CN102831472A - People counting method based on video flowing image processing - Google Patents
People counting method based on video flowing image processing Download PDFInfo
- Publication number
- CN102831472A CN102831472A CN2012102747424A CN201210274742A CN102831472A CN 102831472 A CN102831472 A CN 102831472A CN 2012102747424 A CN2012102747424 A CN 2012102747424A CN 201210274742 A CN201210274742 A CN 201210274742A CN 102831472 A CN102831472 A CN 102831472A
- Authority
- CN
- China
- Prior art keywords
- moving target
- video streaming
- concyclic
- streaming image
- profile
- 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
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides a people counting method based on video flowing image processing. The method comprises the following steps of: S1, acquiring a video flowing image of a monitoring region to be used as an input image; S2, removing noises in the input image by using median filter processing, and carrying out histogram equalization processing to improve the contrast of the input image obtained in the step S1; S3, segmenting a motion target region through using a Gaussian background modeling method according to the input image (obtained in the step S2) with the high contrast; S4, carrying out erosion smoothing operation on the segmented motion target region through an erosion smoothing operator to obtain a motion target outline; S5, scanning the motion target outline obtained in the step S4 by using a plurality of ring modules, carrying out concyclic overlap ratio HO calculation to extract human head outlines; and S6, counting the human head outlines to obtain the number of people. By using people counting method based on the video flowing image processing, the efficiency and the accuracy degree of counting the number of the people in real time in a public area are effectively improved.
Description
Technical field
The invention belongs to video image and handle and distinguishment technical field, particularly a kind of demographic method of handling based on video streaming image.
Background technology
Demographics is an important application of moving object detection and tracking technology, also is frontier that ten minutes is active in the current intelligent vision systematic study.Traditional demographic method is to utilize artificial counting or artificial electronic equipment flip-flop number, along with arrival of information age, develops a kind of intelligentized demographic method and seems very necessary.
Intelligent demographics technology is exactly to utilize computer vision and image process method to set up an intelligent management system; The intervention that does not need the people, or situation that only needs are seldom intervened under; Realize pedestrian's location, tracking and carry out accurate demographics on this basis through the video sequence analysis of video camera being clapped record; Further judge the trend of flow of the people; Accomplished to accomplish daily management and can when abnormal conditions take place, in time make a response again, thereby a kind of add advanced and feasible Intelligent treatment scheme are provided.
The demographic method of handling based on video streaming image not only has very strong scientific research value, also has very strong practical value, and its cost is low, and life cycle is long, and the accuracy rate of statistics is high, does number of research projects over against this both at home and abroad at present.
The patent of invention that Chinese patent 03109626.3, name are called " small insect automatic technique system " discloses the automatic counter system of a kind of small insect.Yet this invention only is the automatic counting of in specific background environment and specific region, realizing the specific objective body.This counting technology to the specific objective body can not satisfy in the various environment in public domain the demand to the people flow rate statistical of continuous variation.
In view of this, be necessary the demographic method in the public domain in the prior art is improved, to address the above problem.
Summary of the invention
The object of the present invention is to provide a kind of demographic method of handling based on video streaming image, in order to improve the efficient and the accuracy of the real-time demographics in the public domain.
For realizing above-mentioned purpose, the invention provides a kind of demographic method of handling based on video streaming image, this method may further comprise the steps:
S1, obtain guarded region video streaming image as input picture;
S2, handle to remove the noise in the said input picture, and carry out histogram equalization and handle, to improve the contrast of the input picture that is obtained among the step S1 through medium filtering;
S3, according to the input picture of the high-contrast of being obtained among the step S2, be partitioned into motion target area through the Gaussian Background modeling;
S4, the motion target area that is partitioned into is corroded level and smooth computing through the corrosion smoothing operator, obtain the moving target profile;
S5, some annulus templates are scanned the moving target profile that obtains among the step S4, and carry out concyclic registration HO and calculate, to extract number of people profile;
S6, to number of people profile counting, to obtain number.
As further improvement of the present invention, said step S1 is specially: the video streaming image that obtains guarded region through video camera is as input picture, and said guarded region is positioned at the oblique below of video camera.
As further improvement of the present invention; Gaussian Background modeling among the said step S3 is specially: use recurrence method to calculate the Gaussian distribution parameter; And adopting different turnover rate α, β to upgrade computing to it, said Gaussian distribution parameter comprises average value mu and variances sigma
2, this renewal operational formula is:
Wherein, α, β ∈ [0,1], and α>β.
As further improvement of the present invention, the level and smooth computing of corrosion is among the said step S4: the corrosion smoothing operator of utilization 3 * 3 sizes corrodes level and smooth computing to motion target area, obtains the moving target profile.
As further improvement of the present invention; Said step S5 is specially: according to the moving target profile that is obtained among the step S4; With some annulus templates said moving target profile is scanned, extracting the moving target point in the annulus template zone, and the moving target point that is extracted and annulus template are carried out concyclic registration HO calculate; Then should concyclic registration HO and setting threshold T make comparisons
If concyclic registration HO more than or equal to setting threshold T, extracts this number of people profile,
If concyclic registration HO less than setting threshold T, does not extract this number of people profile,
The computing formula of said concyclic registration HO is:
Wherein, HCN is the moving target point summation in the annulus template zone, and r1 and r2 are respectively the inside radius and the external radius of this annulus template.
As further improvement of the present invention, said setting threshold T is 80%.
As further improvement of the present invention, the inside radius r1 of said annulus template and the difference between the external radius r2 are 10 pixels.
Compared with prior art; The invention has the beneficial effects as follows: through the present invention; Can carry out the statistics of real-time number to huge places of flow of the people such as supermarket, market, station, banks, thereby improve the efficient and the accuracy of the real-time demographics in the public domain effectively.
Description of drawings
Fig. 1 is a kind of based on the schematic flow sheet in demographic method one embodiment of video streaming image processing for the present invention;
Fig. 2 is the principle of work synoptic diagram of the video streaming image that obtains guarded region shown in Figure 1;
Fig. 3 is that medium filtering shown in Figure 1 is handled the schematic diagram of removing the noise in the input picture;
Fig. 4 is the principle of work synoptic diagram of the level and smooth computing of corrosion shown in Figure 1;
Fig. 5 is a corrosion smoothing operator synoptic diagram shown in Figure 4;
Fig. 6 is the design sketch of the level and smooth computing of corrosion shown in Figure 4;
Fig. 7 is the synoptic diagram of annulus template shown in Figure 1;
The schematic diagram that Fig. 8 scans the moving target profile for annulus template shown in Figure 1.
Embodiment
Below in conjunction with each embodiment shown in the drawings the present invention is elaborated; But should be noted that; These embodiments are not limitation of the present invention; The function that those of ordinary skills do according to these embodiments, method, or structural equivalent transformation or substitute all belong within protection scope of the present invention.
Join shown in Figure 1ly, Fig. 1 is the schematic flow sheet in a kind of demographic method embodiment of handling based on video streaming image of the present invention.In this embodiment, a kind of demographic method of handling based on video streaming image, this method may further comprise the steps:
S1, obtain guarded region video streaming image as input picture.
About the method for the demographics in the public domain, commonly used method based on kinetic characteristic is arranged, based on the method for shape information, based on method of method, neural network method, small echo and the SVMs of the method for pedestrian dummy, structural element, stereoscopic vision etc.
Join shown in Figure 2ly, a kind of demographic method of handling based on video streaming image of the present invention is based on video camera and vertically takes and be applicable to outdoor situations and indoor situation.In this embodiment, this step S1 is specially: the video streaming image that obtains guarded region 30 through video camera 10 is as input picture, said guarded region 30 be positioned at video camera 10 under.
Concrete, video camera 10 be arranged on gateway 20 directly over, the pedestrian can walk up and down in gateway 20 on the direction of arrow 201.The guarded region 30 that video camera 10 is obtained can cover the Zone Full of gateway 20 fully.
In this embodiment, this guarded region 30 is a square, can certainly be rectangle or circular or other shapes.Video camera 10 is positioned at the oblique upper of the central point 301 of guarded region 30, and we can derive the oblique below that this guarded region 30 is positioned at video camera 10 thus.
S2, handle to remove the noise of video streaming image, and carry out histogram equalization and handle, to improve the contrast of the input picture that is obtained among the step S1 through medium filtering.
Medium filtering is a kind of nonlinear smoothing filtering, when it mainly plays the protection image border, can remove the noise in the input picture again.So-called medium filtering is handled, be meant with certain pixel in the input picture (x be that the gray scale of all pixels in the wicket at center is pressed series arrangement from big to small y), with intermediate value as (x, the gray-scale value of y) locating.If in the window even number pixel is arranged, then gets the mean value of two intermediate values.
As shown in Figure 3, Fig. 3 has shown how median filtering algorithm eliminates the noise in the input picture.The pixel (x, gray scale y) that are selected in this input picture of digitized representation among Fig. 3 in the synoptic diagram of left side.Can find out that " 8 " of former figure centre and gray scale on every side differ greatly, it can think a noise.Medium filtering through 3 * 3 windows (promptly 9 pixels are got intermediate value) is handled, to obtain the right side synoptic diagram among Fig. 3.Therefrom we can find out, the noise of left side synoptic diagram is successfully removed among Fig. 3, to obtain the synoptic diagram on right side among Fig. 3.
The medium filtering processing can overcome linear filtering under certain conditions, and (for example: the image detail fuzzy problem that average filter etc.) is brought can effectively filter out the noise that impulse disturbances and image scanning are produced.When keeping the original clear profile of input picture, can especially preferable effect be arranged to the noise remove of input picture again to salt-pepper noise and impulsive noise.
But medium filtering is handled the noise that goes to handle in the input picture and can be caused the contrast of input picture to descend.Therefore, after handling the noise in the input picture, need to improve the contrast of this input picture, be used for successive image and handle through medium filtering.In this embodiment, can adopt histogram equalization to handle, to improve the contrast of the input picture that is obtained among the step S1.
The histogram equalization processing is that the histogram distribution of the input picture after handling the medium filtering among the step S2 changes over equally distributed histogram, to obtain the histogram of overall equalization.Gray level is r in the input picture
kEach pixel mapping gray level in the output image be s
kThe formula of respective pixel be formula (1),
Wherein, r is the gray scale of the input picture after medium filtering is handled, and s is the gradation of image after histogram modification; N is a pixel summation in the image, n
kBe that gray level is r
kNumber of pixels, L is gray level possible in an image sum.
S3, according to the input picture of the height that obtained among step S2 contrast, be partitioned into motion target area through the Gaussian Background modeling.
In the Gaussian distribution background model, static gray values of pixel points satisfies Gaussian distribution, is that benchmark is done the vibration that is no more than certain deviation near it with certain average promptly.Whether each pixel in the scene all adopts Gaussian distribution to set up background model, for each pixel in the new frame video streaming image, be complementary with its background Gaussian distribution according to its pixel value, judges that it is foreground pixel point or background pixel point.Belong to background pixel arbitrary pixel (x, y), it is μ that its gray-scale value I satisfies average, variance is σ
2Gaussian distribution.Therefore, coordinate figure be (x, y), gray-scale value is that the probability that certain pixel of I belongs to background pixel is:
For any pixel (x, y), (x y) less than certain setting threshold U, then is considered to the foreground point if its gray-scale value I belongs to the Probability p of background pixel; If not, think that it is a background dot.In this embodiment, this setting threshold U can be set at 90%.
If lighting change or a long-time static object have taken place in the static scene to begin to move; The pixel that then moves the front position corresponding to this object on the image will be considered to the foreground point; This will produce cumulative mistake in the later stage target following, this just needs Gauss model can respond these conversion.The information that just requires to utilize video sequence to provide comes the parameter of Gauss model is upgraded.After the initialization of background estimating image is accomplished, along with the arrival of the new images of each frame, for the continuous Gaussian distribution parameter of background image updating adaptively.
In this embodiment, the Gaussian Background modeling is improved, use recurrence method to calculate Gaussian distribution mean parameter μ and variances sigma
2, and to μ and variances sigma
2Adopt different turnover rate α, β upgrades, and this frequency more new formula is:
Wherein, α, β ∈ [0,1], and α>β.Preferably, α is 0.005, and β is 0.003.
S4, the motion target area that is partitioned into is corroded level and smooth computing through the corrosion smoothing operator, obtain the moving target profile.
The image of crossing through the S3 step process is a bianry image.Wherein, the pixel that belongs to motion target area is white, and the pixel that belongs to background is a black.Because the motion target area that extracts is entity area basically, so need further handle it.In this embodiment, used the corrosion smoothing operator to corrode level and smooth computing, to obtain the moving target profile.
Ginseng Fig. 4 and shown in Figure 5, in this embodiment, corroding level and smooth computing is the details of picture structure in inspection 3 * 3 windows, wherein P is a current point, P
0-P
7Be its 8 abutment points.The criterion of corroding level and smooth computing is: when P=255 (being that P is a white pixel), if P
0, P
1, P
2, P
3, P
4, P
5, P
6, P
7All equal 255 (promptly all being white pixel), then P becomes 0 (promptly becoming black picture element).
In conjunction with shown in Figure 6, corrode level and smooth computing through the corrosion smoothing operator, extract with motion target area, to obtain the moving target profile shown in right side among Fig. 6 left side among Fig. 6.
S5, some annulus templates are scanned the moving target profile that obtains among the step S4, and carry out concyclic registration HO and calculate, to extract number of people profile.
Processing through the S4 step; The moving target profile has been extracted; The moving target profile has comprised the various piece of human body; And we need look for is the number of people profile of moving target profile, is exactly concyclic characteristic and number of people profile is different from the maximum characteristic of other partial contour of human body, and promptly number of people profile is circle or sub-circular.
In conjunction with Fig. 7 and shown in Figure 8; In this embodiment, said step S5 is specially: according to the moving target profile that is obtained among the step S4, scan with 60 pairs of said moving target profiles of some annulus templates; To extract the moving target point 601 in annulus template 60 zones; And moving target point that is extracted 601 and annulus template 60 are carried out concyclic registration HO calculate, then should concyclic registration HO and setting threshold T make comparisons
If concyclic registration HO more than or equal to setting threshold T, extracts this number of people profile,
If concyclic registration HO less than setting threshold T, does not extract this number of people profile,
The computing formula of said concyclic registration HO is:
Wherein, HCN is moving target point 601 summations in annulus template 60 zones, and r1 and r2 are respectively the inside radius and the external radius of this annulus template 60.
In this embodiment, this annulus template 60 be some with r be the circle of radius combination (be r1 r r2).Wherein, r1 and r2 are the inside radius and the external radiuss of this annulus template 60.More specifically, inside radius r1 and the difference between the external radius r2 in each group annulus template 60 is 10 pixels.
S6, to number of people profile counting, to obtain number.
In this embodiment, through scanning with 60 pairs of moving target profiles of annulus template among the step S5,, then think people's head region, and counting is 1 if the concyclic registration HO of this moving target profile and annulus template 60 is greater than or equal to setting threshold T; Exclude this moving target profile of having confirmed as people's head region simultaneously, continue scanning for the second time then.By that analogy, stop scanning during less than setting threshold T, and stop counting up to the concyclic registration HO that scans.The number of everyone head region that therefore, before scans is exactly the number in the video streaming image that is obtained in the guarded region 30 in special time period.
The listed a series of detailed description of preceding text only is specifying to feasibility embodiment of the present invention; They are not in order to restriction protection scope of the present invention, allly do not break away from equivalent embodiment or the change that skill of the present invention spirit done and all should be included within protection scope of the present invention.
To those skilled in the art, obviously the invention is not restricted to the details of above-mentioned example embodiment, and under the situation that does not deviate from spirit of the present invention or essential characteristic, can realize the present invention with other concrete form.Therefore; No matter from which point; All should regard embodiment as exemplary; And be nonrestrictive, scope of the present invention is limited accompanying claims rather than above-mentioned explanation, therefore is intended to the implication of the equivalents that drops on claim and all changes in the scope are included in the present invention.Should any Reference numeral in the claim be regarded as limit related claim.
In addition; Describing according to embodiment though should be appreciated that this instructions, is not that each embodiment only comprises an independently technical scheme; This narrating mode of instructions only is for clarity sake; Those skilled in the art should make instructions as a whole, and the technical scheme among each embodiment also can form other embodiments that it will be appreciated by those skilled in the art that through appropriate combination.
Claims (7)
1. demographic method of handling based on video streaming image is characterized in that this method may further comprise the steps:
S1, obtain guarded region video streaming image as input picture;
S2, handle to remove the noise in the said input picture, and carry out histogram equalization and handle, to improve the contrast of the input picture that is obtained among the step S1 through medium filtering;
S3, according to the input picture of the high-contrast of being obtained among the step S2, be partitioned into motion target area through the Gaussian Background modeling;
S4, the motion target area that is partitioned into is corroded level and smooth computing through the corrosion smoothing operator, obtain the moving target profile;
S5, some annulus templates are scanned the moving target profile that obtains among the step S4, and carry out concyclic registration HO and calculate, to extract number of people profile;
S6, to number of people profile counting, to obtain number.
2. the demographic method of handling based on video streaming image according to claim 1; It is characterized in that; Said step S1 is specially: the video streaming image that obtains guarded region through video camera is as input picture, and said guarded region is positioned at the oblique below of video camera.
3. the demographic method of handling based on video streaming image according to claim 1; It is characterized in that; Gaussian Background modeling among the said step S3 is specially: use recurrence method to calculate the Gaussian distribution parameter; And adopting different turnover rate α, β to upgrade computing to it, said Gaussian distribution parameter comprises average value mu and variances sigma
2, this renewal operational formula is:
Wherein, α, β ∈ [0,1], and α>β.
4. the demographic method of handling based on video streaming image according to claim 1; It is characterized in that; The level and smooth computing of corrosion is among the said step S4: the corrosion smoothing operator of utilization 3 * 3 sizes corrodes level and smooth computing to motion target area, obtains the moving target profile.
5. the demographic method of handling based on video streaming image according to claim 1; It is characterized in that said step S5 is specially:, said moving target profile is scanned with some annulus templates according to the moving target profile that is obtained among the step S4; To extract the moving target point in the annulus template zone; And the moving target point that is extracted and annulus template are carried out concyclic registration HO calculate, then should concyclic registration HO and setting threshold T make comparisons
If concyclic registration HO more than or equal to setting threshold T, extracts this number of people profile,
If concyclic registration HO less than setting threshold T, does not extract this number of people profile,
The computing formula of said concyclic registration HO is:
Wherein, HCN is the moving target point summation in the annulus template zone, and r1 and r2 are respectively the inside radius and the external radius of this annulus template.
6. the demographic method of handling based on video streaming image according to claim 5 is characterized in that said setting threshold T is 80%.
7. the demographic method of handling based on video streaming image according to claim 5 is characterized in that the inside radius r1 of said annulus template and the difference between the external radius r2 are 10 pixels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012102747424A CN102831472A (en) | 2012-08-03 | 2012-08-03 | People counting method based on video flowing image processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012102747424A CN102831472A (en) | 2012-08-03 | 2012-08-03 | People counting method based on video flowing image processing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102831472A true CN102831472A (en) | 2012-12-19 |
Family
ID=47334594
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012102747424A Pending CN102831472A (en) | 2012-08-03 | 2012-08-03 | People counting method based on video flowing image processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102831472A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104091351A (en) * | 2014-06-27 | 2014-10-08 | 无锡慧眼电子科技有限公司 | People counting method based on clustering method |
| CN105354610A (en) * | 2014-08-18 | 2016-02-24 | 无锡慧眼电子科技有限公司 | Random Hough transform-based people counting method |
| CN105635696A (en) * | 2016-03-22 | 2016-06-01 | 南阳理工学院 | Statistical method and device |
| CN105844649A (en) * | 2016-04-12 | 2016-08-10 | 中国科学院长春光学精密机械与物理研究所 | Statistical method, apparatus and system for the quantity of people |
| CN106228137A (en) * | 2016-07-26 | 2016-12-14 | 广州市维安科技股份有限公司 | A kind of ATM abnormal human face detection based on key point location |
| CN106339673A (en) * | 2016-08-19 | 2017-01-18 | 中山大学 | ATM identity authentication method based on face recognition |
| CN106951820A (en) * | 2016-08-31 | 2017-07-14 | 江苏慧眼数据科技股份有限公司 | Passenger flow statistical method based on annular template and ellipse fitting |
| CN106997459A (en) * | 2017-04-28 | 2017-08-01 | 成都艾联科创科技有限公司 | A kind of demographic method split based on neutral net and image congruencing and system |
| CN107133607A (en) * | 2017-05-27 | 2017-09-05 | 上海应用技术大学 | Demographics' method and system based on video monitoring |
| CN108509913A (en) * | 2018-03-30 | 2018-09-07 | 世纪美映影院技术服务（北京）有限公司 | A kind of occupancy statistical method |
| CN113160268A (en) * | 2021-05-13 | 2021-07-23 | 深圳龙岗智能视听研究院 | Event camera-based method for counting number of moving objects |
| CN113723500A (en) * | 2021-08-27 | 2021-11-30 | 四川启睿克科技有限公司 | Image data expansion method based on feature similarity and linear smoothing combination |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080130952A1 (en) * | 2002-10-17 | 2008-06-05 | Siemens Corporate Research, Inc. | method for scene modeling and change detection |
| CN102147869A (en) * | 2011-03-31 | 2011-08-10 | 上海交通大学 | Pedestrian detection method based on foreground analysis and pattern recognition |
-
2012
- 2012-08-03 CN CN2012102747424A patent/CN102831472A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080130952A1 (en) * | 2002-10-17 | 2008-06-05 | Siemens Corporate Research, Inc. | method for scene modeling and change detection |
| CN102147869A (en) * | 2011-03-31 | 2011-08-10 | 上海交通大学 | Pedestrian detection method based on foreground analysis and pattern recognition |
Non-Patent Citations (3)
| Title |
|---|
| 顾德军: "基于视频图像处理的人数自动统计技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, no. 6, 15 June 2011 (2011-06-15), pages 1 - 37 * |
| 顾德军等: "一种基于人头特征的人数统计方法研究", 《机械制造与自动化》, 20 August 2010 (2010-08-20), pages 134 - 138 * |
| 黄廉真等: "一种新的高速圆形匹配算法", 《哈尔滨工业大学学报》, vol. 44, no. 7, 30 July 2012 (2012-07-30), pages 87 - 91 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104091351A (en) * | 2014-06-27 | 2014-10-08 | 无锡慧眼电子科技有限公司 | People counting method based on clustering method |
| CN104091351B (en) * | 2014-06-27 | 2017-03-15 | 江苏慧眼数据科技股份有限公司 | Number method of counting based on clustering procedure |
| CN105354610A (en) * | 2014-08-18 | 2016-02-24 | 无锡慧眼电子科技有限公司 | Random Hough transform-based people counting method |
| CN105635696A (en) * | 2016-03-22 | 2016-06-01 | 南阳理工学院 | Statistical method and device |
| CN105844649A (en) * | 2016-04-12 | 2016-08-10 | 中国科学院长春光学精密机械与物理研究所 | Statistical method, apparatus and system for the quantity of people |
| CN106228137A (en) * | 2016-07-26 | 2016-12-14 | 广州市维安科技股份有限公司 | A kind of ATM abnormal human face detection based on key point location |
| CN106339673A (en) * | 2016-08-19 | 2017-01-18 | 中山大学 | ATM identity authentication method based on face recognition |
| CN106951820B (en) * | 2016-08-31 | 2019-12-13 | 江苏慧眼数据科技股份有限公司 | Passenger flow statistical method based on annular template and ellipse fitting |
| CN106951820A (en) * | 2016-08-31 | 2017-07-14 | 江苏慧眼数据科技股份有限公司 | Passenger flow statistical method based on annular template and ellipse fitting |
| CN106997459A (en) * | 2017-04-28 | 2017-08-01 | 成都艾联科创科技有限公司 | A kind of demographic method split based on neutral net and image congruencing and system |
| CN106997459B (en) * | 2017-04-28 | 2020-06-26 | 成都艾联科创科技有限公司 | People counting method and system based on neural network and image superposition segmentation |
| CN107133607A (en) * | 2017-05-27 | 2017-09-05 | 上海应用技术大学 | Demographics' method and system based on video monitoring |
| CN107133607B (en) * | 2017-05-27 | 2019-10-11 | 上海应用技术大学 | Demographics' method and system based on video monitoring |
| CN108509913A (en) * | 2018-03-30 | 2018-09-07 | 世纪美映影院技术服务（北京）有限公司 | A kind of occupancy statistical method |
| CN108509913B (en) * | 2018-03-30 | 2021-03-02 | 世纪美映影院技术服务（北京）有限公司 | Indoor people counting method |
| CN113160268A (en) * | 2021-05-13 | 2021-07-23 | 深圳龙岗智能视听研究院 | Event camera-based method for counting number of moving objects |
| CN113723500A (en) * | 2021-08-27 | 2021-11-30 | 四川启睿克科技有限公司 | Image data expansion method based on feature similarity and linear smoothing combination |
| CN113723500B (en) * | 2021-08-27 | 2023-06-16 | 四川启睿克科技有限公司 | Image data expansion method based on combination of feature similarity and linear smoothing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102831472A (en) | People counting method based on video flowing image processing | |
| CN101957997B (en) | Regional average value kernel density estimation-based moving target detecting method in dynamic scene | |
| CN104123544B (en) | Anomaly detection method and system based on video analysis | |
| CN107133969B (en) | A kind of mobile platform moving target detecting method based on background back projection | |
| CN102867177B (en) | A kind of demographic method based on gradation of image coupling | |
| CN101751678B (en) | Method and device for detecting violent crowd movement | |
| CN101794435B (en) | Binary image noise-reduction method based on integral graph and binary image processing system | |
| CN103985108B (en) | Method for multi-focus image fusion through boundary detection and multi-scale morphology definition measurement | |
| CN102855465B (en) | A kind of tracking of mobile object | |
| CN101470809B (en) | Moving object detection method based on expansion mixed gauss model | |
| CN106023245B (en) | Moving target detecting method under the static background measured based on middle intelligence collection similarity | |
| CN102930248A (en) | Crowd abnormal behavior detection method based on machine learning | |
| CN102855466B (en) | A kind of demographic method based on Computer Vision | |
| CN102324016A (en) | Statistical method for high-density crowd flow | |
| CN107742291B (en) | Defect detection method and device for photovoltaic glass | |
| CN101527044A (en) | Automatic segmenting and tracking method of multiple-video moving target | |
| CN111967394B (en) | Forest fire smoke root node detection method based on dynamic and static grid fusion strategy | |
| CN101483763A (en) | Digital video processing method oriented to social security monitoring and apparatus thereof | |
| CN102867349A (en) | People counting method based on elliptical ring template matching | |
| Karpagavalli et al. | Estimating the density of the people and counting the number of people in a crowd environment for human safety | |
| CN103077380B (en) | A kind of demographic method based on video and device | |
| CN101324958A (en) | Method and apparatus for tracking object | |
| CN105405153A (en) | Intelligent mobile terminal anti-noise interference motion target extraction method | |
| CN104376580B (en) | The processing method of non-interesting zone issue in a kind of video frequency abstract | |
| Ye | A robust method for counting people in complex indoor spaces |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20121219 |