CN102110295A - Real-time region detection method - Google Patents
Real-time region detection method Download PDFInfo
- Publication number
- CN102110295A CN102110295A CN2009102451207A CN200910245120A CN102110295A CN 102110295 A CN102110295 A CN 102110295A CN 2009102451207 A CN2009102451207 A CN 2009102451207A CN 200910245120 A CN200910245120 A CN 200910245120A CN 102110295 A CN102110295 A CN 102110295A
- Authority
- CN
- China
- Prior art keywords
- region
- target
- sensitive parameter
- algorithm
- warning region
- 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
Abstract
The invention belongs to the technical field of video image processing, and particularly discloses a real-time region intrusion detection method. In the method, based on a related tracking algorithm, the space region positions in a monitored scene are classified and labeled by using a region classification labeling thought to acquire sensitive parameters of region states, and the sensitive parameter of the space position of a moving target in the monitored scene is used as basis for judging whether the target is intruded. The algorithm has the ingenuity that the classification labeling of the region space positions does not aim at the image data but a corresponding newly-developed data block. Therefore, the interference resistance is greatly increased, more importantly, the classification labeling of the region can be executed only when the region is set or changed, so region marks of each frame do not need to be updated, and the efficiency of algorithm execution is greatly improved. Compared with other homogeneous algorithms, the algorithm avoids complex judgment between the moving region and a warning region; and whether the target is intruded is judged only by using the sensitive parameters of the space position of the moving target in the warning region as basis, so the thought is simple, and the algorithm is easily implemented and has good robustness.
Description
Technical field
The invention belongs to field of video image processing, especially a kind of real-time method for detecting area.
Background technology
Increasingly mature along with the continuous development of intelligent monitoring technology and image processing techniques, original employing manpower carries out the suspicious object monitoring and has not satisfied demand, and can remedy the problem of shortage of manpower to a great extent based on the intelligent safety and defence system of technology such as artificial intelligence and video analysis.The intelligent safety and defence system advantage is that the security personnel need not to go on patrol on the spot, just can obtain on-the-spot actual conditions by the video that supervisory system is passed back, so that make a policy rapidly, therefore wide development space and enormous and latent market is arranged.Warning region intrusion detection, regional stream of people statistics, regional crowd density detection etc. are one important application aspects.The warning region intrusion detection is meant to be discerned the target of swarming into warning region, and regional stream of people's statistics is meant that the turnover stream of people to a certain zone in certain period adds up, and regional crowd density detection is meant to be added up the target in a certain zone.They mainly are to utilize image process method, automatically detect the target invasion in reality scene, make computing machine have certain and understand and the ability of analysis video, thereby target invaded to report to the police or invasion added up reports to the police when satisfying a certain condition.
Traditional method for detecting area, the general judgement that all is based on complexity between moving target and the warning region realizes, even for simple rectangular area, also need a large amount of judgements, so these method calculated amount are big, the thought complexity, especially for picture concave polygon zone, efficient is lower.
Summary of the invention
Based on above problem, the present invention proposes to use the thought of territorial classification mark that the sensitive parameter that key words sorting obtains zone state is carried out in the locus in zone in the monitoring scene, serves as according to judging that whether target invade with the sensitive parameter of locus, moving target region.And,,, just do not need to upgrade regional sensitive parameter as long as warning region does not change so only need to set or obtain the area sensitive parameter when changing in the zone because the key words sorting in zone is not directly at view data.Algorithm thought is simple, is easy to realize, and is all good for the regional robustness of arbitrary polygon (protruding, concave polygon), and very high execution efficient is arranged.
For reaching above purpose, the present invention adopts following technical scheme.
A kind of real-time method for detecting area.Specifically may further comprise the steps:
E, monitoring scene background are set up and are upgraded;
F, the relevant motion detection algorithm of utilization obtain moving target;
G, utilization related objective track algorithm tracking target obtain the current locus of target coordinate;
H, zone are detected;
It is characterized in that, the regional detection algorithm of step D is at first to region-wide sheltering in the monitoring scene, then masking regional and warning region are carried out key words sorting, to obtain the sensitive parameter of warning region locus state, the basis for estimation of whether invading as target with the sensitive parameter of moving target region realizes the zone detection at last.
And, to the region-wide calculating of sheltering of monitoring scene, can be to open up one with the identical array of image size, to its corresponding pixel points assignment.
And, masking regional and warning region are carried out key words sorting, with the calculating of the sensitive parameter that obtains warning region locus state, can be following steps:
(1) number of vertex and the apex coordinate thereof of mark polygonal region desired in input.
(2) need to determine mark warning region vertical direction scope, promptly calculate maximal value and the minimum value of y in all polygon vertex coordinates, with this process range as sweep trace.
(3) every in process range sweep trace is set up orderly limit table.
(4), repeat the following step to every in process range sweep trace:
A sets up the effective edge table of current scan line with orderly limit table;
B takes out an antinode successively from the effective edge table, mark is carried out in the warning region locus in two intersection points obtain the warning region sensitive parameter
C is that next bar sweep trace upgrades the effective edge table, promptly increases the x value of intersection point and the limit that deletion is no longer intersected;
D rearranges the effective edge table.
And, for the basis for estimation of whether invading as target with the sensitive parameter of moving target region, realize the calculating that the zone is detected, can be the sensitive parameter of judging locus, moving target region, if it equals the sensitive parameter of warning region, then represent the target invasion.
Advantage of the present invention and good effect are:
(1) carries out the basis for estimation that mark is used as the target invasion by locus state, significantly reduced number of times and the complexity judged, and have identical execution efficient for concavo-convex polygonal region to the specific region;
(2) by opening up the internal memory of a view data size, can make the key words sorting in zone be independent of view data itself, therefore only need carry out once, as long as warning region does not change for the key words sorting in zone, just need not to upgrade the key words sorting in zone, improved efficiency of algorithm greatly.
Description of drawings
Fig. 1 is that overview flow chart is detected in the zone in real time;
Fig. 2 uses the process flow diagram that the present invention realizes a specific embodiment;
Fig. 3 is a regional polygonal limit table;
Fig. 4 is two kinds of strategies that shorten on the limit;
Embodiment
Below in conjunction with accompanying drawing embodiments of the invention are elaborated: present embodiment has provided detailed process being to realize under the prerequisite with the technical solution of the present invention, but protection scope of the present invention is not limited to following embodiment.
In the present embodiment certain video sequence is carried out the zone and detect, Fig. 1 is that overview flow chart is detected in the zone, and Fig. 2 uses the process flow diagram that the present invention realizes a specific embodiment
As shown in Figure 1, the zone is detected and is generally comprised following steps:
A, input video frame are to video monitoring scene background modeling and renewal;
B, the relevant motion detection algorithm of utilization obtain moving target;
C, utilization related objective track algorithm tracking target obtain the current locus of target coordinate;
D, zone are detected;
As shown in Figure 2, present embodiment is used the present invention and is realized that regional detection module specifically comprises the steps: step 201, sets up the piece image identical with original image size, and its pixel is composed an initial value 0, realizes sheltering region-wide:
MR(x，y)＝0，
(x，y)＝{(x
i，y
i)|i＝1，2L?N}
MR (x is the pixel horizontal coordinate for x, y) expression area sensitive parameter, and y is the pixel vertical coordinate, and N is the pixel sum,
Step 202, the number of vertex and the summit pixel coordinate thereof of mark polygonal region desired in input.
Step 203 need to determine mark warning region vertical direction scope, promptly calculates maximal value and the minimum value of y in all polygon vertex pixel coordinates, with this process range as sweep trace.
(x，1/k，ymax，next)
Its construction process is as follows:
A at first constructs a vertical chained list, and the length of chained list is the maximum scan line number that polygon occupied, and each node of chained list is called a bucket, each bar sweep trace that then corresponding polygon covers.As shown in Figure 3.
B goes into the message sense on every limit and the corresponding bucket of this limit minimum vertical coordinate ymin place.That is to say that if the low end points on certain limit is ymin, then this limit just is placed in the corresponding sweep trace bucket.
The data on every limit of c form a node.
The ascending ordering of x|ymin (y hour x value) press on some limits in the same bucket of d, and is equal as if x|ymin, then according to the ascending ordering of 1/k.
In addition, for solving when sweep trace and polygonal summit are crossing, the problem when intersection point is counted 1 (this algorithm does not have problems when 2 or 0) can be the summit of 1 intersection point to separate those accrued with polygonal some limit shortening.For example can handle non-horizontal sides on the whole Polygonal Boundary with the clockwise direction of appointment.When handling every limit, detect, determine whether this limit and the non-horizontal sides of next bar have the end points y value of monotone increasing or monotone decreasing, promptly whether exist the ymax on certain bar limit to equal the ymin on another limit.If have, the sweep trace that the limit of ymax can be shortened to guarantee the public vertex by connecting two limits only has an intersection point to generate, and is about to ymax and is updated to ymax-1.Certainly also the limit of ymin can be shortened, be about to ymin and be updated to ymin+1, but this mode need calculate respective x |, more complicated behind the ymin+1 participating in the ordering of bucket formula.Fig. 4 illustrates two kinds of strategies that shorten the limit.
A sets up the effective edge table of current scan line with orderly limit table;
The structure of effective edge table and in order the limit epiphase with, effective edge table storage be and limit that current scan line intersects, initialization effective edge table, the effective edge table is put sky, the limit in the ordered list that first is not empty merges with effective table;
B takes out an antinode (x successively from the effective edge table
1, y), (x
2, y), the sensitive parameter of the warning region locus in two intersection points is put 1, obtain the warning region sensitive parameter like this:
MR (x, y)=1, (x
1＜x＜x
2, y represents present level sweep trace vertical coordinate)
C is that next bar sweep trace upgrades the effective edge table, promptly increases the x value of intersection point and the limit that deletion is no longer intersected;
For next bar sweep trace, x (x is the horizontal coordinate of the intersection point on present level sweep trace and limit) is updated to x+1/k, merges the limit in the corresponding bucket of next bar sweep trace in the table of orderly limit simultaneously, be inserted in order in the effective edge table, and the limit of deletion y=ymax forms new effective edge table.
D rearranges the effective edge table.
Step 206, judgement moving target locus coordinate (x, sensitive parameter y), if equal 1, then expression invasion:
Q is 1, and expression can be mated with the sensitive parameter of warning region, and the invasion of expression target is 0 not invasion.
By using the embodiment of institute of the present invention specific implementation, proved absolutely that the present invention proposes to use the thought of territorial classification mark that the sensitive parameter that key words sorting obtains zone state is carried out in the locus in zone in the monitoring scene, with the sensitive parameter of locus, moving target region serves as according to judging that the thought whether target invades is entirely true, and can specific implementation.By the test result of embodiment, illustrated that also using the present invention realizes that the regional effect that detects is can satisfy the actual monitored scene fully to require, and particularly importantly had good execution efficient.
It is emphasized that at last; embodiment of the present invention is illustrative; rather than it is determinate; therefore the present invention is not limited to the embodiment described in the embodiment; every other embodiments that drawn by those skilled in the art's technical scheme according to the present invention belong to the scope of protection of the invention equally.
Claims (4)
1. real-time method for detecting area.Specifically may further comprise the steps:
A, monitoring scene background are set up and are upgraded;
B, the relevant motion detection algorithm of utilization obtain moving target;
C, utilization related objective track algorithm tracking target obtain the current locus of target coordinate;
D, zone are detected;
It is characterized in that, the regional detection algorithm of step D is at first to region-wide sheltering in the monitoring scene, then masking regional and warning region are carried out key words sorting, to obtain the sensitive parameter of warning region locus state, the basis for estimation of whether invading as target with the sensitive parameter of moving target region realizes the zone detection at last.
2. a kind of real-time method for detecting area according to claim 1 to the region-wide calculating of sheltering of monitoring scene, can be to open up one with the identical array of image size, to its corresponding pixel points assignment.
3. a kind of real-time method for detecting area according to claim 1 carries out key words sorting to masking regional and warning region, with the calculating of the sensitive parameter that obtains warning region locus state, can be following steps:
(1) number of vertex and the apex coordinate thereof of mark polygonal region desired in input.
(2) need to determine mark warning region vertical direction scope, promptly calculate maximal value and the minimum value of y in all polygon vertex coordinates, with this process range as sweep trace.
(3) every in process range sweep trace is set up orderly limit table.
(4), repeat the following step to every in process range sweep trace:
A sets up the effective edge table of current scan line with orderly limit table;
B takes out an antinode successively from the effective edge table, mark is carried out in the warning region locus in two intersection points obtain the warning region sensitive parameter;
C is that next bar sweep trace upgrades the effective edge table, promptly increases the x value of intersection point and the limit that deletion is no longer intersected;
D rearranges the effective edge table.
4. a kind of real-time method for detecting area according to claim 1, for the basis for estimation of whether invading as target with the sensitive parameter of moving target region, realize the calculating that the zone is detected, it can be the sensitive parameter of judging locus, moving target region, if it equals the sensitive parameter of warning region, then represent the target invasion.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102451207A CN102110295A (en) | 2009-12-28 | 2009-12-28 | Real-time region detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009102451207A CN102110295A (en) | 2009-12-28 | 2009-12-28 | Real-time region detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102110295A true CN102110295A (en) | 2011-06-29 |
Family
ID=44174442
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009102451207A Pending CN102110295A (en) | 2009-12-28 | 2009-12-28 | Real-time region detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102110295A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101835035A (en) * | 2010-06-04 | 2010-09-15 | 天津市亚安科技电子有限公司 | Regional invasion real-time detection method |
| CN103903269A (en) * | 2014-04-10 | 2014-07-02 | 公安部第三研究所 | Structural description method and system of dome camera monitor video |
-
2009
- 2009-12-28 CN CN2009102451207A patent/CN102110295A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101835035A (en) * | 2010-06-04 | 2010-09-15 | 天津市亚安科技电子有限公司 | Regional invasion real-time detection method |
| CN101835035B (en) * | 2010-06-04 | 2014-02-26 | 天津市亚安科技股份有限公司 | Regional invasion real-time detection method |
| CN103903269A (en) * | 2014-04-10 | 2014-07-02 | 公安部第三研究所 | Structural description method and system of dome camera monitor video |
| CN103903269B (en) * | 2014-04-10 | 2017-10-27 | 公安部第三研究所 | The description method and system of ball machine monitor video |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101835035B (en) | Regional invasion real-time detection method | |
| Albiol et al. | Detection of parked vehicles using spatiotemporal maps | |
| CN104303193B (en) | Target classification based on cluster | |
| AU2009243442B2 (en) | Detection of abnormal behaviour in video objects | |
| CN103069434B (en) | For the method and system of multi-mode video case index | |
| CN106504270B (en) | Method and device for displaying target object in video | |
| US8731241B2 (en) | Activity mapping system | |
| CN101470809B (en) | Moving object detection method based on expansion mixed gauss model | |
| EP2956891B1 (en) | Segmenting objects in multimedia data | |
| Venetianer et al. | Performance evaluation of an intelligent video surveillance system–A case study | |
| CN105631418A (en) | People counting method and device | |
| CN101167086A (en) | Human detection and tracking for security applications | |
| CN101986348A (en) | Visual target identification and tracking method | |
| CN111462155B (en) | Motion detection method, device, computer equipment and storage medium | |
| CN110533955A (en) | A kind of method, terminal device and the computer readable storage medium on determining parking stall | |
| Li et al. | Decoupled appearance and motion learning for efficient anomaly detection in surveillance video | |
| CN111582060B (en) | Automatic line drawing perimeter alarm method, computer equipment and storage device | |
| CN103577875A (en) | CAD (computer-aided design) people counting method based on FAST (features from accelerated segment test) | |
| CN106600628A (en) | Target object identification method and device based on infrared thermal imaging system | |
| CN108647587A (en) | Demographic method, device, terminal and storage medium | |
| Zhang et al. | Counting vehicles in urban traffic scenes using foreground time‐spatial images | |
| CN104159088A (en) | System and method of remote monitoring of intelligent vehicle | |
| Zhang et al. | New mixed adaptive detection algorithm for moving target with big data | |
| CN105469054A (en) | Model construction method of normal behaviors and detection method of abnormal behaviors | |
| CN111797831A (en) | BIM and artificial intelligence based parallel abnormality detection method for poultry feeding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110629 |