime number amount is negatively correlated.Certainly, step 201 is nonessential the step of taking, and the value of M, which can also use, presets numerical value, or
It is the numerical value that change is inputted by user.
202, M sub-regions are divided into the video image, M is the positive integer more than or equal to 2；
It include at least two pixel in each subregion.Wherein, the number for the pixel for including in M sub-regions can phase
Together, for example, each subregion is 3 × 39 pixel arrays or 4 × 4 16 pixel arrays.In some embodiments, M sub-district
The number for the pixel for including in domain can also be different, and be divided into first area and second area in advance for video image, and first
Region and second area are different zones.In sub-zone dividing, the pixel that includes in the subregion of the region division of first area
Number be less than second area region division subregion in include pixel number, first area, which can be, is judged as fortune
The less or inactive region of moving-target activity, second area can be the region for being judged as that moving target activity is more.
203, the mean value of each subregion pixel is sought respectively；
Specifically, the calculating process of the mean value for 1 sub-regions pixel are as follows: calculate all pixels in 1 sub-regions
With it is rear, then by the result of calculating except the total number of pixel in the subregion.
204, according to the mean value of each subregion pixel mixed Gauss model of building each subregion different pixels.
In calculating, to each subregion X in imageI, tIt is modeled with the mixed Gauss model that multiple Gaussian Profiles are constituted,Wherein, ηk(XI, t, μI, t, k, ∑I, t, k) it is high
This distribution function, μI, t, kFor mean value, ∑I, t, kFor covariance matrix, ωI, t, kFor weight,K value is got over
Greatly, more complicated background can be more described, but also increases calculation amount, influences live effect, so K generally takes 3~5.XI, tIt can be with
Indicate the subregion of the i-th row t column.By the corresponding multiple Gaussian Profiles of the mean value of each subregion pixel, as the subregion
The corresponding multiple Gaussian Profiles of middle all pixels.Relative to using the multiple Gausses for successively calculating and constructing each pixel respectively
Distribution, by once-through operation, can calculate the corresponding multiple Gaussian Profiles of all pixels in a sub-regions, greatly reduce fortune
Calculation amount.
205, the parameter of the Gaussian mixture model-universal background model of each pixel is updated based on each pixel of current frame image, and to working as
Prior image frame pixel analysis obtains foreground pixel；
It is updated in step in the Gaussian mixture model-universal background model of background, it is mixed to update Gauss using the method for adaptive RTS threshold adjustment
Close the parameter of background model.
206, denoising is carried out to the foreground pixel using wavelet threshold denoising method and generates the first filtering image；
207, denoising is carried out to first filtering image using the closed operation in Mathematical Morphology and generates current frame image
Moving target.
In background modeling stage etch 204, for ease of calculating and improving the speed of modeling, first to video frame images into
Row piecemeal divides sub-region processes, the value of image block subregion pixel is then replaced with the mean value of image block subregion pixel, most
Reconstruct background model is gone with image block subregion averaging method afterwards.
The third aspect, according to method shown in fig. 1 or fig. 2, another embodiment of the disclosure additionally provides a kind of movement
Object detecting device, as shown in figure 3, described device specifically includes that
Unit 10 is established, for establishing Gaussian mixture model-universal background model based on video image pixel；
Analytical unit 20, for updating the parameter of Gaussian mixture model-universal background model based on current frame image pixel, and to current
The analysis of frame image pixel obtains foreground pixel；
First removes dryness unit 30, generates first for carrying out denoising to the foreground pixel using wavelet threshold denoising method
Filtering image；
Second removes dryness unit 40, for carrying out denoising life to first filtering image using the closed operation in Mathematical Morphology
At the moving target of current frame image.
In some embodiments, establishing unit includes:
Division module, for being divided into M sub-regions to the video image, M is the positive integer more than or equal to 2；
Module is sought, for seeking the mean value of each subregion pixel respectively；
Module is constructed, for constructing each subregion difference according to the multiple Gaussian Profiles of mean value of each subregion pixel
The mixed Gauss model of pixel.
In some embodiments, establishing unit includes:
Determining module, for obtain execute the moving target detecting method based on Gaussian mixture model-universal background model device it is hard
Part configuration information determines M value according to the hardware configuration information.
Described device includes pocessor and storage media, above-mentioned to establish unit, analytical unit, first remove dryness unit, second
It removes dryness unit etc. to be stored in a storage medium as program unit, the above-mentioned journey being stored in a storage medium is executed by processor
Sequence unit realizes corresponding function.
Include kernel in above-mentioned processor, is gone in storage medium to transfer corresponding program unit by kernel.Kernel can be set
Set one or more.
In technical solution provided in an embodiment of the present invention, the Gaussian mixture model-universal background model pair based on the foundation of video image pixel
After current frame image pixel analysis obtains foreground pixel, first the foreground pixel is denoised using wavelet threshold denoising method
The first filtering image is generated, then denoising is carried out to first filtering image using the closed operation in Mathematical Morphology and generates present frame
The moving target of image.Compared with the existing technology, the noise jamming generated under dynamic background can be eliminated, so that the fortune detected
Moving-target is more complete, to improve moving object detection accuracy.
The moving object detection device based on Gaussian mixture model-universal background model that the embodiment of the third aspect provides, can be to
The moving target detecting method based on Gaussian mixture model-universal background model provided by the embodiment of first aspect or second aspect is executed,
The relevant meaning being used for and specific embodiment may refer to the correlation in the embodiment of first aspect or second aspect
Description, is no longer described in detail herein.
Fourth aspect, embodiment of the disclosure provide a kind of storage medium, and the storage medium includes the program of storage,
Wherein, equipment where controlling the storage medium when described program is run execute described in first aspect or second aspect based on
The moving target detecting method of Gaussian mixture model-universal background model.
Storage medium may include the non-volatile memory in computer-readable medium, random access memory (RAM)
And/or the forms such as Nonvolatile memory, if read-only memory (ROM) or flash memory (flashRAM), memory include at least one
Storage chip.
5th aspect, embodiment of the disclosure provide the moving object detection device based on Gaussian mixture model-universal background model,
Described device includes storage medium；And one or more processor, the storage medium are coupled with the processor, the place
Reason device is configured as executing the program instruction stored in the storage medium；Described program instruction operation when execute first aspect or
Based on the moving target detecting method of Gaussian mixture model-universal background model described in second aspect.
Embodiment of the disclosure reduces the root-mean-square error of image while improving the Y-PSNR of detection image,
And obtain better visual effect.
It should be understood by those skilled in the art that, embodiment of the disclosure can provide as method, system or computer program
Product.Therefore, embodiment of the disclosure can be used complete hardware embodiment, complete software embodiment or combine software and hardware
The form of the embodiment of aspect.Moreover, it wherein includes that computer is available that embodiment of the disclosure, which can be used in one or more,
It is real in the computer-usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form for the computer program product applied.
The application is the flow chart of the method for reference embodiment of the disclosure, equipment (system) and computer program product
And/or block diagram describes.It should be understood that each process in flowchart and/or the block diagram can be realized by computer program instructions
And/or the combination of the process and/or box in box and flowchart and/or the block diagram.It can provide these computer programs to refer to
Enable the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to generate
One machine so that by the instruction that the processor of computer or other programmable data processing devices executes generate for realizing
The device for the function of being specified in one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
In a typical configuration, calculating equipment includes one or more processors (CPU), input/output interface, net
Network interface and memory.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/
Or the forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flashRAM).Memory is computer-readable medium
Example.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable
Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM),
Digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage or other magnetic storage devices
Or any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, it calculates
Machine readable medium does not include temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.
It should also be noted that, the terms "include", "comprise" or its any other variant are intended to nonexcludability
It include so that the process, method, commodity or the equipment that include a series of elements not only include those elements, but also to wrap
Include other elements that are not explicitly listed, or further include for this process, method, commodity or equipment intrinsic want
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including element
There is also other identical elements in process, method, commodity or equipment.
It will be understood by those skilled in the art that embodiment of the disclosure can provide as method, system or computer program product.
Therefore, complete hardware embodiment, complete software embodiment or combining software and hardware aspects can be used in embodiment of the disclosure
The form of embodiment.Moreover, it wherein includes computer available programs generation that embodiment of the disclosure, which can be used in one or more,
The meter implemented in the computer-usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) of code
The form of calculation machine program product.
The above is only embodiments herein, are not intended to limit this application.To those skilled in the art,
Various changes and changes are possible in this application.It is all within the spirit and principles of the present application made by any modification, equivalent replacement,
Improve etc., it should be included within the scope of the claims of this application.
Claims (10)
1. a kind of moving target detecting method based on Gaussian mixture model-universal background model characterized by comprising
Gaussian mixture model-universal background model is established based on video image pixel；
Based on current frame image pixel update Gaussian mixture model-universal background model parameter, and to current frame image pixel analysis obtain before
Scene element；
Denoising is carried out to the foreground pixel using wavelet threshold denoising method and generates the first filtering image；
The moving target that denoising generates current frame image is carried out to first filtering image using the closed operation in Mathematical Morphology.
2. moving target detecting method according to claim 1, which is characterized in that
The wavelet threshold denoising method is half Threshold Denoising Method of small echo, Wavelet Soft-threshold Denoising Method, small echo hard -threshold are gone
Any one in method for de-noising.
3. moving target detecting method according to claim 1 or 2, which is characterized in that established based on video image pixel
Gaussian mixture model-universal background model, comprising:
M sub-regions are divided into the video image, M is the positive integer more than or equal to 2；
The mean value of each subregion pixel is sought respectively；
The mixed Gaussian mould of each subregion different pixels is constructed with multiple Gaussian Profiles according to the mean value of each subregion pixel
Type.
4. moving target detecting method according to claim 3, which is characterized in that be based on current frame image pixel more new peak
The parameter of this mixture model-universal background model, specifically:
The parameter of the Gaussian mixture model-universal background model of each pixel is updated based on each pixel of current frame image.
5. moving target detecting method according to claim 3, which is characterized in that be divided into M to the video image
Subregion includes: before
The hardware configuration information for executing the device of the moving target detecting method based on Gaussian mixture model-universal background model is obtained, according to institute
It states hardware configuration information and determines M value.
6. a kind of moving object detection device based on Gaussian mixture model-universal background model characterized by comprising
Unit is established, for establishing Gaussian mixture model-universal background model based on video image pixel；
Analytical unit, for updating the parameter of Gaussian mixture model-universal background model based on current frame image pixel, and to current frame image
Pixel analysis obtains foreground pixel；
First removes dryness unit, generates the first filtering figure for carrying out denoising to the foreground pixel using wavelet threshold denoising method
Picture；
Second removes dryness unit, generates currently for carrying out denoising to first filtering image using the closed operation in Mathematical Morphology
The moving target of frame image.
7. moving object detection device according to claim 6, which is characterized in that establishing unit includes:
Division module, for being divided into M sub-regions to the video image, M is the positive integer more than or equal to 2；
Module is sought, for seeking the mean value of each subregion pixel respectively；
Module is constructed, for constructing each subregion different pixels according to the multiple Gaussian Profiles of mean value of each subregion pixel
Mixed Gauss model.
8. moving object detection device according to claim 7, which is characterized in that establishing unit includes:
Determining module is matched for obtaining the hardware of device of moving target detecting method of the execution based on Gaussian mixture model-universal background model
Confidence breath, determines M value according to the hardware configuration information.
9. a kind of storage medium, which is characterized in that the storage medium includes the program of storage, wherein run in described program
When control the storage medium where equipment perform claim require any one of 1 to 5 described in moving target detecting method.
10. a kind of moving object detection device based on Gaussian mixture model-universal background model, which is characterized in that described device includes storage
Medium；And one or more processor, the storage medium are coupled with the processor, the processor is configured to executing
The program instruction stored in the storage medium；Perform claim requires described in any one of 1 to 5 when described program instruction operation
Moving target detecting method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910287686.XA CN109949337A (en) | 2019-04-11 | 2019-04-11 | Moving target detecting method and device based on Gaussian mixture model-universal background model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910287686.XA CN109949337A (en) | 2019-04-11 | 2019-04-11 | Moving target detecting method and device based on Gaussian mixture model-universal background model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109949337A true CN109949337A (en) | 2019-06-28 |
Family
ID=67014301
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910287686.XA Pending CN109949337A (en) | 2019-04-11 | 2019-04-11 | Moving target detecting method and device based on Gaussian mixture model-universal background model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109949337A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111539975A (en) * | 2020-04-09 | 2020-08-14 | 普联技术有限公司 | Method, device and equipment for detecting moving target and storage medium |
| CN113392677A (en) * | 2020-03-12 | 2021-09-14 | 阿里巴巴集团控股有限公司 | Target object detection method and device, storage medium and terminal |
| CN113554685A (en) * | 2021-08-02 | 2021-10-26 | 中国人民解放军海军航空大学航空作战勤务学院 | Method and device for detecting moving target of remote sensing satellite, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778595A (en) * | 2016-12-12 | 2017-05-31 | 河北工业大学 | The detection method of abnormal behaviour in crowd based on gauss hybrid models |
| CN107368784A (en) * | 2017-06-15 | 2017-11-21 | 西安理工大学 | A kind of novel background subtraction moving target detecting method based on wavelet blocks |
-
2019
- 2019-04-11 CN CN201910287686.XA patent/CN109949337A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778595A (en) * | 2016-12-12 | 2017-05-31 | 河北工业大学 | The detection method of abnormal behaviour in crowd based on gauss hybrid models |
| CN107368784A (en) * | 2017-06-15 | 2017-11-21 | 西安理工大学 | A kind of novel background subtraction moving target detecting method based on wavelet blocks |
Non-Patent Citations (1)
| Title |
|---|
| 邹斐: "煤矿井下运动目标检测与跟踪研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113392677A (en) * | 2020-03-12 | 2021-09-14 | 阿里巴巴集团控股有限公司 | Target object detection method and device, storage medium and terminal |
| CN111539975A (en) * | 2020-04-09 | 2020-08-14 | 普联技术有限公司 | Method, device and equipment for detecting moving target and storage medium |
| CN111539975B (en) * | 2020-04-09 | 2023-12-22 | 成都市联洲国际技术有限公司 | Method, device, equipment and storage medium for detecting moving object |
| CN113554685A (en) * | 2021-08-02 | 2021-10-26 | 中国人民解放军海军航空大学航空作战勤务学院 | Method and device for detecting moving target of remote sensing satellite, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108550161A (en) | A kind of dimension self-adaption core correlation filtering fast-moving target tracking method | |
| US9098919B2 (en) | Detecting motion in a high resolution video | |
| CN107507225B (en) | Moving object detection method, device, medium and computing equipment | |
| CN109949337A (en) | Moving target detecting method and device based on Gaussian mixture model-universal background model | |
| CN102521840B (en) | Moving target tracking method, system and terminal | |
| JP2009544090A (en) | Image processing for change detection | |
| CN111311611B (en) | Real-time three-dimensional large-scene multi-object instance segmentation method | |
| Huang et al. | An efficient optical flow based motion detection method for non-stationary scenes | |
| CN111681198A (en) | Morphological attribute filtering multimode fusion imaging method, system and medium | |
| Lian et al. | A novel method on moving-objects detection based on background subtraction and three frames differencing | |
| CN110060278B (en) | Method and device for detecting moving target based on background subtraction | |
| Zhao et al. | Real-time stereo on GPGPU using progressive multi-resolution adaptive windows | |
| CN117274515A (en) | Visual SLAM method and system based on ORB and NeRF mapping | |
| CN107451595A (en) | Infrared image salient region detection method based on hybrid algorithm | |
| Wu et al. | Superpixel-based image noise variance estimation with local statistical assessment | |
| CN114724218A (en) | Video detection method, device, equipment and medium | |
| CN108830828A (en) | A kind of method for detecting change of remote sensing image and device | |
| CN108765463A (en) | A kind of moving target detecting method calmodulin binding domain CaM extraction and improve textural characteristics | |
| Wu et al. | Direct sparse odometry with planes | |
| CN113902772A (en) | Crowd counting method and device, computer storage medium and processor | |
| Ljubičić et al. | SSIMS-Flow: Image velocimetry workbench for open-channel flow rate estimation | |
| Ji et al. | DRV-SLAM: An Adaptive Real-Time Semantic Visual SLAM Based on Instance Segmentation Toward Dynamic Environments | |
| CN111383216B (en) | Method and device for detecting change between images | |
| Gong et al. | Near-real-time stereo matching with slanted surface modeling and sub-pixel accuracy | |
| CN113658180B (en) | Surface defect region segmentation method and device based on spatial context guidance |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190628 |