CN103473787A - On-bridge-moving-object detection method based on space geometry relation - Google Patents
On-bridge-moving-object detection method based on space geometry relation Download PDFInfo
- Publication number
- CN103473787A CN103473787A CN2013103212197A CN201310321219A CN103473787A CN 103473787 A CN103473787 A CN 103473787A CN 2013103212197 A CN2013103212197 A CN 2013103212197A CN 201310321219 A CN201310321219 A CN 201310321219A CN 103473787 A CN103473787 A CN 103473787A
- Authority
- CN
- China
- Prior art keywords
- bridge
- terrestrial reference
- real
- time
- unique point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention discloses an on-bridge-moving-object detection method based on a space geometry relation. The method includes: selecting a bridge from a satellite picture and a river bank with obvious shape features in a river area which the bridge is in as landmarks; manufacturing a landmark reference picture according to the satellite picture which the bridge landmark and the river bank landmark are in; selecting river-bank landmark feature points and bridge-landmark feature points according to the landmark reference picture and at the same time establishing a feature library for spacial constraint relation between the river-bank landmark feature points and the bridge-landmark feature points; acquiring a real-time picture and carrying out perspective transformation on the landmark reference picture according to imaging position parameters of an aircraft so that positions of the bridge-landmark feature points in the real-time picture are obtained; setting average position coordinate of the bridge-landmark feature points in the real-time picture after the perspective transformation and determining a rectangular subarea according to prior knowledge about size of the bridge landmark in the real-time picture. The on-bridge-moving-object detection method based on the space geometry relation is capable of solving a technical problem that under a background of a diara or a bridge, instantaneity of detection of a moving object to be detected is poor.
Description
Technical field
The invention belongs to the identification of imaging automatic target and the technical field that navigational guidance intersects, more specifically, relate to moving target detecting method on a kind of bridge based on the space geometry relation.
Background technology
On bridge Detection for Moving Target modern military and civilian in all occupy very important status, it is the main task of Detection for Moving Target on bridge all the time that reliably and accurately high-quality target information is provided.For this reason, domestic and international many researchists are devoted to the research of this problem always.On bridge, moving object detection system generally is based on the processing to image sequence, tries hard to identify target from complicated background, and the characteristics of motion of target is predicted, realize to target continuously, follow the tracks of accurately.
To the detection of point source moving target, the wide coverage of what is called " Detect before Track " and two class methods researchs such as " first follow the tracks of afterwards and detect " is arranged.In real scene, particularly, under complex background condition, exist the Moving Objects of full range of sizes and the object of various movement velocitys to need our determination and analysis in the lump.Yet existing most methods and algorithm are single time scales, according to the detection frame by frame of consecutive frame.Like this when the imaging platform gradually near target, the yardstick of target in figure in real time will change, on traditional bridge, moving object detection is only considered a kind of yardstick, will cause so in actual applications undetectedly, makes algorithm not possess very strong adaptability.
Moreover on existing bridge, the moving object detection recognition methods is all with regard to moving target itself, and the residing background of target and context restrictions are not considered.On traditional bridge, moving object detection is not considered the restriction range of moving target, and full figure is searched for, so not only can cause huge algorithm expense, make the real-time of moving object detection on bridge not ensure, also cause a large amount of false-alarms and false dismissal.In addition, full figure is carried out to moving object detection on bridge, the algorithm expense is very large, makes the real-time of moving object detection on bridge not ensure.
Summary of the invention
Above defect or Improvement requirement for prior art, the invention provides moving target detecting method on a kind of bridge based on the space geometry relation, its purpose is continent, the existing middle of the river, under the bridge background, moving target to be identified is small and weak, there is more background interference, the technical matters that the moving object detection real-time is poor, this method can detect identification to the moving target under bridge terrestrial reference constraint under the moving platform condition, can meet unmanned vehicle and the autonomous accurately detection of the moving target of people's aircraft in the highway scene of ground identification is arranged, accuracy is high, and there is good real-time.
For achieving the above object, according to one aspect of the present invention, provide moving target detecting method on a kind of bridge based on the space geometry relation, comprised the following steps:
(1) choose in bridge and this bridge place river region the obvious riverbank of shape facility as terrestrial reference from satellite photo;
(2) prepare the terrestrial reference reference diagram according to the satellite photo at bridge terrestrial reference and terrestrial reference place, riverbank;
(3) choose riverbank terrestrial reference unique point according to the terrestrial reference reference diagram
and bridge terrestrial reference unique point T (x
t, y
t), set up the space constraint relationship characteristic storehouse Δ between riverbank terrestrial reference unique point and bridge terrestrial reference unique point simultaneously
i(Δ x
i, Δ y
i), i=1 wherein, 2 ..., j means riverbank terrestrial reference unique point sequence number, the sum that j is riverbank terrestrial reference unique point,
with
be respectively horizontal stroke, the ordinate of riverbank terrestrial reference unique point in the terrestrial reference reference diagram, x
tand y
tbe respectively horizontal stroke, the ordinate of bridge terrestrial reference unique point in the terrestrial reference reference diagram,
(4) obtain real-time figure, and according to aircraft imaging attitude parameter, the terrestrial reference reference diagram is carried out to perspective transform, to obtain bridge terrestrial reference unique point T (x
t, y
t) position result in real-time figure
i;
(5) the mean place coordinate result=(result of perspective transform axle casing terrestrial reference unique point in real-time figure is set
1+ result
2+ ...+result
j)/j;
(6) point centered by coordinate result, and according to bridge be marked on size in real-time figure priori determine the region of interest of a rectangular sub-regions territory as bridge terrestrial reference constraint;
(7) real-time figure is carried out to binary segmentation, to obtain segmentation result figure;
(8), in the segmentation result figure obtained in step (7), carry out the detection of bridge constraint according to the region of interest of resulting bridge terrestrial reference constraint in step (6), to determine bridge constraint Area
r;
(9) under moving platform, the real-time figure of two frames is carried out to registration, and the bridge constraint Area extracted in step (8)
rin carry out multiple dimensioned moving target window selection, to obtain window area corresponding to each pixel in the real-time figure of every frame;
(10) the motion significance measure value in the different windows zone of using the multiple dimensioned moving target detecting method calculation procedure of space-time (9) to obtain, determine that window area corresponding to maximum motion significance measure value is as the motion salient region, and obtain the optimal time interval of this motion salient region;
(11) utilize optimal time interval
, each motion salient region extracts moving target in zone in multiple frame cumulation difference method and labeling method bridge constraint that step (8) is extracted, to complete the detection to moving target on bridge.
Preferably, choose j riverbank terrestrial reference unique point from each riverbank terrestrial reference, each bridge terrestrial reference is chosen a bridge terrestrial reference unique point, the point that wherein terrestrial reference unique point in riverbank is arc shaped riverbank curvature of a curve maximum, the centre of form that bridge terrestrial reference unique point is this bridge zone.
What preferably, binary segmentation adopted is the OTSU algorithm.
Preferably, the detection of bridge constraint is to adopt Hough transformation.
Preferably, step (9) comprises following sub-step:
(9-1) select the initial time interval of delta t, adopt the SIFT method to scheme in real time f (x, y, t to two frames
c) and f (x, y, t
c+ Δ t) carry out registration, wherein x is the horizontal ordinate of the real-time figure of a frame wherein, and y is its ordinate, t
ccurrent image frame for real-time figure;
(9-2) the bridge constraint Area extracted in step (8)
rin determine M window, its size is followed successively by from small to large: S
min_x* S
min_y, (S
min_x+ Δ S
x) * (S
min_y+ Δ S
y) ..., (S
min_x+ (M-1) Δ S
x) * (S
min_y+ (M-1) Δ S
y), wherein M is positive integer, S
min_xthe minimum value that means length of window, S
min_ythe minimum value that means window width, Δ S
xmean length increment, Δ S
ymean the width increment;
(9-3) utilize minimum window
respectively to two frames scheme in real time f (x, y,
ct) and f (x, y, t
c+ Δ t) pursue the pixel traversal, to obtain respectively window area corresponding to each pixel (x, y) in the real-time figure of every frame
with
.
Preferably, step (10) specifically comprises following sub-step:
(10-1) calculate minimum window (S
min_x, S
min_y) the overlay area Ω that locates at pixel (x, y)
x,ymotion significance measure value Value (x, y);
(10-2) for each movement mark pixel (x ', y '), determine maximal value in corresponding M the motion significance measure value Value from it (x ', y '), overlay area corresponding to this maximal value is the motion salient region, is designated as ω
x ', y 'thereby, the bridge constraint Area extracted for step (8)
rin N movement mark pixel, can obtain N motion salient region, be designated as
;
(10-3) the bridge constraint Area extracted in step (8)
rin, calculate each motion salient region ω
x ', y 'optimal time interval
.
Preferably, step (11) is specially, for each motion salient region
(i=1,2 ... N), its best interFrameGap
in previous step, obtain, to each motion salient region
, get image to f
t(x, y) with
(x, y) or
(x, y), in the zone of two width images
inside do the multiple frame cumulation difference, symmetrical two frames in front and back and present frame are accumulated respectively difference, cumulative error partial image in the bridge constraint that extraction step (8) extracts respectively, then strengthen the difference between the cumulative error partial image by image co-registration, thereafter, by the OTSU algorithm, the cumulative error partial image is carried out to Threshold segmentation, morphology processing and mark, to extract the moving target in the cumulative error partial image, finally obtain the moving object detection result.
In general, the above technical scheme of conceiving by the present invention compared with prior art, can obtain following beneficial effect:
1, owing to having adopted step (5) and step (6) to extract the terrestrial reference constraint at target place, rather than full figure is carried out to moving object detection, thereby the detectability of moving-target is fixed in constraint, effectively got rid of a large amount of background interference, largely reduced false-alarm.
2, owing to having adopted step (5) and step (6) to extract the terrestrial reference constraint at target place, rather than full figure is carried out to moving object detection, thereby reduced moving object detection algorithm expense, ensured real-time;
3, owing to having adopted step (7) to carrying out multiple dimensioned motion window selection in the terrestrial reference constraint, make moving object detection there is multiple dimensioned characteristic, can detect under the moving platform condition target of difference is arranged time-empty position, size.
The accompanying drawing explanation
Fig. 1 is the overview flow chart that the present invention is based on moving target detecting method on the bridge of space geometry relation.
Fig. 2 is bridge and zone, continent, middle of the river satellite photo.
Fig. 3 is continent, the middle of the river, bridge terrestrial reference reference diagram.
Fig. 4 is that continent, the middle of the river, bridge terrestrial reference characteristic quantity are chosen.
Fig. 5 illustrates the real-time figure obtained in step (3).
Fig. 6 is the local region of interest of bridge terrestrial reference constraint.
Fig. 7 is figure as a result after real-time figure binary segmentation.
Fig. 8 is that result is extracted in bridge terrestrial reference constraint.
Fig. 9 is that the 10th frame and the 19th frame SIFT angle point extract result.
Figure 10 is that the lower motion salient region of the 19th frame figure bridge terrestrial reference constraint extracts result.
Figure 11 is the lower moving Object Segmentation result of the 19th frame figure bridge terrestrial reference constraint.
Figure 12 is the real-time figure final detection result of the 19th frame.
Embodiment
In order to make purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.In addition, below in each embodiment of described the present invention involved technical characterictic as long as form each other conflict, just can mutually not combine.
As shown in Figure 1, the present invention is based on moving target detecting method on the bridge of space geometry relation comprises the following steps:
(1) choose in bridge and this bridge place river region the obvious riverbank of shape facility as terrestrial reference from satellite photo; The satellite photo of Fig. 2 for obtaining, this image resolution ratio is 0.8 meter, and size is 541 * 541 pixels, and the present invention chooses continent, the middle of the river as the riverbank terrestrial reference;
(2) prepare the terrestrial reference reference diagram according to the satellite photo at bridge terrestrial reference and terrestrial reference place, riverbank, the process that specifically prepares the terrestrial reference reference diagram is disclosed at the Chinese patent " a kind of plane for forward sight navigational guidance terrestrial reference is selected and reference map preparation method " (patent No. ZL200910273308.2) of the applicant's submission, do not repeat them here, Fig. 3 is generated terrestrial reference reference diagram;
(3) choose riverbank terrestrial reference unique point LSP according to the terrestrial reference reference diagram
i (i=1 wherein, 2 ..., j means riverbank terrestrial reference unique point sequence number) and bridge terrestrial reference unique point T (x
t, y
t), set up the space constraint relationship characteristic storehouse Δ between riverbank terrestrial reference unique point and bridge terrestrial reference unique point simultaneously
i(Δ x
i, Δ y
i); Wherein
with
be respectively horizontal stroke, the ordinate of riverbank terrestrial reference unique point in the terrestrial reference reference diagram, x
tand y
tbe respectively horizontal stroke, the ordinate of bridge terrestrial reference unique point in the terrestrial reference reference diagram,
the sum that wherein j is riverbank terrestrial reference unique point, its span is 3 to 5; Particularly, choose 3 to 5 riverbank terrestrial reference unique points from each riverbank terrestrial reference, each bridge terrestrial reference is chosen a bridge terrestrial reference unique point, the point that wherein terrestrial reference unique point in riverbank is arc shaped riverbank curvature of a curve maximum, the centre of form that bridge terrestrial reference unique point is this bridge zone.For example, as shown in Figure 4, choose the point of maximum curvature of continent, 4 middles of the river bank line as riverbank terrestrial reference unique point, the 1st riverbank terrestrial reference unique point coordinate is (328,80), is designated as LSP
1; The 2nd riverbank landmark point coordinate is (332,103), is designated as LSP
2; The 3rd riverbank landmark point coordinate is (239,162), is designated as LSP
3; The 4th riverbank landmark point coordinate is (262,173), is designated as LSP
4, the coordinate (320,400) that the coordinate of bridge terrestrial reference unique point T is bridge zone orthogonal projection centre of form in the terrestrial reference reference diagram, these 5 unique points are marked with one group of solid round dot in Fig. 4.Space constraint relationship characteristic storehouse between riverbank terrestrial reference unique point and bridge terrestrial reference unique point is: { Δ
1(8,320), Δ
2(12,297), Δ
3(81,238), Δ
4(58,227) }.
(4) obtain real-time figure, and according to aircraft imaging attitude parameter, the terrestrial reference reference diagram is carried out to perspective transform, to obtain bridge terrestrial reference unique point T (x
t, y
t) position result in real-time figure
i; Particularly, the forward sight state is arrived in the continent, the middle of the river of preparation in step (2) and the perspective transform of bridge terrestrial reference reference diagram, the Chinese patent that the process of concrete perspective transform has been submitted to the applicant " method of position identification ground stereoscopic buildings is three-dimensionally demarcated in a kind of utilization " (patent No.: disclosed 200910063624.7), do not repeat them here.When perspective transform, the aiming point of aircraft optical axis is respectively the LSP in Fig. 4
1, LSP
2, LSP
3and LSP
4, when the aiming point of optical axis is LSP
1the time, Fig. 5 is after perspective transform, and bridge terrestrial reference unique point is that coordinate on the 320 pixels tall plane of delineation that is 256 pixels is (147,117) at width, and with result
1mean (as shown in Figure 6), when the aiming point of optical axis is LSP
2the time, Fig. 5 is after perspective transform, and bridge terrestrial reference unique point is that coordinate on the 320 pixels tall plane of delineation that is 256 pixels is (163,117) at width, and with result
2mean (as shown in Figure 6), when the aiming point of optical axis is LSP
3the time, Fig. 5 is after perspective transform, and bridge terrestrial reference unique point is that coordinate on the 320 pixels tall plane of delineation that is 256 pixels is (143,122) at width, with result
3mean (as shown in Figure 6), when the aiming point of optical axis is LSP
4the time, Fig. 5 is after perspective transform, and bridge terrestrial reference unique point is that coordinate on the 320 pixels tall plane of delineation that is 256 pixels is (151,122) at width, with result
4mean (as shown in Figure 6).
(5) the mean place coordinate result=(result of perspective transform axle casing terrestrial reference unique point in real-time figure is set
1+ result
2+ ...+result
j)/j; Result=(result in above example
1+ result
2+ ...+result
4)/4, its value is (151,120), foundation when this coordinate is follow-up location region of interest;
(6) point centered by coordinate result, and according to bridge be marked on size in real-time figure priori determine the region of interest of a rectangular sub-regions territory as bridge terrestrial reference constraint, as shown in Figure 6, this rectangular sub-regions field width degree is 245 pixels, is highly 28 pixels;
(7) real-time figure is carried out to binary segmentation, to obtain segmentation result figure; In the present embodiment, adopt the OTSU algorithm to carry out binary segmentation, its result as shown in Figure 7;
(8), in the segmentation result figure obtained in step (7), carry out the detection of bridge constraint according to the region of interest of resulting bridge terrestrial reference constraint in step (6), to determine bridge constraint Area
r, as shown in Figure 8; In the present embodiment, the detection of bridge constraint is to adopt Hough transformation;
(9) under moving platform, the real-time figure of two frames is carried out to registration, and the bridge constraint Area extracted in step (8)
rin carry out multiple dimensioned moving target window selection, to obtain window area corresponding to each pixel in the real-time figure of every frame; This step comprises following sub-step:
(9-1) select the initial time interval of delta t, adopt yardstick invariant features conversion (Scale-invariant feature transform is called for short SIFT) method to scheme in real time f (x, y, t to two frames
c) and f (x, y, t
c+ Δ t) carry out registration; Wherein x is the horizontal ordinate of the real-time figure of a frame wherein, and y is its ordinate, t
cfor the current image frame of real-time figure, Δ t is greater than 5 positive integer, and preferably, its value is 10 frame interFrameGaps, and the registration results of the 10th frame and the 19th frame realtime graphic as shown in Figure 9.
(9-2) the bridge constraint Area extracted in step (8)
rin determine M window, its size is followed successively by from small to large: S
min_x* S
min_y, (S
min_x+ Δ S
x) * (S
min_y+ Δ S
y) ..., (S
min_x+ (M-1) Δ S
x) * (S
min_y+ (M-1) Δ S
y), wherein M is positive integer, S
min_xmean the minimum value of length of window, its span is to be greater than 2 pixels, S
min_ymean the minimum value of window width, its span is to be greater than 2 pixels, Δ S
xmean length increment, its value equals 2, Δ S
ymean the width increment, its value equals 2;
(9-3) utilize minimum window
respectively to two frames scheme in real time f (x, y,
ct and f (x, y, t
c+ Δ t) pursue the pixel traversal, to obtain respectively window area corresponding to each pixel (x, y) in the real-time figure of every frame
with
;
(10) the motion significance measure value in the different windows zone of using the multiple dimensioned moving target detecting method calculation procedure of space-time (9) to obtain, determine that window area corresponding to maximum motion significance measure value is as the motion salient region, and obtain the optimal time interval of this motion salient region; This step specifically comprises following sub-step:
(10-1) calculate minimum window (S
min_x, S
min_y) the overlay area Ω that locates at pixel (x, y)
x,ymotion significance measure value Value (x, y); Disclosed in the Chinese patent " the multiple dimensioned moving target detecting method of space-time " (application number 201210591104.5) that the process that concrete motion significance measure value is derived has been submitted to the applicant, do not repeated them here.Wherein, if motion significance measure value Value (x, y) is more than or equal to predetermined threshold, regional Ω
x,ybelong to candidate's motion salient region, its corresponding pixel is movement mark pixel (x ', y '), otherwise, regional Ω
x,ydo not belong to candidate's motion salient region, the span of subscribing threshold value is 0 to 1, is preferably 0.6;
(10-2) for each movement mark pixel (x ', y '), determine maximal value in corresponding M the motion significance measure value Value from it (x ', y '), overlay area corresponding to this maximal value is the motion salient region, is designated as ω
x ', y 'thereby, the bridge constraint Area extracted for step (8)
rin N movement mark pixel, can obtain N motion salient region, be designated as
, i=1 wherein, 2 .N..Be that in the real-time figure bridge of the 19th frame constraint, the motion salient region extracts result as shown in figure 10.
(10-3) the bridge constraint Area extracted in step (8)
rin, calculate each motion salient region ω
x ', y 'optimal time interval
; The computation process of concrete optimal time interval is referring to described in Chinese patent " the multiple dimensioned moving target detecting method of space-time ".
(11) utilize optimal time interval
, each motion salient region extracts moving target in zone in multiple frame cumulation difference method and labeling method bridge constraint that step (8) is extracted, to complete the detection to moving target on bridge.
For each motion salient region
(i=1,2 ... N), its best interFrameGap
in previous step, obtain.To each motion salient region
, get image to f
t(x, y) with
(x, y) or
(x, y), in the zone of two width images
inside do the multiple frame cumulation difference, the process of multiple frame cumulation difference is referring to described in Chinese patent " the multiple dimensioned moving target detecting method of space-time ".Symmetrical two frames in front and back and present frame are accumulated respectively difference, cumulative error partial image in the bridge constraint that extraction step (8) extracts respectively, then strengthen the difference between the cumulative error partial image by image co-registration, thereafter, by the OTSU algorithm, the cumulative error partial image is carried out to Threshold segmentation, morphology processing and mark, to extract the moving target in the cumulative error partial image, finally obtain the moving object detection result.Be the lower moving Object Segmentation result of the real-time figure highway of the 19th frame terrestrial reference constraint as shown in figure 12.
Those skilled in the art will readily understand; the foregoing is only preferred embodiment of the present invention; not in order to limit the present invention, all any modifications of doing within the spirit and principles in the present invention, be equal to and replace and improvement etc., within all should being included in protection scope of the present invention.
Claims (7)
1. moving target detecting method on the bridge based on the space geometry relation, is characterized in that, comprises the following steps:
(1) choose in bridge and this bridge place river region the obvious riverbank of shape facility as terrestrial reference from satellite photo;
(2) prepare the terrestrial reference reference diagram according to the satellite photo at bridge terrestrial reference and terrestrial reference place, riverbank;
(3) choose riverbank terrestrial reference unique point according to the terrestrial reference reference diagram
and bridge terrestrial reference unique point T (x
t, y
t), set up the space constraint relationship characteristic storehouse △ between riverbank terrestrial reference unique point and bridge terrestrial reference unique point simultaneously
i(△ x
i, △ y
i), i=1 wherein, 2 ..., j means riverbank terrestrial reference unique point sequence number, the sum that j is riverbank terrestrial reference unique point,
with
be respectively horizontal stroke, the ordinate of riverbank terrestrial reference unique point in the terrestrial reference reference diagram, x
tand y
tbe respectively horizontal stroke, the ordinate of bridge terrestrial reference unique point in the terrestrial reference reference diagram,
(4) obtain real-time figure, and according to aircraft imaging attitude parameter, the terrestrial reference reference diagram is carried out to perspective transform, to obtain bridge terrestrial reference unique point T (x
t, y
t) position result in real-time figure
i;
(5) the mean place coordinate result=(result of perspective transform axle casing terrestrial reference unique point in real-time figure is set
1+ result
2+ ...+result
j)/j;
(6) point centered by coordinate result, and according to bridge be marked on size in real-time figure priori determine the region of interest of a rectangular sub-regions territory as bridge terrestrial reference constraint;
(7) real-time figure is carried out to binary segmentation, to obtain segmentation result figure;
(8), in the segmentation result figure obtained in step (7), carry out the detection of bridge constraint according to the region of interest of resulting bridge terrestrial reference constraint in step (6), to determine bridge constraint Area
r;
(9) under moving platform, the real-time figure of two frames is carried out to registration, and the bridge constraint Area extracted in step (8)
rin carry out multiple dimensioned moving target window selection, to obtain window area corresponding to each pixel in the real-time figure of every frame;
(10) the motion significance measure value in the different windows zone of using the multiple dimensioned moving target detecting method calculation procedure of space-time (9) to obtain, determine that window area corresponding to maximum motion significance measure value is as the motion salient region, and obtain the optimal time interval of this motion salient region;
2. moving target detecting method on bridge according to claim 1, it is characterized in that, step (3) is specially, choose j riverbank terrestrial reference unique point from each riverbank terrestrial reference, each bridge terrestrial reference is chosen a bridge terrestrial reference unique point, the point that wherein terrestrial reference unique point in riverbank is arc shaped riverbank curvature of a curve maximum, the centre of form that bridge terrestrial reference unique point is this bridge zone.
3. moving target detecting method on bridge according to claim 1, is characterized in that, what binary segmentation adopted is the OTSU algorithm.
4. moving target detecting method on bridge according to claim 1, is characterized in that, the detection of bridge constraint is to adopt Hough transformation.
5. moving target detecting method on bridge according to claim 1, is characterized in that, step (9) comprises following sub-step:
(9-1) select initial time interval △ t, adopt the SIFT method to scheme in real time f (x, y, t to two frames
c) and f (x, y, t
c+ △ t) carry out registration, wherein x is the horizontal ordinate of the real-time figure of a frame wherein, and y is its ordinate, t
ccurrent image frame for real-time figure;
(9-2) the bridge constraint Area extracted in step (8)
rin determine M window, its size is followed successively by from small to large: S
min_x* S
min_y, (S
min_x+ △ S
x) * (S
min_y+ △ S
y) ..., (S
min_x+ (M-1) △ S
x) * (S
min_y+ (M-1) △ S
y), wherein M is positive integer, S
min_xthe minimum value that means length of window, S
min_ythe minimum value that means window width, △ S
xmean length increment, △ S
ymean the width increment;
6. moving target detecting method on bridge according to claim 5, is characterized in that, step (10) specifically comprises following sub-step:
(10-1) calculate minimum window (S
min_x, S
min_y) the overlay area Ω that locates at pixel (x, y)
x,ymotion significance measure value Value (x, y);
(10-2) for each movement mark pixel (x ', y '), determine maximal value in corresponding M the motion significance measure value Value from it (x ', y '), overlay area corresponding to this maximal value is the motion salient region, is designated as ω
x ', y 'thereby, the bridge constraint Area extracted for step (8)
rin N movement mark pixel, can obtain N motion salient region, be designated as
7. moving target detecting method on bridge according to claim 6, is characterized in that, step (11) is specially, for each motion salient region
(i=1,2 ... N), its best interFrameGap
in previous step, obtain, to each motion salient region
get image to f
t(x, y) with
or
zone at two width images
inside do the multiple frame cumulation difference, symmetrical two frames in front and back and present frame are accumulated respectively difference, cumulative error partial image in the bridge constraint that extraction step (8) extracts respectively, then strengthen the difference between the cumulative error partial image by image co-registration, thereafter, by the OTSU algorithm, the cumulative error partial image is carried out to Threshold segmentation, morphology processing and mark, to extract the moving target in the cumulative error partial image, finally obtain the moving object detection result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310321219.7A CN103473787B (en) | 2013-07-29 | 2013-07-29 | A kind of based on moving target detecting method on the bridge of space geometry relation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310321219.7A CN103473787B (en) | 2013-07-29 | 2013-07-29 | A kind of based on moving target detecting method on the bridge of space geometry relation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103473787A true CN103473787A (en) | 2013-12-25 |
CN103473787B CN103473787B (en) | 2015-08-26 |
Family
ID=49798621
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310321219.7A Expired - Fee Related CN103473787B (en) | 2013-07-29 | 2013-07-29 | A kind of based on moving target detecting method on the bridge of space geometry relation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103473787B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103985117A (en) * | 2014-04-28 | 2014-08-13 | 上海融军科技有限公司 | Method for capturing and determining object based on remote sensing image |
CN108062574A (en) * | 2017-12-31 | 2018-05-22 | 厦门大学 | A kind of Weakly supervised object detection method based on particular category space constraint |
CN108288268A (en) * | 2018-01-23 | 2018-07-17 | 华中科技大学 | Inshore ship detection method in high-resolution remote sensing image based on Projection Analysis |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101620671A (en) * | 2009-08-14 | 2010-01-06 | 华中科技大学 | Method for indirectly positioning and identifying three-dimensional buildings by using riverway landmarks |
US20100097494A1 (en) * | 2008-10-21 | 2010-04-22 | Qualcomm Incorporated | Multimode GPS-Enabled Camera |
CN102201057A (en) * | 2011-05-04 | 2011-09-28 | 华中科技大学 | Method for detecting multiscale space moving object |
-
2013
- 2013-07-29 CN CN201310321219.7A patent/CN103473787B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100097494A1 (en) * | 2008-10-21 | 2010-04-22 | Qualcomm Incorporated | Multimode GPS-Enabled Camera |
CN101620671A (en) * | 2009-08-14 | 2010-01-06 | 华中科技大学 | Method for indirectly positioning and identifying three-dimensional buildings by using riverway landmarks |
CN102201057A (en) * | 2011-05-04 | 2011-09-28 | 华中科技大学 | Method for detecting multiscale space moving object |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103985117A (en) * | 2014-04-28 | 2014-08-13 | 上海融军科技有限公司 | Method for capturing and determining object based on remote sensing image |
CN108062574A (en) * | 2017-12-31 | 2018-05-22 | 厦门大学 | A kind of Weakly supervised object detection method based on particular category space constraint |
CN108062574B (en) * | 2017-12-31 | 2020-06-16 | 厦门大学 | Weak supervision target detection method based on specific category space constraint |
CN108288268A (en) * | 2018-01-23 | 2018-07-17 | 华中科技大学 | Inshore ship detection method in high-resolution remote sensing image based on Projection Analysis |
CN108288268B (en) * | 2018-01-23 | 2020-07-24 | 华中科技大学 | Method for detecting ship landing in high-resolution remote sensing image based on projection analysis |
Also Published As
Publication number | Publication date |
---|---|
CN103473787B (en) | 2015-08-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3519770B1 (en) | Methods and systems for generating and using localisation reference data | |
CN109931939B (en) | Vehicle positioning method, device, equipment and computer readable storage medium | |
CN111462200A (en) | Cross-video pedestrian positioning and tracking method, system and equipment | |
CN106548173B (en) | A kind of improvement no-manned plane three-dimensional information acquisition method based on classification matching strategy | |
KR101569919B1 (en) | Apparatus and method for estimating the location of the vehicle | |
Qu et al. | Landmark based localization in urban environment | |
CN109446973B (en) | Vehicle positioning method based on deep neural network image recognition | |
CN103093459B (en) | Utilize the method that airborne LiDAR point cloud data assisted image mates | |
CN104077760A (en) | Rapid splicing system for aerial photogrammetry and implementing method thereof | |
CN106096497B (en) | A kind of house vectorization method for polynary remotely-sensed data | |
CN108416798B (en) | A kind of vehicle distances estimation method based on light stream | |
Aeschliman et al. | Tracking vehicles through shadows and occlusions in wide-area aerial video | |
CN111598952A (en) | Multi-scale cooperative target design and online detection and identification method and system | |
CN105444741A (en) | Double view window based route characteristic identifying, deviation measuring, and accurate positioning method | |
CN101620672B (en) | Method for positioning and identifying three-dimensional buildings on the ground by using three-dimensional landmarks | |
CN103456027B (en) | Time sensitivity target detection positioning method under airport space relation constraint | |
CN113223064B (en) | Visual inertial odometer scale estimation method and device | |
Xiao et al. | Geo-spatial aerial video processing for scene understanding and object tracking | |
Chen et al. | Improving image-based characterization of road junctions, widths, and connectivity by leveraging OpenStreetMap vector map | |
CN109358315B (en) | Auxiliary target indirect positioning method and system | |
CN103473787B (en) | A kind of based on moving target detecting method on the bridge of space geometry relation | |
CN103456026B (en) | A kind of Ground moving target detection method under highway terrestrial reference constraint | |
CN113971697A (en) | Air-ground cooperative vehicle positioning and orienting method | |
CN106780541B (en) | A kind of improved background subtraction method | |
CN108846363A (en) | A kind of subregion vehicle bottom shadow detection method based on divergence expression scanning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150826 Termination date: 20180729 |