CN101127887A - Intelligent vision monitoring method and device - Google Patents
Intelligent vision monitoring method and device Download PDFInfo
- Publication number
- CN101127887A CN101127887A CNA200710035636XA CN200710035636A CN101127887A CN 101127887 A CN101127887 A CN 101127887A CN A200710035636X A CNA200710035636X A CN A200710035636XA CN 200710035636 A CN200710035636 A CN 200710035636A CN 101127887 A CN101127887 A CN 101127887A
- Authority
- CN
- China
- Prior art keywords
- target
- circuit board
- processing circuit
- image
- embedded processing
- 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
- Closed-Circuit Television Systems (AREA)
- Image Analysis (AREA)
Abstract
The utility model discloses an intelligent visual monitoring method and device; wherein, the method achieves the intelligent visual monitoring by means of detecting a plurality of mobile targets and tracing the mobile targets; the device comprises a main control computer, a panoramic image acquisition device, a ball-shaped high-speed dome camera, a box body positioned on the monitoring site, an embedded processing circuit board and a power supply unit which are fixed in the box body. The panoramic image acquisition device and the ball-shaped video camera are connected with the embedded processing circuit board via a video output cable. The embedded processing circuit board is connected with the main control computer via the network connection line. The main control computer detects a plurality of mobile targets in the picture using the corresponding picture computing method and realizes the positioning and tracing of the mobile targets in three-dimensional space in fixed settings according to specific priority level control strategy. The utility model can automatically trace a plurality of mobile targets in fixed settings, and can be widely used in the fields of the security, the machine vision, the telemonitoring and other fields. The utility model has the advantage of good versatility.
Description
Technical field
The present invention is mainly concerned with the vision monitoring field, refers in particular to a kind of intelligent vision monitoring method and device.
Background technology
Along with the development of computer and image sensor technologies,, be widely used in different social sectors already by camera supervised dynamic scene.Traffic monitoring on from the guard monitor of community and critical facility to city and highway, from the intellectual weapon that detects of military target, video camera plays important effect as the extension of human vision.Intelligent vision monitoring is exactly the method that will use computer vision, under the situation that does not need human intervention, by the image sequence of video camera recording is analyzed automatically, realization is to location, identification and the tracking of target in the dynamic scene, and analyze and judge the behavior of target on this basis, can when taking place, abnormal conditions in time make a response again thereby accomplish to finish daily management.Intelligent vision monitoring relates generally to contents such as camera calibration, object identification, motion segmentation and tracking, multiple-camera fusion, high-level semantic understanding, is the forward position research direction of computer vision field.It is with a wide range of applications and huge potential economic worth, has caused many scientific research institutions and researcher's great interest.The U.S.'s 9.11 incidents more make intelligent vision monitoring seem urgent day by day, and in important project is also listed it one after another in countries in the world therefore.
The dynamic scene vision monitoring is an emerging application direction of computer vision field.Vision monitoring is different from traditional supervisory control system and is that it is intelligent.Briefly, not only replace human eye, and replace people, contributor, finish monitoring or control task, thereby alleviate people's burden with computer with video camera.This intelligent vision monitoring device is gathered the vision signal of two video cameras, finish monitoring on-the-spot a plurality of motion target detection and tracking by embedded processing device and main control computer, all processes does not need manual operation, has realized the auto-tracking shooting to moving target.Have high using value and application prospects.
Summary of the invention
The technical problem to be solved in the present invention just is: at the technical problem of existing moving target vision monitoring existence, the invention provides a kind of automaticity height, tracking velocity is fast, precision is high, applied widely intelligent vision monitoring device, thereby realized in moving scene track up to a plurality of moving targets, the present invention can be widely used in the vision security protection, machine vision, fields such as vision-based detection are a kind of intelligent vision monitoring devices with highly versatile.
For solving the problems of the technologies described above, the solution that the present invention proposes is: a kind of intelligent vision monitoring method is characterized in that step is:
(1), multiple mobile object detects: at first first two field picture after the input video denoising is made as initial background, to the t moment image I from the image scanner input
t(x, the y) noise that adopts median filter filtering part to be caused by transducer is with itself and t moment background image H
t(x y) subtracts each other and with binaryzation as a result, handles the back detection through morphologic filtering and whether have moving target, and renewal function obtains H according to testing result
T+1(x y) as the t+1 background image constantly after upgrading, thereby realizes the background real-time update, and described renewal function is
N represents the detected target sum of t frame, K in the formula
iBe i target, a, b are update coefficients; Then background image and present image are subtracted each other and threshold value turns to bianry image, further adopt morphology, cut apart, extract moving target by the connected region in neighbours territory with " spot " noise remove in the binary image, and with area less than threshold value T
aTarget cast out, obtain detected moving target in the present frame, and annotate with target boundary rectangle collimation mark;
(2), motion target tracking: at first use Region Segmentation method handle by the motion target area rectangle frame mark that obtains in the step (1), calculate the centre of form coordinate of that maximum rectangular area then, calculate the deviation of the target centre of form and picture centre, control the action of image scanner according to this deviation, target is remained near the central authorities of visual field, this is a feedback control procedure, constantly detects the centre of form of target and the deviation of picture centre, up to the central authorities of target in the visual field.
The Region Segmentation step of moving target is as follows in the described step:
(1), initialization: remove scanning flag, the right margin of left=monitor area, the left margin of right=monitor area, number of targets ir=0;
(2), from top to bottom (y) from left to right (x) line by line scan do not beat the target picture point (x, y), if find to begin continuously from it some when not beating the target point and belonging to variations (target) zone, note top=y gives a some mark, otherwise commentaries on classics (6);
(3), in not beating target point, find the left margin point x of target left from this impact point, if x＜left, then left:=x; Find right margin point x more in the same way to the right, if x＞right, then right=x; And to the some mark that scans;
(4), calculate mid point: x:=(left+right)/2, and move down a row y:=y+1, (x y), gives a some mark to invocation point; Change (3) if this point belongs to impact point, otherwise note bottom:=y-1 obtains a fresh target,, change (2) if target area is given up it less than the minimum value that allows;
(5), relatively, if mutual distance is then incorporated this target into respective objects in allowing the scope (as 3 pixels) that merges, otherwise add object queue, ir=ir+1 changes (2) with other target area in fresh target and the object queue;
(6), the zone of mutual distance in the object queue in allowing the merging scope merged, object queue to the end.
A kind of intelligent vision monitoring device, it is characterized in that: it comprises main control computer, panoramic picture harvester, is with the ball-shaped camera of high speed The Cloud Terrace, is placed on the on-the-spot box body of monitoring and is fixed on box body interior embedded processing circuit board and power subsystem, described panoramic picture harvester links to each other with the embedded processing circuit board by the video output cable with ball-shaped camera, and the embedded processing circuit board links to each other with main control computer by network connection.
Described panoramic picture harvester adopts rifle type video camera, and the embedded processing circuit board is by the motion state of the ball-shaped camera of RS485 serial bus control band high speed The Cloud Terrace.
Described embedded processing circuit board inserts the local area network (LAN) at main control computer place by network connection.
Described panoramic picture harvester directly links to each other with the embedded processing circuit board with the RS485 universal serial bus by the video output cable with the ball-shaped camera of band high speed The Cloud Terrace.
Compared with prior art, advantage of the present invention just is:
1, intelligent vision monitoring method of the present invention and device have been realized the high degree of intelligence vision monitoring, and most of existing supervising device can only realize monitoring on-the-spot shooting, have only small part can realize track up to single moving target.Used the multiple mobile object detection algorithm in this device and adopted certain priority control strategy, so can realize monitoring the track up of on-the-spot a plurality of moving targets.
2, existing vision monitoring apparatus can only be realized the location to the two dimensional surface coordinate of moving target.Intelligent vision monitoring method of the present invention and device use two visual imaging devices, can utilize two width of cloth space plane images of the on-the-spot same asynchronism(-nization) angle of monitoring that photographs to produce three dimensional space coordinate system, can realize demarcation the three dimensional space coordinate of monitoring on-the-spot moving target.
3, existing vision monitoring apparatus such as need connect the PC main frame and also need to be the host configuration image pick-up card, and use vision cable to be connected to image collecting device.Though had the image collecting device that can work alone to occur, can only realize simple single motion target tracking.Intelligent vision monitoring method of the present invention and device are finished image collecting function by the embedded processing circuit board, and main control computer is finished a plurality of motion target detection and identification, and according to certain priority level motion target tracking is taken.Main control computer and embedded processing circuit two parts have been realized optimum organization.
4, existing visual monitor system need use vision cable to be connected with video server or hard disk DVR etc., and when image collecting device increased, wiring was very loaded down with trivial details.The embedded processing circuit board inserts main control computer place local area network (LAN) by network connection in intelligent vision monitoring method of the present invention and the device, and main control computer is by the network connection access to LAN.This device has made full use of network technology, makes the interconnection between subsystem simple and convenient.
Description of drawings
Fig. 1 is the schematic flow sheet of intelligent vision monitoring method of the present invention;
Fig. 2 is a single moving object detection algorithm flow schematic diagram of the present invention;
Fig. 3 is a single motion target tracking algorithm flow schematic diagram of the present invention;
Fig. 4 is the frame structure schematic diagram of intelligent vision monitoring device of the present invention;
Fig. 5 is a core processor circuit principle schematic in the embedded processing circuit board of the present invention;
Fig. 6 is a Network Transmission circuit theory intention in the embedded processing circuit board of the present invention;
Fig. 7 is a power circuit principle intention in the embedded processing circuit board of the present invention;
Fig. 8 is memory and an image acquisition circuit principle intention in the embedded processing circuit board of the present invention.
Marginal data
1, box body 2, embedded processing circuit board
3, power subsystem 4, panoramic picture harvester
5, ball-shaped camera 6, support
7, support 8, main control computer
9, video output cable 10, network connection
11, RS485 universal serial bus 12, network connection
Embodiment
Below with reference to the drawings and specific embodiments the present invention is described in further details.
As Fig. 1, Fig. 2 and shown in Figure 3, the present invention is directed to static at gamma camera is to realize under the fixed background that multiple mobile object detects, obtaining and upgrading and adopt background subtraction and time difference to combine to detect the method for target background images.In the static monitoring of camera (fixed scene), as adopt background subtraction method detection moving target can obtain the complete movable information of target, particularly at target travel accuracy in detection height under the situation more slowly, but the adaptability that background dynamics is changed is low, as detecting target with the time difference method, then can improve adaptability to the environment dynamic change, but to the object variations difficult accurately detection of part slowly.The present invention is fully analyzing on the basis of first and second class detection method, and the method that adopts background subtraction and time difference to combine can effectively detect moving target and can adapt to the dynamic change of monitoring scene.Intelligent vision monitoring method of the present invention is divided into two big steps: multiple mobile object detects and motion target tracking.
The detailed process that detects of multiple mobile object as shown in Figure 2: first two field picture after the input video denoising is made as initial background and (supposes that first two field picture does not comprise moving target, this hypothesis tallies with the actual situation, as input is that chromatic image is converted into gray scale image earlier), to t moment image I from the image scanner input
t(x, the y) noise that adopts median filter filtering part to be caused by transducer is with itself and t moment background image H
t(x y) subtracts each other and with binaryzation as a result, handles the back detection through morphologic filtering and whether have moving target, and renewal function obtains H according to testing result
T+1(x y) as the t+1 background image constantly after upgrading, thereby realizes the background real-time update.Renewal function:
N represents the detected target sum of t frame, K in the formula
iBe i target.Update coefficients a, b realize the context update under the different situations: H
T+1(x is y) corresponding to I
t(x, the area update coefficient a that comprises moving target in y) is about 1, and a is more near 1, and context update is slow more, i.e. H
T+1(x, y) in this range image fundamental sum H
t(x, y) the corresponding region image is consistent, and this is that background is almost constant in this zone because only reflect target travel in this zone of present frame; H
T+1(x is y) corresponding to I
t(x does not comprise motion target area in y), can think that this regional background is in the dynamic change, and update coefficients b is made as much smaller than 1, and b approaches 0 more, and context update is fast more, i.e. H
T+1(x, y) in this range image fundamental sum I
t(x, y) the corresponding region image is consistent.As seen adopt renewal function energy real-time update background, need not set up background model, can adapt to light variation in the scene preferably, improve the accuracy of target detection through one period initialization time.
Subtracting each other also by background image and present image, threshold value turns to bianry image, because making target cut apart, the slight change of visual field is subjected to the noise effect in the primary data very big, therefore can adopt morphology earlier " open back " closing " computing with " spot " noise remove in the binary image; the connected region by the neighbours territory is cut apart; extract moving target, and with area less than threshold value T
a(establish T in the experiment
a=30) target is cast out, and obtains detected moving target in the present frame, and annotates with target boundary rectangle collimation mark.
The detailed process of motion target tracking as shown in Figure 3: detect a plurality of moving targets of monitoring scene in host PC after, also need to realize tracing and monitoring to single moving target.Tracking to single moving target is firm and hard existing by flush type circuit.The present invention proposes a kind of motion target tracking method fast and effectively and on the flush type circuit plate, realize: at first with the Region Segmentation method the motion target area rectangle frame mark that transmits by host PC, calculate the centre of form coordinate of that maximum rectangular area then, calculate the deviation of the target centre of form and picture centre, control the action of camera pan-tilt according to this deviation, target is remained near the central authorities of visual field.This is a feedback control procedure, constantly detects the centre of form of target and the deviation of picture centre, up to the central authorities of target in the visual field.
For the Region Segmentation of moving target, the present invention adopt a kind of simply, Region Segmentation Algorithm fast: mid point guiding scan method becomes the rectangular area with image segmentation.Algorithm steps is as follows:
(1) initialization: remove scanning flag, the right margin of left=monitor area, the left margin of right=monitor area, number of targets ir=0
(2) from top to bottom (y) from left to right (x) line by line scan do not beat the target picture point (x, y), if find to begin continuously from it some when not beating the target point and belonging to variations (target) zone, note top=y gives a some mark, otherwise commentaries on classics (6);
(3) in not beating target point, find the left margin point x of target left from this impact point, if x＜left, then left:=x; Find right margin point x more in the same way to the right, if x＞right, then right=x; And to the some mark that scans;
(4) calculate mid point: x:=(left+right)/2, and move down a row y:=y+1, (x y), gives the some mark to invocation point.Change (3) if this point belongs to impact point, otherwise note bottom:=y-1 obtains a fresh target,, change (2) if target area is given up it less than the minimum value that allows.
(5) relatively, if mutual distance is then incorporated this target into respective objects in allowing the scope (as 3 pixels) that merges, otherwise add object queue, ir=ir+1 changes (2) with other target area in fresh target and the object queue;
(6) zone of mutual distance in the object queue in allowing the merging scope merged, get object queue to the end.
Above-mentioned mid point guiding scanning dividing method only needs image of scanning, and computation complexity is the linear scale of image size.
For the tracking of moving target, in the present invention, to choose a maximum regular-shape motion target image, and target image is regarded as a uniform thin plate of density as target image, the center of gravity of obtaining like this is called the centre of form of target image.
The position of the centre of form is the definite point of on the targeted graphical, and when targeted attitude changed, this change in location was less, so follow the tracks of when following the tracks of with the centre of form more steady; And the anti-clutter interference capability is strong, is a kind of method with the most use in the TV tracker system.
The definition of the centre of form is:
Wherein,
Wherein
It is target centre of form coordinate; Integral domain Ω is whole object-image region.
Because the result of binaryzation, target image Ω is " 1 " with interior signal amplitude, and the signal amplitude beyond the target image is " 0 ", like this, can be rewritten into centre of form solution formula
Wherein,
Wherein: when (x, when y) belonging to the target area, B (x, y)=1; As (x, B when y) not belonging to the target area (x.y)=0; A, b, c, d is for following the tracks of the window boundary coordinate.
In digitized processing, coordinate x, y is quantized, x, y round numbers, the discrete form of the centre of form is like this:
Wherein,
Wherein: when (x, and B when y) belonging to the target area (x, y)=1; When (x, and B when y) not belonging to the target area (x, y)=0; A, b, c, c is for following the tracks of the window boundary coordinate.
Summation is carried out in following the tracks of window, belongs to the point on the target, makes B that (x y)=1, promptly participates in summation; Do not belong to the point on the target, then make B that (x y)=0, does not promptly participate in summation;
Q
xExactly the x value of each picture element on the target image add up obtain and; Q
yExactly the y value of each picture element on the target image add up obtain and; M is exactly the sum of the picture element that comprises on the target image.
Calculate the deviation of the centre of form coordinate and the picture centre coordinate of moving target at last, and send to the ball machine head by serial ports by the control information that a kind of proportionate relationship is converted into the position coordinates deviation integrated camera The Cloud Terrace.
As Fig. 4, Fig. 5, Fig. 6, Fig. 7 and shown in Figure 8, intelligent vision monitoring device of the present invention comprises image collecting device 5, the power subsystem 3 of embedded processing circuit board 2, main control computer 8, panoramic picture harvester 4, band cloud platform.Mounting bracket 6 is installed in panoramic picture harvester 4 angle that can collect panoramic picture in the monitoring scene.Thereby mounting bracket 7 with the image collecting device 5 of band cloud platform be installed in monitoring on-the-spot directly over can follow the tracks of shooting to the moving target in each orientation in the space.Shown in circuit theory Fig. 5,6,7,8: embedded processing circuit board 2 mainly is made up of core processor circuit, Network Transmission circuit, power circuit, memory and four unit modules of image acquisition circuit, image acquisition circuit is with the analog video signal digitlization that collects, the view data that the memory circuitry storage of collected arrives, the Network Transmission circuit sends to main control computer 8 with view data by local area network (LAN), and receives the ball machine control code that main control computer 8 sends.Embedded processing circuit board 2 is fixed in the box body 1, and box body is installed on the on-the-spot wall of monitoring by screw.Panoramic picture harvester 4 directly is connected with embedded processing circuit board 2 by video output cable 9.The image collecting device 5 of band cloud platform directly is connected with embedded processing circuit board 2 by video output cable 9, and embedded processing circuit board 2 is connected with the image collecting device 5 of band cloud platform by RS485 universal serial bus 11, the movement velocity and the direction of control ball machine head.Among the present invention, panoramic picture harvester 4 uses general area array camera to meet the demands, and the image collecting device 5 of band cloud platform needs the The Cloud Terrace of configuration high-speed and the camera with autozoom function.Embedded processing circuit board 2 and main control computer 8 all are connected on the hub of local area network (LAN) by network connection.It is on-the-spot or on-the-spot away from monitoring that main control computer 8 can be placed on monitoring as required.
Comprised the built-in image collection device in this visual monitor system, this device is made up of core processor circuit, Network Transmission circuit, power circuit, memory and four functional modules of image acquisition circuit.Fig. 5 is the core processor circuit, and this circuit is the core of built-in image collection device, and all data processing and control all will just can be finished through this module.The master chip that adopts is the BF533 (element of U1 by name among Fig. 4) of ADI company.BF5335 passes through 8 parallel PPI interface PPI[0:3] and PF[12:15] cooperate the digital picture of the synchronization pulse PPI_CLK collection video decoding chip of a video decoding chip output to export.16 data/address bus DATA[0:15] with 19 bit address bus ADR[1:19] be connected with network controller.In addition, BF533 has independent SDRAM control interface to cooperate data and address bus to be connected with outside SDRAM.
Fig. 6 is the Network Transmission circuit, this circuit mainly receives digital video signal that BF533 sends over and sends the cradle head control signal that receives main control computer simultaneously by network by network, the network chip that uses is LAN91C111 (element of U6 by name among Fig. 3), with the main connecting line of BF533 be exactly 16 data/address bus DATA[0:15] and 19 bit address bus ADR[1:19].
Fig. 7 is a power circuit, and the power supply that this circuit is mainly finished 12V changes into the needed working power of each module of embedded system, and power circuit mainly provides the output of 3 road power supplys, the chip power supply voltage of 3.3V (element of U9 by name among Fig. 4); 1.2V BF533 kernel supply power voltage (among Fig. 4 by name U10 element); 1.8V aanalogvoltage supplying video decoding chip (among Fig. 4 by name U11 element).
Fig. 8 is memory and image acquisition circuit, this circuit is divided into two parts again, the image acquisition circuit major function is to gather analog video signal and be converted into digital vision signal output to send to BF533, master chip is TVP5150A (element of U8 by name among Fig. 5), cooperates a synchronization pulse PPI_CLK to be connected with BF533 by 8 digital video outputs.SDRAM memory circuit major function is the view data that storage BF533 collects, master chip is MT48LC16M16A2TG (element of U4 by name among Fig. 5), control signal wire is connected with the SDRAM control interface of BF533, and data and address wire are connected on the data and address bus of BF533.The FLASH chip is used for the operation code of storage system, and master chip is M29W800DB (element of U5 by name among Fig. 5), and data and address wire are connected on the data and address bus of BF533 too.
Operation principle: after system powers on, embedded processing circuit board 2 carries out initial work to each element circuit module, system is entered after the suitable operating state, embedded processing circuit board 2 control image acquisition circuit are sent to main control computer 8 with digitlization rear video image by network then with the in-site modeling video image digitlization that the image collecting device 5 of panoramic picture harvester 4 and band cloud platform is collected.The corresponding image algorithm of main control computer 8 usefulness carries out calculation process with two width of cloth images of the synchronization that image collecting device collects from two different angles, from digital picture, extract the three dimensional space coordinate of monitoring on-the-spot moving target and calculating each moving target, main control computer 8 is according to the priority level (as: size that configures then, movement velocity etc.) make a strategic decision out and need a moving target of track up, and the dimensional orientation of this moving target is converted into can be by the control routine of ball machine (image collecting device 5 of band cloud platform) identification, by network control routine is sent to embedded processing circuit board 2, embedded processing circuit board 2 is sent to the ball machine by RS485 universal serial bus 11 with control routine, is finally finished by the ball machine tracking of moving target is made a video recording.
Through the actual motion test, apparatus of the present invention can realize effective track up at 5 meters/ moving target below the S to the people in the normal walking or movement velocity, this be because in realizing the tracing process of a moving target only the run duration of ball machine just taken 500Ms, this time can't change.So when moving target or people's movement velocity surpasses 5 meters/S, will follow the tracks of the phenomenon of losing.Problem hereto, apparatus of the present invention can be good at solving: after each generation tracking is lost, main control computer 8 is searching moving target in the middle of the on-the-spot panoramic picture of two width of cloth that panoramic picture harvester 4 continuous acquisition arrive just, if do not search moving target, illustrate that this moment, moving target left the monitoring scene, main control computer 8 to the motion control code that 2 transmissions of embedded processing circuit board reset image collecting device 5 positions of band cloud platform, makes it to move to initial position by embedded processing circuit board 2 by the image collecting device 5 that RS485 universal serial bus 11 sends to the band cloud platform by network.If search moving target, main control computer 8 calculates after moving target orientation and the ball machine control routine and sends corresponding motion control code by network to embedded processing circuit board 2, and the image collecting device 5 that sends to the band cloud platform by RS485 universal serial bus 11 by embedded processing circuit board 2 makes it to recover the tracking to moving target.
Used imaging device in apparatus of the present invention, so the problem of nighttime imaging inevitably can occur.At this problem, apparatus of the present invention are the camera lens that imaging device has disposed band infrared LED lamp, make the imaging device can be in the images acquired at night.Through actual motion, apparatus of the present invention still can normally be moved under 0Lux illumination, and finish the track up to moving target.If realize nighttime imaging effect more clearly, then can be at the on-the-spot infrared light supply of placing of monitoring.Intelligent vision monitoring device of the present invention can be applied to safety-security area and realize monitoring on-the-spot round-the-clock automatic post, can be applied to the binocular vision of field of machine vision simulation human eye, can also be applied to field of visual inspection as front end intelligent vision sensor device, be that a kind of practicality is very strong, application is the intelligent vision monitoring device very widely.
Claims (7)
1. intelligent vision monitoring method is characterized in that step is:
(1), multiple mobile object detects: at first first two field picture after the input video denoising is made as initial background, to the t moment image I from the image scanner input
t(x, the y) noise that adopts median filter filtering part to be caused by transducer is with itself and t moment background image H
t(x y) subtracts each other and with binaryzation as a result, handles the back detection through morphologic filtering and whether have moving target, and renewal function obtains H according to testing result
T+1(x y) as the t+1 background image constantly after upgrading, thereby realizes the background real-time update, and described renewal function is
N represents the detected target sum of t frame, K in the formula
iBe i target, a, b are update coefficients; Then background image and present image are subtracted each other and threshold value turns to bianry image, further adopt morphology, cut apart, extract moving target by the connected region in neighbours territory with " spot " noise remove in the binary image, and with area less than threshold value T
aTarget cast out, obtain detected moving target in the present frame, and annotate with target boundary rectangle collimation mark;
(2), motion target tracking: at first use Region Segmentation method handle by the motion target area rectangle frame mark that obtains in the step (1), calculate the centre of form coordinate of that maximum rectangular area then, calculate the deviation of the target centre of form and picture centre, control the action of image scanner according to this deviation, target is remained near the central authorities of visual field, this is a feedback control procedure, constantly detects the centre of form of target and the deviation of picture centre, up to the central authorities of target in the visual field.
2. intelligent vision monitoring method according to claim 1 is characterized in that: the Region Segmentation step of moving target is as follows in the described step:
(1), initialization: remove scanning flag, the right margin of left=monitor area, the left margin of right=monitor area, number of targets ir=0;
(2), from top to bottom (y) from left to right (x) line by line scan do not beat the target picture point (x, y), if find to begin continuously from it some when not beating the target point and belonging to variations (target) zone, note top=y gives a some mark, otherwise commentaries on classics (6);
(3), in not beating target point, find the left margin point x of target left from this impact point, if x＜left, then left:=x; Find right margin point x more in the same way to the right, if x＞right, then right=x; And to the some mark that scans;
(4), calculate mid point: x:=(left+right)/2, and move down a row y:=y+1, (x y), gives a some mark to invocation point; Change (3) if this point belongs to impact point, otherwise note bottom:=y-1 obtains a fresh target,, change (2) if target area is given up it less than the minimum value that allows;
(5), relatively, if mutual distance is then incorporated this target into respective objects in allowing the scope (as 3 pixels) that merges, otherwise add object queue, ir=ir+1 changes (2) with other target area in fresh target and the object queue;
(6), the zone of mutual distance in the object queue in allowing the merging scope merged, object queue to the end.
3. intelligent vision monitoring device, it is characterized in that: it comprises main control computer (8), panoramic picture harvester (4), is with the ball-shaped camera (5) of high speed The Cloud Terrace, is placed on the on-the-spot box body (1) of monitoring and is fixed on box body (1) interior embedded processing circuit board (2) and power subsystem (3), described panoramic picture harvester (4) links to each other with embedded processing circuit board (2) by video output cable (9) with ball-shaped camera (5), and embedded processing circuit board (2) links to each other with main control computer (8) by network connection (10).
4. intelligent vision monitoring device according to claim 3, it is characterized in that: described panoramic picture harvester (4) adopts rifle type video camera, and embedded processing circuit board (2) is by the motion state of the ball-shaped camera (5) of RS485 universal serial bus (11) control band high speed The Cloud Terrace.
5. according to claim 3 or 4 or described intelligent vision monitoring device, it is characterized in that: described embedded processing circuit board (2) inserts the local area network (LAN) at main control computer (8) place by network connection (10).
6. according to claim 3 or 4 described intelligent vision monitoring devices, it is characterized in that: described panoramic picture harvester (4) directly links to each other with embedded processing circuit board (2) with RS485 universal serial bus (11) by video output cable (9) with the ball-shaped camera (5) of band high speed The Cloud Terrace.
7. intelligent vision monitoring device according to claim 5 is characterized in that: described panoramic picture harvester (4) directly links to each other with embedded processing circuit board (2) with RS485 universal serial bus (11) by video output cable (9) with the ball-shaped camera (5) of band high speed The Cloud Terrace.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA200710035636XA CN101127887A (en) | 2007-08-28 | 2007-08-28 | Intelligent vision monitoring method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA200710035636XA CN101127887A (en) | 2007-08-28 | 2007-08-28 | Intelligent vision monitoring method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101127887A true CN101127887A (en) | 2008-02-20 |
Family
ID=39095786
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA200710035636XA Pending CN101127887A (en) | 2007-08-28 | 2007-08-28 | Intelligent vision monitoring method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101127887A (en) |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101820531A (en) * | 2010-03-16 | 2010-09-01 | 江苏瀚远科技股份有限公司 | Video anti-counterfeiting monitoring device and monitoring method |
| CN101827204A (en) * | 2010-04-19 | 2010-09-08 | 成都索贝数码科技股份有限公司 | Method and system for detecting moving object |
| CN101908214A (en) * | 2010-08-10 | 2010-12-08 | 长安大学 | Moving object detection method with background reconstruction based on neighborhood correlation |
| CN102034112A (en) * | 2010-12-17 | 2011-04-27 | 浙江大学 | Method for identifying moving and static targets by using phased array three-dimensional acoustic image pickup sonar |
| CN102118611A (en) * | 2011-04-15 | 2011-07-06 | 中国电信股份有限公司 | Digital video surveillance method, digital video surveillance system and digital video surveillance platform for moving object |
| CN102223518A (en) * | 2010-04-14 | 2011-10-19 | 鸿富锦精密工业（深圳）有限公司 | Stereoscopic digital imaging monitoring system and stereoscopic digital imaging monitoring method |
| CN102244776A (en) * | 2011-07-06 | 2011-11-16 | 浙江浙大华是科技有限公司 | Automatic tracking laser thermal imaging monitoring system |
| CN102446355A (en) * | 2011-09-05 | 2012-05-09 | 西安理工大学 | Method for detecting target protruding from plane based on double viewing fields without calibration |
| CN102602681A (en) * | 2012-01-13 | 2012-07-25 | 天津工业大学 | Machine vision based online deviation fault detecting method for conveying belts |
| CN102833465A (en) * | 2012-07-24 | 2012-12-19 | 武汉大千信息技术有限公司 | Criminal investigation video pretreatment method based on movement detection |
| CN102853821A (en) * | 2012-08-31 | 2013-01-02 | 浙江中是科技有限公司 | Ship GPS positioning and checking method and ship GPS interactive system |
| CN102945105A (en) * | 2012-12-02 | 2013-02-27 | 周恺弟 | Multipoint portable optical touch screen and positioning method thereof |
| CN103024281A (en) * | 2013-01-11 | 2013-04-03 | 重庆大学 | Infrared and visible video integration system |
| CN103079059A (en) * | 2013-01-27 | 2013-05-01 | 中国科学技术大学 | Single-gun-camera-multi-dome-camera linkage method based on grid combination interpolation |
| CN103200394A (en) * | 2013-04-07 | 2013-07-10 | 南京理工大学 | Target image real time transmission and tracking method based on digital signal processor (DSP) and target image real time transmission and tracking device based on digital signal processor (DSP) |
| CN103297760A (en) * | 2013-06-04 | 2013-09-11 | 四川艾普视达数码科技有限公司 | Monitoring device with panoramic monitoring and fixed-point thinning camera shooting functions |
| CN103425355A (en) * | 2013-07-08 | 2013-12-04 | 周恺弟 | Omnidirectional camera structured portable optical touch screen and positioning calibrating method thereof |
| CN103514609A (en) * | 2013-07-06 | 2014-01-15 | 中国科学技术大学 | Image processing method based on fed back moving object segmentation |
| CN103945181A (en) * | 2014-04-04 | 2014-07-23 | 高新兴科技集团股份有限公司 | Network video information processing system for security and protection monitoring |
| CN104129615A (en) * | 2014-05-20 | 2014-11-05 | 大连日佳电子有限公司 | Precise conveyer belt control system |
| CN104270620A (en) * | 2014-10-27 | 2015-01-07 | 成都新舟锐视科技有限公司 | Real-time linkage intelligent tracking and early-warning system integrated with multiple cameras |
| CN104639917A (en) * | 2015-03-05 | 2015-05-20 | 苏州科达科技股份有限公司 | Video monitoring system and control method thereof |
| CN104943405A (en) * | 2015-07-03 | 2015-09-30 | 北人机器人系统（苏州）有限公司 | Visual marking system for automotive chassis swinging arm piece |
| CN105120213A (en) * | 2015-08-19 | 2015-12-02 | 江西钜维科技有限公司 | Noise pollution intelligent monitoring system |
| CN105959575A (en) * | 2016-07-08 | 2016-09-21 | 成都易瞳科技有限公司 | Zooming panoramic camera |
| CN107054654A (en) * | 2017-05-09 | 2017-08-18 | 广东容祺智能科技有限公司 | A kind of unmanned plane target tracking system and method |
| CN107067440A (en) * | 2017-03-28 | 2017-08-18 | 苏州工业园区服务外包职业学院 | A kind of image calibration method and system |
| WO2018103005A1 (en) * | 2016-12-07 | 2018-06-14 | 深圳前海达闼云端智能科技有限公司 | Electronic apparatus, and monitoring method and device for monitoring target |
| CN109451283A (en) * | 2018-12-21 | 2019-03-08 | 核动力运行研究所 | A kind of steam generator water chamber ambient image acquisition system |
| CN110751678A (en) * | 2018-12-12 | 2020-02-04 | 北京嘀嘀无限科技发展有限公司 | Moving object detection method and device and electronic equipment |
-
2007
- 2007-08-28 CN CNA200710035636XA patent/CN101127887A/en active Pending
Cited By (44)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101820531A (en) * | 2010-03-16 | 2010-09-01 | 江苏瀚远科技股份有限公司 | Video anti-counterfeiting monitoring device and monitoring method |
| CN102223518A (en) * | 2010-04-14 | 2011-10-19 | 鸿富锦精密工业（深圳）有限公司 | Stereoscopic digital imaging monitoring system and stereoscopic digital imaging monitoring method |
| CN101827204A (en) * | 2010-04-19 | 2010-09-08 | 成都索贝数码科技股份有限公司 | Method and system for detecting moving object |
| CN101827204B (en) * | 2010-04-19 | 2013-07-17 | 成都索贝数码科技股份有限公司 | Method and system for detecting moving object |
| CN101908214A (en) * | 2010-08-10 | 2010-12-08 | 长安大学 | Moving object detection method with background reconstruction based on neighborhood correlation |
| CN102034112A (en) * | 2010-12-17 | 2011-04-27 | 浙江大学 | Method for identifying moving and static targets by using phased array three-dimensional acoustic image pickup sonar |
| CN102118611B (en) * | 2011-04-15 | 2013-01-02 | 中国电信股份有限公司 | Digital video surveillance method, digital video surveillance system and digital video surveillance platform for moving object |
| CN102118611A (en) * | 2011-04-15 | 2011-07-06 | 中国电信股份有限公司 | Digital video surveillance method, digital video surveillance system and digital video surveillance platform for moving object |
| CN102244776A (en) * | 2011-07-06 | 2011-11-16 | 浙江浙大华是科技有限公司 | Automatic tracking laser thermal imaging monitoring system |
| CN102244776B (en) * | 2011-07-06 | 2013-03-20 | 浙江浙大华是科技有限公司 | Automatic tracking laser thermal imaging monitoring system |
| CN102446355B (en) * | 2011-09-05 | 2013-11-06 | 西安理工大学 | Method for detecting target protruding from plane based on double viewing fields without calibration |
| CN102446355A (en) * | 2011-09-05 | 2012-05-09 | 西安理工大学 | Method for detecting target protruding from plane based on double viewing fields without calibration |
| CN102602681B (en) * | 2012-01-13 | 2014-01-08 | 天津工业大学 | Machine vision based online deviation fault detecting method for conveying belts |
| CN102602681A (en) * | 2012-01-13 | 2012-07-25 | 天津工业大学 | Machine vision based online deviation fault detecting method for conveying belts |
| CN102833465A (en) * | 2012-07-24 | 2012-12-19 | 武汉大千信息技术有限公司 | Criminal investigation video pretreatment method based on movement detection |
| CN102853821A (en) * | 2012-08-31 | 2013-01-02 | 浙江中是科技有限公司 | Ship GPS positioning and checking method and ship GPS interactive system |
| CN102853821B (en) * | 2012-08-31 | 2015-06-03 | 浙江中起科技有限公司 | Ship GPS positioning and checking method and ship GPS interactive system |
| CN102945105B (en) * | 2012-12-02 | 2016-03-02 | 狒特科技（北京）有限公司 | A kind of multipoint portable optical touch screen and localization method thereof |
| CN102945105A (en) * | 2012-12-02 | 2013-02-27 | 周恺弟 | Multipoint portable optical touch screen and positioning method thereof |
| CN103024281A (en) * | 2013-01-11 | 2013-04-03 | 重庆大学 | Infrared and visible video integration system |
| CN103079059A (en) * | 2013-01-27 | 2013-05-01 | 中国科学技术大学 | Single-gun-camera-multi-dome-camera linkage method based on grid combination interpolation |
| CN103079059B (en) * | 2013-01-27 | 2015-10-28 | 中国科学技术大学 | A kind of based on the single gunlock-many ball machine interlock method of grid in conjunction with interpolation |
| CN103200394B (en) * | 2013-04-07 | 2016-04-20 | 南京理工大学 | Based on the target image real-time Transmission of DSP and tracking and device thereof |
| CN103200394A (en) * | 2013-04-07 | 2013-07-10 | 南京理工大学 | Target image real time transmission and tracking method based on digital signal processor (DSP) and target image real time transmission and tracking device based on digital signal processor (DSP) |
| CN103297760A (en) * | 2013-06-04 | 2013-09-11 | 四川艾普视达数码科技有限公司 | Monitoring device with panoramic monitoring and fixed-point thinning camera shooting functions |
| CN103514609B (en) * | 2013-07-06 | 2016-03-09 | 中国科学技术大学 | A kind of image processing method of the moving meshes based on feedback |
| CN103514609A (en) * | 2013-07-06 | 2014-01-15 | 中国科学技术大学 | Image processing method based on fed back moving object segmentation |
| CN103425355B (en) * | 2013-07-08 | 2016-09-07 | 狒特科技（北京）有限公司 | The portable optical touch screen of a kind of omnidirectional camera structure and location calibration steps thereof |
| CN103425355A (en) * | 2013-07-08 | 2013-12-04 | 周恺弟 | Omnidirectional camera structured portable optical touch screen and positioning calibrating method thereof |
| CN103945181A (en) * | 2014-04-04 | 2014-07-23 | 高新兴科技集团股份有限公司 | Network video information processing system for security and protection monitoring |
| CN104129615A (en) * | 2014-05-20 | 2014-11-05 | 大连日佳电子有限公司 | Precise conveyer belt control system |
| CN104270620A (en) * | 2014-10-27 | 2015-01-07 | 成都新舟锐视科技有限公司 | Real-time linkage intelligent tracking and early-warning system integrated with multiple cameras |
| CN104639917B (en) * | 2015-03-05 | 2017-12-15 | 苏州科达科技股份有限公司 | A kind of video monitoring system and its control method |
| CN104639917A (en) * | 2015-03-05 | 2015-05-20 | 苏州科达科技股份有限公司 | Video monitoring system and control method thereof |
| CN104943405A (en) * | 2015-07-03 | 2015-09-30 | 北人机器人系统（苏州）有限公司 | Visual marking system for automotive chassis swinging arm piece |
| CN105120213A (en) * | 2015-08-19 | 2015-12-02 | 江西钜维科技有限公司 | Noise pollution intelligent monitoring system |
| CN105959575A (en) * | 2016-07-08 | 2016-09-21 | 成都易瞳科技有限公司 | Zooming panoramic camera |
| CN105959575B (en) * | 2016-07-08 | 2022-02-08 | 成都易瞳科技有限公司 | Variable-focus panoramic camera |
| WO2018103005A1 (en) * | 2016-12-07 | 2018-06-14 | 深圳前海达闼云端智能科技有限公司 | Electronic apparatus, and monitoring method and device for monitoring target |
| CN107067440A (en) * | 2017-03-28 | 2017-08-18 | 苏州工业园区服务外包职业学院 | A kind of image calibration method and system |
| CN107067440B (en) * | 2017-03-28 | 2020-05-19 | 苏州工业园区服务外包职业学院 | Image calibration method and system |
| CN107054654A (en) * | 2017-05-09 | 2017-08-18 | 广东容祺智能科技有限公司 | A kind of unmanned plane target tracking system and method |
| CN110751678A (en) * | 2018-12-12 | 2020-02-04 | 北京嘀嘀无限科技发展有限公司 | Moving object detection method and device and electronic equipment |
| CN109451283A (en) * | 2018-12-21 | 2019-03-08 | 核动力运行研究所 | A kind of steam generator water chamber ambient image acquisition system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101127887A (en) | Intelligent vision monitoring method and device | |
| US11486548B2 (en) | System for detecting crack growth of asphalt pavement based on binocular image analysis | |
| CN107818326B (en) | A kind of ship detection method and system based on scene multidimensional characteristic | |
| CN101969548B (en) | Active video acquiring method and device based on binocular camera shooting | |
| Zhao et al. | Detection, tracking, and geolocation of moving vehicle from uav using monocular camera | |
| CN102036054B (en) | Intelligent video monitoring system based on three-dimensional virtual scene | |
| CN102811343B (en) | Intelligent video monitoring system based on behavior recognition | |
| Song et al. | Real-time tunnel crack analysis system via deep learning | |
| Bodor et al. | Vision-based human tracking and activity recognition | |
| CN101561270B (en) | Embedded omnidirectional ball vision object detection and mobile monitoring system and method | |
| CN109444911A (en) | A kind of unmanned boat waterborne target detection identification and the localization method of monocular camera and laser radar information fusion | |
| CN102447835A (en) | Non-blind-area multi-target cooperative tracking method and system | |
| CN102148921B (en) | Multi-target tracking method based on dynamic group division | |
| CN101576952B (en) | Method and device for detecting static targets | |
| Xu et al. | Real-time detecting method of marine small object with underwater robot vision | |
| CN102982341A (en) | Self-intended crowd density estimation method for camera capable of straddling | |
| CN113592905B (en) | Vehicle driving track prediction method based on monocular camera | |
| CN103279791A (en) | Pedestrian counting method based on multiple features | |
| CN113012195B (en) | Method for river surface flow velocity estimation based on optical flow calculation and readable storage medium | |
| CN1480716A (en) | Automatic system for tracking and monitoring infrared realtime image temp of quick multiple objective | |
| CN104570764A (en) | Verification platform for airborne target indicating system | |
| CN106504274A (en) | A kind of visual tracking method and system based under infrared camera | |
| CN106558069A (en) | A kind of method for tracking target and system based under video monitoring | |
| CN103281518A (en) | Multifunctional networking all-weather intelligent video monitoring system | |
| CN109946564A (en) | A kind of distribution network overhead line inspection data collection method and cruising inspection system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20080220 |