CN105469394A - Complex-environment-based intelligent target tracking method - Google Patents
Complex-environment-based intelligent target tracking method Download PDFInfo
- Publication number
- CN105469394A CN105469394A CN201510805380.0A CN201510805380A CN105469394A CN 105469394 A CN105469394 A CN 105469394A CN 201510805380 A CN201510805380 A CN 201510805380A CN 105469394 A CN105469394 A CN 105469394A
- Authority
- CN
- China
- Prior art keywords
- gauss model
- pixel
- image
- detail view
- current time
- 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 18
- 230000003044 adaptive effect Effects 0.000 claims abstract description 6
- 238000004458 analytical method Methods 0.000 claims abstract description 6
- 238000001914 filtration Methods 0.000 claims description 15
- 230000000877 morphologic effect Effects 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 230000006978 adaptation Effects 0.000 claims description 3
- 230000000052 comparative effect Effects 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 abstract 2
- 238000001556 precipitation Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
- G06T2207/20064—Wavelet transform [DWT]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a complex-environment-based intelligent target tracking method. The method comprises: video streaming data at current time are collected and each frame of video image in the video streaming data is analyzed successively; a Gaussian mixture model is started; a background Gaussian model and a foreground Gaussian model are established; the motion interference in the video is filtered; rain-snow stripe noises are removed by wavelet transform and Fourier transform; a target is locked, a tracking operation is started to be executed, and noise removing processing is carried out continuously during the tracking process. According to the invention, on the basis of the adaptive analysis algorithm by the Gaussian mixture model, various motion interferences like leaf shaking caused by wind in a picture in a complex environment are filtered effectively; and rain-snow noises in the picture are also filtered effectively by combining wavelet transform and Fourier transform.
Description
Technical field
The present invention relates to field of video monitoring, particularly relate to a kind of Intelligent target tracking based on complex environment.
Background technology
Reach its maturity with perfect along with technology such as computing machine, network, communication, Streaming Medias, the application of video in social production life is increasingly extensive.In the computer vision systems such as intelligent video monitoring, navigation, remote sensing, Car license recognition, video is topmost information source.But these systemic-functions generally all consider design based on normal weather condition, although the application of outdoor video supervisory system widely, can not play its due effect well in the inclement weather such as rain, snow.In order to ensure that outdoor monitoring system can adapt to various weather condition, also normally can work under the mal-conditions such as sleet, so be necessary very much to carry out removal research to the sleet in air, the image obtained is repaired, eliminate the interference of weather conditions, improve the stability of system.
Summary of the invention
The object of the invention is to overcome the deficiencies in the prior art, a kind of Intelligent target tracking based on complex environment is provided, based on passing through mixed Gauss model adaptive analysis algorithm, the various motion artifacts of picture in effective filtration complex environment, the leaf shake etc. caused as blown, combined with wavelet transformed and Fourier transform can precipitation noises effectively in filtering picture.
The object of the invention is to be achieved through the following technical solutions: a kind of Intelligent target tracking based on complex environment, comprises the following steps:
The video stream data of S1, collection current time, analyzes each frame video image in video stream data successively, has judged whether abnormal frame image, if exist, carry out Exception Type detection;
S2, startup mixed Gauss model, set up X Gauss model to each pixel of the every frame video image in video stream data;
S3, according to the weight of Gauss model, this X Gauss model to be sorted, and accumulative summation is carried out to weight, if the weight of top n Gauss model and be greater than predetermined threshold value, then this top n Gauss model is set to background pixel point, set up background Gauss model, a rear X-N Gauss model is set to foreground pixel point, sets up prospect Gauss model;
S4, be that each pixel mates corresponding Gauss model according to the average of the history pixel of current time pixel and this pixel and variance, calculate the absolute value of the difference of the average of current time pixel and its history pixel, if the ratio of the variance of the history pixel of this absolute value and this pixel is less than preset matching threshold value, then the Gauss model of this current time pixel with background pixel point is mated, if this ratio is not less than preset matching threshold value, then the Gauss model of current time pixel with foreground pixel point is mated;
S5, according to the change of background in video stream data and the comparative result in S4, mixed Gauss model adaptive updates model parameter, upgrades the weights of X Gauss model, draws and the background Gauss model parameter that matching degree is the highest complete complex background modeling;
S6, the foreground image mated of prospect Gauss model is carried out small echo direct transform process, if foreground image is coloured image, then respectively wavelet analysis process is carried out to it by R, G, B triple channel, obtain the approximate diagram LL of foreground image, horizontal direction detail view HL, vertical direction detail view LH and diagonal detail view HH, extract vertical direction detail view, Fourier's direct transform is carried out to vertical direction detail view, and the HFS deleting its default frequency range completes high-frequency filtering;
New vertical direction detail view is obtained by Fourier inversion reduction after S7, high-frequency filtering process, foreground image is reconstructed again by inverse wavelet transform, if former foreground image is coloured image, is then merged by R, G, B triple channel again and obtain the foreground image after filtering.
S8, by Morphological scale-space, extract independently motion target area, lock onto target, start to perform and follow the tracks of operation, in the process of target following, repeated execution of steps S1-S7.
Further, described small echo direct transform is lifting wavelet transform.
Further, described default frequency range adjusts according to sleet size adaptation.
The invention has the beneficial effects as follows: the present invention is based on by mixed Gauss model adaptive analysis algorithm, the various motion artifacts of picture in effective filtration complex environment, the leaf shake etc. caused as blown, combined with wavelet transformed and Fourier transform can precipitation noise effectively in filtering picture and other noises, and pass through Morphological scale-space, extract independently motion target area, lock onto target, start to perform and follow the tracks of operation.
Accompanying drawing explanation
Fig. 1 is schematic flow sheet of the present invention.
Embodiment
Below in conjunction with accompanying drawing, technical scheme of the present invention is described in further detail, but protection scope of the present invention is not limited to the following stated.
As shown in Figure 1, This embodiment describes a kind of Intelligent target tracking based on complex environment, it comprises following multiple step.
The video stream data of S1, collection current time, analyzes each frame video image in video stream data successively, has judged whether abnormal frame image, if exist, carry out Exception Type detection.
S2, startup mixed Gauss model, set up X Gauss model to each pixel of the every frame video image in video stream data.
S3, according to the weight of Gauss model, this X Gauss model to be sorted, and accumulative summation is carried out to weight, if the weight of top n Gauss model and be greater than predetermined threshold value, then this top n Gauss model is set to background pixel point, set up background Gauss model, a rear X-N Gauss model is set to foreground pixel point, sets up prospect Gauss model.
S4, be that each pixel mates corresponding Gauss model according to the average of the history pixel of current time pixel and this pixel and variance, calculate the absolute value of the difference of the average of current time pixel and its history pixel, if the ratio of the variance of the history pixel of this absolute value and this pixel is less than preset matching threshold value, then the Gauss model of this current time pixel with background pixel point is mated, if this ratio is not less than preset matching threshold value, then the Gauss model of current time pixel with foreground pixel point is mated.
S5, according to the change of background in video stream data and the comparative result in S4, mixed Gauss model adaptive updates model parameter, upgrade the weights of X Gauss model, draw the background Gauss model parameter that matching degree is the highest, complete complex background modeling, effectively can filter the various motion artifacts of picture in complex environment, the leaf shake etc. caused as blown.
Further, due in the video pictures of sleety weather, sleet generally shows as nicking form, and therefore, the present invention, can precipitation noise effectively in filtering picture and other noises by combined with wavelet transformed and Fourier transform process.
S6, the foreground image mated of prospect Gauss model is carried out small echo direct transform process, if foreground image is coloured image, then respectively wavelet analysis process is carried out to it by R, G, B triple channel, obtain the approximate diagram LL of foreground image, horizontal direction detail view HL, vertical direction detail view LH and diagonal detail view HH, extract vertical direction detail view, Fourier's direct transform is carried out to vertical direction detail view, and the HFS deleting its default frequency range completes high-frequency filtering.
New vertical direction detail view is obtained by Fourier inversion reduction after S7, high-frequency filtering process, foreground image is reconstructed again by inverse wavelet transform, if former foreground image is coloured image, then merged by R, G, B triple channel again and obtain the foreground image after filtering, remove sleet fringes noise and other fringes noises of picture in complex environment.
S8, by Morphological scale-space, extract independently motion target area, lock onto target, start to perform and follow the tracks of operation, in the process of target following, repeated execution of steps S1-S7, realizes continuing tracking target in the picture after removing motion artifacts and precipitation noise.
In the present invention, described small echo direct transform can adopt lifting wavelet transform.
In the present invention, described default frequency range adjusts according to sleet size adaptation, with the precipitation noise in the different force of rain situation of filtering.
Describe in an illustrative manner according to a kind of Intelligent target tracking based on complex environment of the present invention above with reference to accompanying drawing.But; those skilled in the art are to be understood that; for a kind of Intelligent target tracking based on complex environment that the invention described above proposes; various improvement can also be made on the basis not departing from content of the present invention; or equivalent replacement is carried out to wherein portion of techniques feature; within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.Therefore, protection scope of the present invention should be determined by the content of appending claims.
Claims (3)
1., based on an Intelligent target tracking for complex environment, it is characterized in that, comprise the following steps:
The video stream data of S1, collection current time, analyzes each frame video image in video stream data successively, has judged whether abnormal frame image, if exist, carry out Exception Type detection;
S2, startup mixed Gauss model, set up X Gauss model to each pixel of the every frame video image in video stream data;
S3, according to the weight of Gauss model, this X Gauss model to be sorted, and accumulative summation is carried out to weight, if the weight of top n Gauss model and be greater than predetermined threshold value, then this top n Gauss model is set to background pixel point, set up background Gauss model, a rear X-N Gauss model is set to foreground pixel point, sets up prospect Gauss model;
S4, be that each pixel mates corresponding Gauss model according to the average of the history pixel of current time pixel and this pixel and variance, calculate the absolute value of the difference of the average of current time pixel and its history pixel, if the ratio of the variance of the history pixel of this absolute value and this pixel is less than preset matching threshold value, then the Gauss model of this current time pixel with background pixel point is mated, if this ratio is not less than preset matching threshold value, then the Gauss model of current time pixel with foreground pixel point is mated;
S5, according to the change of background in video stream data and the comparative result in S4, mixed Gauss model adaptive updates model parameter, upgrades the weights of X Gauss model, draws and the background Gauss model parameter that matching degree is the highest complete complex background modeling;
S6, the foreground image mated of prospect Gauss model is carried out small echo direct transform process, if foreground image is coloured image, then respectively wavelet analysis process is carried out to it by R, G, B triple channel, obtain the approximate diagram LL of foreground image, horizontal direction detail view HL, vertical direction detail view LH and diagonal detail view HH, extract vertical direction detail view, Fourier's direct transform is carried out to vertical direction detail view, and the HFS deleting its default frequency range completes high-frequency filtering;
New vertical direction detail view is obtained by Fourier inversion reduction after S7, high-frequency filtering process, foreground image is reconstructed again by inverse wavelet transform, if former foreground image is coloured image, is then merged by R, G, B triple channel again and obtain the foreground image after filtering;
S8, by Morphological scale-space, extract independently motion target area, lock onto target, start to perform and follow the tracks of operation, in the process of target following, repeated execution of steps S1-S7.
2. a kind of Intelligent target tracking based on complex environment according to claim 1, is characterized in that: described small echo direct transform is lifting wavelet transform.
3. a kind of Intelligent target tracking based on complex environment according to claim 1, is characterized in that: described default frequency range adjusts according to sleet size adaptation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510805380.0A CN105469394B (en) | 2015-11-20 | 2015-11-20 | A kind of Intelligent target tracking based on complex environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510805380.0A CN105469394B (en) | 2015-11-20 | 2015-11-20 | A kind of Intelligent target tracking based on complex environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105469394A true CN105469394A (en) | 2016-04-06 |
| CN105469394B CN105469394B (en) | 2018-01-19 |
Family
ID=55607053
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510805380.0A Active CN105469394B (en) | 2015-11-20 | 2015-11-20 | A kind of Intelligent target tracking based on complex environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105469394B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023175A (en) * | 2016-05-13 | 2016-10-12 | 哈尔滨工业大学（威海） | Method for distinguishing thermal wake of underwater navigation body based on three-dimensional mode |
| CN107462884A (en) * | 2017-07-25 | 2017-12-12 | 上海航征测控系统有限公司 | A kind of moving target detecting method and system based on frequency modulated continuous wave radar |
| CN114594533A (en) * | 2022-05-10 | 2022-06-07 | 武汉大学 | Video rainfall monitoring method and device based on self-adaptive Gaussian mixture algorithm |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020031247A1 (en) * | 2000-09-13 | 2002-03-14 | Tomohiko Matsuura | Image processing apparatus, image processing method, program for implementing said method, and storage medium therefor |
| CN104933728A (en) * | 2015-07-13 | 2015-09-23 | 天津理工大学 | Mixed motion target detection method |
| CN105023248A (en) * | 2015-06-25 | 2015-11-04 | 西安理工大学 | Low-SNR (signal to noise ratio) video motion target extraction method |
-
2015
- 2015-11-20 CN CN201510805380.0A patent/CN105469394B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020031247A1 (en) * | 2000-09-13 | 2002-03-14 | Tomohiko Matsuura | Image processing apparatus, image processing method, program for implementing said method, and storage medium therefor |
| CN105023248A (en) * | 2015-06-25 | 2015-11-04 | 西安理工大学 | Low-SNR (signal to noise ratio) video motion target extraction method |
| CN104933728A (en) * | 2015-07-13 | 2015-09-23 | 天津理工大学 | Mixed motion target detection method |
Non-Patent Citations (2)
| Title |
|---|
| MIN HUANG 等: "An Algorithm of the Target Detection and Tracking of the Video", 《PAROO/C PERDOIAC EEDNIAG IENNEEGRININEGE》 * |
| 郑宇杰 等: "一种基于多特征融合的目标跟踪新方法", 《信息化研究》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023175A (en) * | 2016-05-13 | 2016-10-12 | 哈尔滨工业大学（威海） | Method for distinguishing thermal wake of underwater navigation body based on three-dimensional mode |
| CN106023175B (en) * | 2016-05-13 | 2018-11-23 | 哈尔滨工业大学（威海） | Method for distinguishing thermal wake of underwater navigation body based on three-dimensional mode |
| CN107462884A (en) * | 2017-07-25 | 2017-12-12 | 上海航征测控系统有限公司 | A kind of moving target detecting method and system based on frequency modulated continuous wave radar |
| CN114594533A (en) * | 2022-05-10 | 2022-06-07 | 武汉大学 | Video rainfall monitoring method and device based on self-adaptive Gaussian mixture algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105469394B (en) | 2018-01-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107240084B (en) | Method and device for removing rain from single image | |
| CN107248148B (en) | Image noise reduction method and system | |
| CN105184743B (en) | A kind of image enchancing method based on non-linear Steerable filter | |
| Xu et al. | Shadow removal with blob-based morphological reconstruction for error correction | |
| CN104794688A (en) | Single image defogging method and device based on depth information separation sky region | |
| CN104200437A (en) | Image defogging method | |
| CN105869174B (en) | A kind of Sky Scene image partition method | |
| CN110717922A (en) | Image definition evaluation method and device | |
| CN105469394A (en) | Complex-environment-based intelligent target tracking method | |
| CN111161167B (en) | Single image defogging method based on middle channel compensation and self-adaptive atmospheric light estimation | |
| CN102663702A (en) | Natural image denoising method based on regional division | |
| CN108668109A (en) | Image monitoring method based on computer vision | |
| CN116823686B (en) | Night infrared and visible light image fusion method based on image enhancement | |
| CN108921132A (en) | Unmanned aerial vehicle onboard cable detection system | |
| CN104794696A (en) | Image motion blur removing method and device | |
| CN111179186A (en) | Image denoising system for protecting image details | |
| CN105574826A (en) | Thin cloud removing method of remote sensing image | |
| CN109118440A (en) | Single image to the fog method based on transmissivity fusion with the estimation of adaptive atmosphere light | |
| CN111325688A (en) | Unmanned aerial vehicle image defogging method fusing morphological clustering and optimizing atmospheric light | |
| CN112053298A (en) | Image defogging method | |
| CN103996028B (en) | A kind of vehicle Activity recognition method | |
| CN110400270B (en) | License plate defogging method utilizing image decomposition and multiple correction fusion | |
| CN102509271A (en) | Image restoration method based on multi-dimensional decomposition, iteration enhancement and correction | |
| Kusnik et al. | On the robust technique of mixed Gaussian and impulsive noise reduction in color digital images | |
| Liu et al. | Removing rain from single image based on details preservation and background enhancement |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |