CN103994724A - Method for monitoring two-dimensional displacement and strain of structure based on digital image processing technology - Google Patents
Method for monitoring two-dimensional displacement and strain of structure based on digital image processing technology Download PDFInfo
- Publication number
- CN103994724A CN103994724A CN201410201870.5A CN201410201870A CN103994724A CN 103994724 A CN103994724 A CN 103994724A CN 201410201870 A CN201410201870 A CN 201410201870A CN 103994724 A CN103994724 A CN 103994724A
- Authority
- CN
- China
- Prior art keywords
- target
- circle
- image
- monitoring
- displacement
- 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
Landscapes
- Length Measuring Devices By Optical Means (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a method for monitoring the two-dimensional displacement and the strain of a structure based on the digital image processing technology. The method is characterized in that a circular pattern is adopted as a target, wherein the target is a shooting object used for being shot to acquire a digital image, the displacement or the strain of a corresponding point, to be monitored, of the structure can be determined by processing the digital image, the circular pattern guarantees the non-directional feature of the cross section of the target, namely, the target does not need to be strictly adjusted to be right during monitoring of physical quantities of the structure, in common words, the situation that due to inclination of the target, the circular pattern becomes rhombic from being square like a square pattern is avoided, and the circular pattern can be kept circular all the time; a square pattern with the known side length is adopted as a calibration plate, wherein the calibration plate is a shooting object for calibration of a current vision monitoring system; a circular circle center is selected as the feature of the image of the target to be extracted, wherein the feature is a monitoring calculation point selected for determining the corresponding displacement or strain of the structure from the shot digital image of the target.
Description
Technical field
The present invention discloses a kind of contactless structure two-dimension displacement and strain monitoring method, particularly a kind of contactless structure two-dimension displacement based on Digital Image Processing and the realization flow of strain monitoring method and the method.
Background technology
For guaranteeing structural safety and the people's lives and property and safety, the health monitoring of carrying out structure is very necessary.The content of monitoring structural health conditions is a lot, as load monitoring, temperature monitoring, acceleration monitoring, displacement monitoring, strain monitoring etc.Wherein, the displacement of structure and the strain monitoring important content of monitoring structural health conditions especially.By to building displacement and the strain monitoring of structure, can know whether structure meets design requirement, and accomplish timely correction in the time there is larger displacement or strain; By the displacement to existed building and strain monitoring, can judge that whether structure is in a safe condition, so that can be in time, it be reinforced etc. to processing.In addition, the displacement information of structure still carries out the extremely valuable raw data of structural experiment and analysis, is one of Structure dynamic characteristics.
Displacement structure (strain) monitoring method has: traditional non-image monitoring method and the displacement monitoring method based on digital picture.Wherein, traditional non-image monitoring method comprises again wired, wireless senser monitoring method and contactless monitoring method.But these displacement monitoring methods with regard to current: sensor-based method has contacting of sensor and tested point surely.And for placing sensor in some inaccessible positions of structure, thereby cannot carry out the monitoring of displacement or strain; Secondly, large civil structure generally all needs to place a hundreds of sensor, so the sensor-based system in Practical Project is very complicated, the amount of assigning a work is large; Finally, the method based on wired sensor has too much signal transmssion line and exists, though and be not vulnerable to environmental interference with transmission line data-signal based on wireless senser, the data that often record are incomplete and be mingled with obvious noise.Traditional contactless measurement, as electronic total station and gps satellite mensuration, in experiment and some Practical Projects, also have successful precedent, but they all need to adopt very expensive machinery and equipment and professional person's operation, and the signal collecting is subject to environmental interference large.
Due to shortcoming and the problem of the non-image monitoring method of tradition, many experts and scholars displacement monitoring method based on Digital Image Processing that begins one's study in recent years, and obtained many achievements.As the civil engineering work two-dimension displacement monitoring method based on scaling board ccd image treatment technology of the propositions such as the Zhao Wenguang of the Central China University of Science and Technology, by to the sport car of certain highway bridge and jump car test show the method can measure bridge under dynamic load and action of static load laterally and vertical displacement, and can export in real time.Its visual monitoring system scheme of installation is as shown in Figure 1:
The propositions such as all rare talent of Tongji University utilize the digital picture of the impact point light source of fixed point placement to carry out tunnel transverse and longitudinal deformation monitoring method, show that by simulated experiment the method can obtain the deflection in tunnel, and deformation monitoring result is more accurate, reliable.Meanwhile, this scholar also proposes in the monitoring of actual tunnel deformation, can an imageing sensor be installed in tunnel wall and at its dead ahead fixed target pointolite every suitable distance, thereby realize the deformation monitoring to subway tunnel between station.Its subway tunnel DEFORMATION MONITORING SYSTEM schematic diagram is as shown in Figure 2: a great waves of China Seismological Bureau Engineering Mechanical Institute propose to utilize the digital picture of the double-colored target of black and white to carry out optical displacement measurement, and carried out the displacement measurement test under the measurement of horizontal vibration platform vibration displacement and bridge pier slow test, twice test findings all shows that its optical displacement measurement method proposing can realize the measurement to structure two-dimension displacement well.The double-colored target of its black and white is as shown in Figure 3:
The propositions such as scholar Jong Jae Lee utilize size known and carry out two-dimension displacement monitoring containing the target board of four white dots.They using under black background containing the target board of 4 white dots as reference object, any point in 4 spots carries out the two-dimension displacement monitoring to bridge as monitoring objective.By steel box girder bridge test show to utilize two-dimension displacement visual monitoring method gained bridge displacement time-history curves can with vibration measurement with laser device gained bridge displacement time-history curves quite well.Its target board schematic diagram is as shown in Figure 4:
But above each method still has some problems.First, this type of displacement monitoring method based on digital image processing techniques is in most cases all for structures such as bridge, tunnel, side slopes, but the surface condition of these structures is all comparatively simple, and whether these vision displacement monitoring methods of mentioning more than in the time that structure is comparatively complicated are suitable for and also should do patulous research; The second, can be found out by above-mentioned several vision displacement monitoring methods, the method for their design is only applicable to two dimension (or one dimension) displacement monitoring of structure, but cannot obtain as the strain of another important indicator of structure; The 3rd, above first three planted designed reference object in displacement monitoring method two classes: the target (scaling board) and the LED lamp source that are designed with special pattern.Wherein, special pattern can be designed to chequered with black and white grid or black and white double-colored, but such design makes target (scaling board), cross section has had directivity, while setting it as reference object, must guarantee that the level on target (scaling board) transverse and longitudinal limit is with vertical, place if there is target the situation of inclination, displacement monitoring precision will obviously reduce.Meanwhile, also there is this problem in the method for the proposition such as Jong Jae Lee.In the time adopting LED lamp source as target, requirement for shooting environmental is comparatively strict, preferably complete black shooting environmental, therefore the design limitation of this target the scope of application of its method, it only can be used in night displacement monitoring or the monitoring of subway tunnel internal modification.
Summary of the invention
This method proposes a kind of structure two-dimension displacement and strain monitoring method based on digital image processing techniques, it is characterized in that: target adopts circular pattern, described target refers to by it is taken pictures, obtain digital picture and Digital Image Processing and can determine the displacement of the corresponding tested point of structure or the reference object of dependent variable, the non-directional of target cross section guaranteed in the selection of circular pattern, in the time carrying out the monitoring of structure physical quantity, do not need to accomplish strictly ajusting of target, popular is exactly that circular pattern can not put and tilt to cause square change rhombus because of target as square pattern, it is circular that circle can remain, adopt the known square pattern of the length of side as scaling board, described scaling board refers to the reference object of demarcating for current visual monitoring service system, choose the circular center of circle and extract as target image feature, described feature refers to as determine corresponding displacement structure or the selected monitoring calculation point of dependent variable from the target digital picture of taking.
Described target and scaling board can be made up of paper, presspaper, plank, plastic plate, lightweight sheet steel etc., arrange in structural active to be measured thereby realize it.
Described target and scaling board should be firmly secured to body structure surface, and guarantee during structural vibration, not have relative displacement between itself and structure tested point.
Described scaling board square pattern size can be set to 10cm × 10cm, and color is the black figure of white background.
The color of described target is set as required.As light body structure surface arranges the black figure of white background, dark body structure surface arranges white with black figure, and night, shooting condition, arranged fluorescence or LED lamp source target.
The work of described visual monitoring system calibrating only need to be carried out in the time that two-dimension displacement is monitored, and should before the monitoring of structure two-dimension displacement starts, launch.Require square scaling board used to ajust, in order to avoid scaling board tilts to cause system calibrating error to increase as far as possible.The strain monitoring of structure is without carrying out system calibrating work.
When structure partial strain monitoring, the each circle in target is arranged by a line, and the center of circle of each circle is on a line.
Can not be overlapped between each circular index point when structure partial strain monitoring, and also should leave enough gaps between adjacent two circles.
This method is to realize the scheme of extracting employing in the center of circle to further comprise:
Described target image carries out greyscale transformation and pre-service;
Pretreated target image is carried out to image binaryzation;
Bianry image is carried out to rim detection;
Edge detected image is further processed and rounded edge fine positioning;
Utilize the edge image after fine positioning to carry out center of circle matching;
When carry out be structure partial strain monitoring time, also need be before rounded edge fine positioning except above step, carry out each circular index point affiliated area and cut apart.Find the image range of each circle individualism, and within the scope of this, complete round edge fine positioning and center of circle matching.
Structure two-dimension displacement based on Digital Image Processing and the enforcement of strain monitoring method, the method comprises: scaling board and target are arranged in to structure tested point successively; Form collecting device by digital camera, Digital Video or camera and carry out the shooting of scaling board and target digital picture; Correlated digital image processing techniques by computer realization to captured digital picture, thus resolution and the target circular image center of circle two-dimensional pixel coordinate of current monitoring system obtained; Determine the two-dimension displacement of structure measuring point according to systemic resolution and center of circle two-dimensional pixel coordinate; Determine structure partial strain according to the center of circle two-dimensional pixel coordinate of adjacent target circle.
Preferably, first, to being arranged in the scaling board of structure position to be measured, utilize digital camera repeatedly to take pictures; Secondly, in the situation that ensureing that structure and camera relative position are constant, scaling board is replaced with to target, and by it being taken pictures or making a video recording to obtain the monitoring information of structure measuring point in shooting time; Next, the scaling board image obtaining is carried out to following digital image processing techniques: digital picture reads, two dimension median filter pre-service, image binaryzation based on process of iteration selected threshold and the further denoising of image followed the tracks of based on area.Obtain the length in pixels on square each limit in scaling board, and combine and then determine monitoring system resolution with the known physics length of side.The target image obtaining is carried out to following digital image processing techniques: digital picture reads, two dimension median filter pre-service, image binaryzation based on process of iteration selected threshold, Method Based on Multi-Scale Mathematical Morphology rim detection, the further denoising of image of following the tracks of based on area, edge fine positioning and the matching of the least square method center of circle based on least radius method, extracts the center of circle two-dimensional pixel coordinate in target.Finally, utilize systemic resolution and target center of circle frame difference method displacement Computing Principle, determine structure measuring point two-dimension displacement; Utilize pixel coordinate and the mean strain computing formula in adjacent two centers of circle in target, determine the mean strain between the adjacent measuring point of structure.
The beneficial effect of this method is: adopt the structure two-dimension displacement and the strain monitoring method that propose, can realize the contactless monitoring to structure two-dimension displacement and local train, and monitoring accuracy is higher, can meet the error requirements in Practical Project.Whole monitoring system simple structure, easy for installation, cost is low, is easy to realize.In whole observation process, can not treat geodesic structure and impact, also can realize the dynamic and static monitoring to displacement structure, and significantly change the monitoring of displacement.
The advantage of choosing the center of circle and be target circle feature is that diameter of a circle needn't be known.On the other hand, because displacement structure is relevant with selected circle marker spot diameter with strain monitoring precision---when circular index point increases, monitoring accuracy can improve.So it is too small that target circular diameter can not be selected, conventionally in 50mm left and right.And concrete diameter value needn't be determined, while only needing to guarantee to monitor, target circular diameter used unanimously at every turn.
Beneficial effect of the present invention is: target and the scaling board simple structure of the method design, be easy to realize.On the other hand, designed system calibrating method simply, easily row and degree of accuracy better.Finally, this method can also can be as the local train monitoring of structure as the two-dimension displacement monitoring of structure: when in target only when a circular index point, can carry out the monitoring of structure two-dimension displacement.In the time having multiple circular index point in target, can carry out structure partial strain monitoring.
Brief description of the drawings
Fig. 1 is the monitoring system scheme of installation that the people such as the Zhao Wenguang of the Central China University of Science and Technology propose.
Fig. 2 is the subway tunnel DEFORMATION MONITORING SYSTEM schematic diagram that the people such as all rare talent of Tongji University propose.
Fig. 3 is the double-colored target of black and white that a great waves of China Seismological Bureau Engineering Mechanical Institute propose.
Fig. 4 is the target schematic diagram that the people such as Jong Jae Lee propose.
Fig. 5 is this structure two-dimension displacement and strain monitoring method flow diagram based on Digital Image Processing.
Fig. 6 is this method scaling board schematic diagram.
Fig. 7 is the black figure target of this method white background schematic diagram.
Fig. 8 is this method white with black figure target schematic diagram.
Fig. 9 is this visual monitoring system schematic in simulated experiment.
Figure 10 is the black figure target of this method white background bianry image schematic diagram.
Figure 11 is this method white with black figure target bianry image schematic diagram.
Figure 12 is that this method target image has noise schematic diagram after binaryzation.
Figure 13 is Morphology edge detection effect schematic diagram.
Figure 14 is the Morphology edge detection design sketch that has noise target bianry image.
Figure 15 is the further denoising effect figure of digital picture.
Figure 16 is target circle least square method center of circle fitting result chart.
Figure 17 is strain monitoring target schematic diagram.
Embodiment
Monitor flow process according to shown in Fig. 5.First, carry out the digital image capture of scaling board and target, thereby obtain the raw data of displacement and strain monitoring.Image collecting device adopts general digital camera, video camera or camera.
When carry out be the static displacement monitoring of structure time, can select the mode of low speed continuous shooting to obtain the target digital image in whole monitoring time; When carry out be structure dynamic displacement and strain monitoring time, should select the mode of capture video to obtain the target digital image in whole monitoring time.Both differences are, the former can directly obtain the image of a frame frame, and the raw data that the latter obtains is video.
In the time that the data that collect are video information, converted frames image manipulation should be carried out be converted into the image of a frame frame.
In this method, the extraction of characteristics of image mainly realizes by following 7 steps:
Digital picture reads: the digital picture of being obtained by image capture device belongs to the true color image in bitmap images, claims again RGB image.What use is RGB color space data type, and the color data of its each picture point all includes the component of three passage R, G, B.On the other hand, true color image is again the bitmap images of 24,
Can show 256 (2
8=256) plant tone, wherein minimum value is 0, and maximal value is 255.So, carry out when digital picture reads producing a three-dimensional matrice for true color image, as: m × n × 3.Particularly, every one dimension represents a Color Channel component; Every one dimension is all a two-dimensional array; In each two-dimensional array, element minimum value is 0, and maximal value is 255.Fig. 6～8 also represent target true color image.
Gray level image reads and can obtain one 2 dimension matrixes it, as: m × n.Wherein each element has represented the gradation of image value of corresponding picture point within the scope of certain grey scale change, and concrete grey scale change scope is: 0～255,0 represents brightness minimum, i.e. black; 255 represent brightness maximum, pure white; Between 0～255, show as gray scale.
Between true color image and gray level image, can mutually change.
Digital picture pre-service: the multiplicative noise in digital picture utilizes two dimension median filter method to remove.
Because two dimension median filter method is the basis using each pixel gray-scale value in gray level image as denoising, therefore need to convert read true color image to gray level image before two dimension median filter carrying out.Secondly, due to scaling board used in this method and target figure all very simple, therefore select Filtering Template size 3 × 3.
The enforcement of gray level image two dimension median filter: as starting point, allow in turn Filtering Template slide in gray level image while being positioned at first pixel position of gray level image taking the central point of Filtering Template.Finally, record Filtering Template in the time of current location, the pixel value of the each pixel of former gray level image in Filtering Template, and sort according to the size of value, the pixel value of fetch bit in the middle of sequence is as the pixel value of current processed pixels point (being Filtering Template central point).The mathematic(al) representation of this image pre-processing method is:
Digital picture binaryzation: described method adopts " image segmentation of choosing based on threshold value ".Wherein, threshold value choose employing process of iteration.
The enforcement of image binaryzation: adopt gray scale intermediate value in the tonal range of pretreated gray level image as initial threshold T
0.Whole gray level image matrix naturally can be by this initial threshold T
0be divided into two parts: all gray-scale value is greater than (or equaling) initial threshold T
0picture point set and all gray-scale values be less than initial threshold T
0picture point set.So, from this two parts picture point set, all can draw its gray scale intermediate value T separately
kand T
k+1if both equate, T so
kor T
k+1just the optimal threshold can be defined as image binaryzation time; If both are not etc., so can be by gray-scale value (T
k+ T
k+1threshold value is divided as new matrix area in)/2, repeats former steps, until T
kand T
k+1equate, the iterative process that whole threshold value is chosen just can stop.
Utilize definite optimal threshold, carry out the binaryzation of gray level image: if the gray-scale value of a certain picture point is greater than selected threshold value in gray level image, this gray level is designated as 1 so; If the gray-scale value of a certain picture point is less than selected threshold value in gray level image, this gray level is designated as 0 so.Known image after image binaryzation only remains two gray levels: 1 and 0.Wherein, 1 represents white, and 0 represents black.Digital picture after binaryzation is called bianry image.Obviously, the color of its prospect of bianry image and background depends in former gray level image strong and weak distribution of color between the two, and in the time that foreground color is deeper than background color (as shown in Figure 7), the image after image binaryzation as shown in figure 10; Of light color during in background color (as shown in Figure 8) when prospect, the image after image binaryzation as shown in figure 11.
For prevent through after edge detecting technology after, in image array, there is negative gray level, should by unified target bianry image be form shown in Figure 11.In the time that the image after binaryzation is Figure 10 situation, also tackles this bianry image and carry out inverse processing, make it become form shown in Figure 11.
Morphology edge detection: this method employing Nonlinear Processing technology---Morphology edge detection operator carries out the rim detection of target bianry image, adopts two kinds of square structure element S E
1and SE
2.
Described structural element is that mathematical morphology extracts image information " probe ".The size of structural element generally should be less than pending image and set simple structure.Because reference object in this method is circular target pattern, belong to simple graph, therefore adopt the structural element of 3 × 3 or 5 × 5 general sizes to be enough to.
The designed anti-noise type Method Based on Multi-Scale Mathematical Morphology of this method edge detection operator is as shown in Equation 2:
Wherein, edge represents the rounded edge image array detecting, A represents the true color image of former target to carry out the bianry image matrix after binaryzation, represents closing operation of mathematical morphology, and o represents morphology opening operation, and ⊕ represents dilation operation, SE
1, SE
2represent to be respectively of a size of
Two-dimensional rectangle structural element.First this anti-noise type Method Based on Multi-Scale Mathematical Morphology edge detection operator is to utilize small scale structures element S E
1bianry image matrix A is carried out to closing operation of mathematical morphology, thereby weed out the negative-going noise in bianry image, and obtain new image array ASE
1; Secondly, with ASE
1for objective matrix utilizes large scale structural element SE
2carry out morphology opening operation to weed out remaining positive-going noise in image, and obtain image array ((ASE
1) oSE
2); Next, again utilize large scale structural element SE
2do closed operation and carry out smoothed image, and obtain the primary image matrix target M finally detecting for target rounded edge
2=((ASE
1) oSE
2) SE
2; Finally, to this image array target M
2utilize small scale structures element S E
1image array M after doing dilation operation and can expanding
1.Now, be M
1m before image array and expansion
2the difference operation of image array obtains edge matrix, and the represented image of this matrix is the edge image of former target image.
The rim detection design sketch of described anti-noise type Method Based on Multi-Scale Mathematical Morphology edge detection operator as shown in figure 13.Wherein, the part that gray-scale value is 1 in figure white portion be the edge of rounded edge point position and part residual noise; The part of gray-scale value 0 i.e. black region in figure.
The described target rim detection design sketch containing residual noise is as shown in figure Figure 14.
The advantage of described anti-noise type Method Based on Multi-Scale Mathematical Morphology edge detection operator: the designed Multiscale Edge Detection Operator of this method takes full advantage of that small scale structures element denoising ability is weak but marginal point station-keeping ability is strong, and the strong marginal point station-keeping ability of large scale structural element denoising ability is slightly weaker than the characteristic of small scale structures element, therefore Image Edge-Detection is respond well.Can fully accomplish the objects such as edge extracting is accurate, noise suppression effect is good.
The further denoising of digital picture: the further Denoising Algorithm of the designed digital picture of this method is called " area track algorithm ".Its principle is: though in a width target rounded edge detected image, have picture noise exist, because noise is all very discrete distribution, thus its separately shared area be all very tiny compared to target rounded edge area occupied.So, can carry out using the area of target complete in figure as a key character track identification of target rounded edge, set a suitable area threshold, the target that all areas is less than to this threshold value all thinks noise and rejects, and experimental result shows that this method has good denoising effect.
Described target refers in the target image after Morphology edge detection, the region that the picture point that all gray level is 1 forms.
The further denoising effect figure of digital picture as shown in figure 15.
Marginal point fine positioning: because there is certain width at the edge that utilizes Morphology edge detection operator to obtain, therefore in order to ensure the accuracy of barycenter (center of circle) location, first tackle target rounded edge and carry out fine positioning, dwindle the border width scope in target rounded edge image.
This method is considered target image situation: picture noise most possibly appears at the outside of circle marker point edge but not inner side, i.e. the inner noiseless impact of circular image.Carry out target rounded edge point fine positioning therefore designed " least radius method ".
Its principle is: any point inside circle marker point edge, along searching for directly over it, the pixel coordinate of the pixel that is 1 by first run into gray level is recorded, and this point is first rounded edge point that utilization " least radius method " institute fine positioning arrives.Next, using this point as edge fine positioning starting point, successively rounded edge is carried out to fine positioning in the direction of the clock, and record the corresponding in the drawings two-dimensional pixel coordinate (x of marginal point after whole fine positionings
i, y
i).
Any point inside described circle marker point edge, be chosen as in the method by single order center moments method the tentatively definite center of circle, its pixel coordinate is denoted as (u
c, v
c).Can be by calculating with following formula 3～7:
m
00＝Ε
xΕ
yf(x
i,y
i) (3)
m
10＝Ε
xΕ
yxf(x
i,y
i) (4)
m
01＝Ε
xΕ
yyf(x
i,y
i) (5)
u
c＝m
10/m
00 (6)
v
c＝m
01/m
00 (7)
In above formula, f (x
i, y
i) presentation video function, (x
i, y
i) represent the pixel coordinate of i pixel, m
00, m
10, m
01presentation video function f (x respectively
i, y
i) zeroth order square and two first moments.
Center of circle matching: the quick center point fitting of non-iteration that this method adopts Pratt to propose carries out center of circle matching, its essence is the least square center fitting process of a kind of general equation taking circle as objective function.
The described general equation taking circle as the form of objective function F (A, B, C) is:
The center of circle fitting theory of this method is: by asking parameter A, and B, C makes the value minimum of F (A, B, C) carry out matching central coordinate of circle (u
c, v
c).Because objective function F (A, B, C) is a nonnegative function, this function has minimum value and minimum value fixes on extreme point place, as long as just find the extreme point of objective function can obtain unknown A, B, C, thus further realize central coordinate of circle (u
c, v
c) matching.Because extreme point is to make objective function be those points of zero to the partial derivative of each variable, therefore to this objective function F (A, B, C), its partial derivative is respectively:
Simultaneous solution formula 9～11, just can draw A, B, the value of tri-parameters of C.
Central coordinate of circle fitting formula is:
u
c＝A/-2 v
c＝B/-2 (12)
This method with the advantage of standard equation compared with objective function carries out the matching of the least square method center of circle taking circle is: taking the standard equation of circle when objective function solves its minimum value, ask after partial derivative the high order power of each unknown quantity in gained system of equations to reach three times to each parameter, so calculate very consuming time.And when the objective function that employing formula 8 builds, amount A to be asked, B, the high order power of C is only for once, and computing velocity can promote greatly.
Target circle least square method center of circle fitting effect as shown in figure 16.In figure, red circle position is the center of circle simulating.
Displacement Computing Principle, utilizes least square method to determine after central coordinate of circle, just can directly calculate not the barycenter pixel displacement amount of the lower same circular target of taking in the same time; Combine and just can obtain barycenter actual displacement amount, i.e. the structure measuring point displacement amount of circular target installation position simultaneously with through the resolution R of system calibrating gained.
Strain Computing Principle, carrying out structural strain calculating meeting calculates slightly different with displacement, in a main width target image that is just to take, have multiple circles and have (as Figure 17), now need first to orient the scope of each circle, carrying out topography cuts apart, to find n only to contain circular topography's matrix, the number of n is consistent with the circular number photographing in image.
By obtaining the analysis of Figure 17: although in same target image, have multiple circular existence, between each circle, be again separate do not have associated, therefore can utilize this character to find n target circle and simulate the central coordinate of circle of each circle.Designed idiographic flow is as follows: (1) search does not comprise the edge of hole connected domain, and the contour description of each connected domain out.This is described result and is made up of two parts: B and L.Wherein, B is cell element array, and for storing the edge coordinate value of each connected domain, its matrix length is the number that can form connected domain detecting in image, i.e. circular number; L is mark matrix, and its can be according to the edge coordinate information of each connected domain represented in B matrix, by original image matrix corresponding to the numerical value at the each connected domain marginal position place in B matrix according to certain rule again life value.As: be 1 with values all on relevant position in the corresponding L matrix of first element (being the coordinate figure matrix at first connected domain edge) of cell element matrix B; Be 2 with values all on relevant position in the corresponding L matrix of second element (i.e. the edge coordinate value matrix of second connected domain) of cell element matrix B, by that analogy, finally obtain whole mark matrix L.(2) this expression mode of L matrix has also ensured that the value of each marginal point in same connected domain is all identical.So, can in same a line of L mark matrix, have how many not identical numbers by differentiating, just can determine and in current shooting image, have how many circular index points.(3) in known markers matrix L, the element position in size and the cell element matrix B of element value has relation one to one, the edge coordinate information for each circular target of simultaneously preserving in B matrix, both comprehensive utilizations just can extract the edge coordinate of certain circular target of wanting.After suitably being amplified, this edge extent just can obtain the only local matrix containing a circular target image information.
When having realized after the local segmentation of single circular target image, center point fitting when the each topography being partitioned into just all can adopt displacement monitoring is determined central coordinate of circle, and final location, the center of circle of realizing all circles.Further, the pixel distance between two adjacent circular index points also can calculate; Finally, utilize the change amount of pixel distance before and after malformation between adjacent two circular index points just can obtain the mean strain value of structure in this segment limit
mathematic(al) representation is:
Wherein,
for mean strain between adjacent two circular index points, L is the adjacent two circle marker spacings before malformation, the adjacent two circle marker spacings of the same L after L ' expression malformation.
Visual monitoring system calibrating, needs to utilize scaling board (as Fig. 6) to carry out on-site proving and determines at current camera and the resolution R that treats visual monitoring system under the relative position of geodesic structure before two-dimension displacement monitoring.Scaling board can be arranged in structure tested point position, and should accomplish to ajust in the time placing scaling board as far as possible.After system calibrating is finished, should be in the situation that guaranteeing that between camera and structure, relative position does not change, remove scaling board and target on same position rearranges.
The designed visual monitoring system calibrating of this method idiographic flow is as follows: obtain the known scaling board digital picture of square size by digital camera, video camera or the first-class image capture device of making a video recording 1.; 2. pair scaling board carries out the shooting of continuous several times, is averaged to reduce systematic error thereby realize data; For obtained a series of scaling board images carry out that digital picture reads, the treatment technology of pre-service and binaryzation; 4. obtained scaling board bianry image is carried out to inverse processing, and scaling board bianry image matrix after the inverse treatment technology that carries out the further denoising of digital picture as target image matrix, wherein, the further noise-removed technology of described digital picture has done detailed introduction above; 5. through the scaling board bianry image matrix of further denoising, in whole image array, be only 1 at the part gray-scale value corresponding to square region, and the remainder gray level of matrix is 0.So, can determine four summits of projected square part in scaling board respectively: in whole picture points that gray-scale value is 1, have the picture point that has the minimum row coordinate of maximum row in the picture point that has minimum row maximum column coordinate in the picture point of minimum ranks coordinate, whole picture points that gray-scale value is 1, whole picture points that gray-scale value is 1, and in gray-scale value whole picture points that are 1, have the picture point of maximum ranks coordinate; 6. utilize four apex coordinates of square region to calculate square region along level and vertical pixel length of side value, be designated as respectively L
x1, L
x2, L
y1, L
y2.Unit is pixel (pixel).7. process respectively for several scaling board images, obtain many group square region along level and vertical pixel length of side value.To pixel length of side value and all vertical pixel length of side value averaged respectively, finally can obtain the average level pixel length of side value L of square region to the whole levels that obtain
xwith average vertical pixel length of side value L
y; 8. because actual scaling board square region is known (mm of unit) along level and vertical length of side value, thus utilize the actual length of side value of horizontal direction divided by horizontal direction pixel length of side value can proper anterior optic monitoring system along horizontal direction systemic resolution Rx; Utilize vertical actual length of side value divided by vertical pixel length of side value can proper anterior optic monitoring system along vertical systemic resolution Ry.Visual monitoring systemic resolution unit is: mm/pixel.
Claims (9)
1. structure two-dimension displacement and the strain monitoring method based on digital image processing techniques, it is characterized in that: target adopts circular pattern, described target refers to by it is taken pictures, obtain digital picture and Digital Image Processing and can determine the displacement of the corresponding tested point of structure or the reference object of dependent variable, the non-directional of target cross section guaranteed in the selection of circular pattern, in the time carrying out the monitoring of structure physical quantity, do not need to accomplish strictly ajusting of target, popular is exactly that circular pattern can not put and tilt to cause square change rhombus because of target as square pattern, it is circular that circle can remain, adopt the known square pattern of the length of side as scaling board, described scaling board refers to the reference object of demarcating for current visual monitoring service system, choose the circular center of circle and extract as target image feature, described feature refers to as determine corresponding displacement structure or the selected monitoring calculation point of dependent variable from the target digital picture of taking.
2. structure two-dimension displacement and the strain monitoring method based on digital image processing techniques as claimed in claim 1, is characterized in that: described target and scaling board can be made up of paper, presspaper, plank, plastic plate, lightweight sheet steel etc.
3. the structure two-dimension displacement based on digital image processing techniques and the strain monitoring method as described in claim 1-2, it is characterized in that: described target and scaling board should be firmly secured to body structure surface, and guarantee during structural vibration, not have relative displacement between itself and structure tested point.
4. structure two-dimension displacement and the strain monitoring method based on digital image processing techniques as claimed in claim 1, is characterized in that: the system calibrating of monitoring system carries out before the monitoring of structure two-dimension displacement starts, and square scaling board used should be ajusted.
5. structure two-dimension displacement and the strain monitoring method based on digital image processing techniques as claimed in claim 1, is characterized in that: when structural strain monitoring, the each circle in target is arranged by a line, and the center of circle of each circle is on a line.
6. structure two-dimension displacement and the strain monitoring method based on digital image processing techniques as claimed in claim 1, is characterized in that: can not be overlapped between each circular index point when structure partial strain monitoring, and also should leave enough gaps between adjacent two circles.
7. the structure two-dimension displacement based on digital image processing techniques and the strain monitoring method as described in claim 1-6, is characterized in that: it is as follows that the scheme adopting is extracted in the center of circle:
Described target image carries out greyscale transformation and pre-service;
Pretreated target image is carried out to image binaryzation;
Bianry image is carried out to rim detection;
Edge detected image is further processed and rounded edge fine positioning;
Utilize the edge image after fine positioning to carry out center of circle matching;
When carry out be structure partial strain monitoring time, also need be before rounded edge fine positioning except above step, carry out each circular index point affiliated area and cut apart;
Find the image range of each circle individualism, and within the scope of this, complete round edge fine positioning and center of circle matching.
8. the structure two-dimension displacement based on digital image processing techniques and the strain monitoring method as described in claim 1-7, is characterized in that: scaling board and target are arranged in to structure tested point successively; Form collecting device by digital camera, Digital Video or camera and carry out the shooting of scaling board and target digital picture; Correlated digital image processing techniques by computer realization to captured digital picture, thus resolution and the target circular image center of circle two-dimensional pixel coordinate of current monitoring system obtained; Determine the two-dimension displacement of structure measuring point according to systemic resolution and center of circle two-dimensional pixel coordinate; Determine structure partial strain according to the center of circle two-dimensional pixel coordinate of adjacent target circle.
9. the structure two-dimension displacement based on digital image processing techniques and the strain monitoring method as described in claim 1-7, is characterized in that: first, to being arranged in the scaling board of structure position to be measured, utilize digital camera repeatedly to take pictures; Secondly, in the situation that ensureing that structure and camera relative position are constant, scaling board is replaced with to target, and by it being taken pictures or making a video recording to obtain the monitoring information of structure measuring point in shooting time; Next, the scaling board image obtaining is carried out to following digital image processing techniques: digital picture reads, two dimension median filter pre-service, image binaryzation based on process of iteration selected threshold and the further denoising of image followed the tracks of based on area; Obtain the length in pixels on square each limit in scaling board, and combine and then determine monitoring system resolution with the known physics length of side; The target image obtaining is carried out to following digital image processing techniques: digital picture reads, two dimension median filter pre-service, image binaryzation based on process of iteration selected threshold, Method Based on Multi-Scale Mathematical Morphology rim detection, the further denoising of image of following the tracks of based on area, edge fine positioning and the matching of the least square method center of circle based on least radius method, extracts the center of circle two-dimensional pixel coordinate in target; Finally, utilize systemic resolution and target center of circle frame difference method displacement Computing Principle, determine structure measuring point two-dimension displacement; Utilize pixel coordinate and the mean strain computing formula in adjacent two centers of circle in target, determine the mean strain between the adjacent measuring point of structure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410201870.5A CN103994724B (en) | 2014-05-13 | 2014-05-13 | Structure two-dimension displacement and strain monitoring method based on digital image processing techniques |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410201870.5A CN103994724B (en) | 2014-05-13 | 2014-05-13 | Structure two-dimension displacement and strain monitoring method based on digital image processing techniques |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103994724A true CN103994724A (en) | 2014-08-20 |
| CN103994724B CN103994724B (en) | 2019-07-02 |
Family
ID=51308959
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410201870.5A Expired - Fee Related CN103994724B (en) | 2014-05-13 | 2014-05-13 | Structure two-dimension displacement and strain monitoring method based on digital image processing techniques |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103994724B (en) |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104374538A (en) * | 2014-12-19 | 2015-02-25 | 广州大学 | Multi-dimensional vibration test system |
| CN104463888A (en) * | 2014-12-19 | 2015-03-25 | 中国科学院武汉岩土力学研究所 | Crack width change monitoring device and method |
| CN105371769A (en) * | 2015-11-04 | 2016-03-02 | 西北工业大学 | Method and device of measuring dynamic crack lengths |
| CN105486244A (en) * | 2016-01-28 | 2016-04-13 | 广州大学 | Two-dimensional strain measuring system based on ultrahigh precision digital image measurement |
| CN106123785A (en) * | 2016-07-25 | 2016-11-16 | 肖锐 | A kind of arch dam monitoring system for Hydraulic and Hydro-Power Engineering |
| CN106441138A (en) * | 2016-10-12 | 2017-02-22 | 中南大学 | Deformation monitoring method based on vision measurement |
| CN106595497A (en) * | 2017-01-23 | 2017-04-26 | 中车建设工程有限公司 | Building surface crack real-time monitoring and early warning system and early warning method thereof |
| CN106739050A (en) * | 2016-12-30 | 2017-05-31 | 西安航天精密机电研究所 | The detection method of tire tag position in a kind of Productive Process of Tyre Vulcanizing |
| CN107044844A (en) * | 2015-11-16 | 2017-08-15 | 通用电气公司 | method for monitoring component |
| CN107121126A (en) * | 2017-05-16 | 2017-09-01 | 华中科技大学 | A kind of roadbed settlement monitoring method and system based on image procossing |
| CN109459127A (en) * | 2018-11-27 | 2019-03-12 | 华南农业大学 | One kind being based on the contactless blade wind shake measurement method of MATLAB image procossing |
| CN103994724B (en) * | 2014-05-13 | 2019-07-02 | 滕军 | Structure two-dimension displacement and strain monitoring method based on digital image processing techniques |
| CN109990936A (en) * | 2019-03-12 | 2019-07-09 | 高新兴创联科技有限公司 | High speed railway track stress automated watch-keeping facility and method |
| CN110044289A (en) * | 2019-04-17 | 2019-07-23 | 上海同岩土木工程科技股份有限公司 | The multibreak face convergent deformation video monitoring device in tunnel and method based on autozoom |
| CN110220635A (en) * | 2019-07-12 | 2019-09-10 | 南昌航空大学 | Fluorescence type fluid pressure measurement method |
| CN110234971A (en) * | 2017-07-10 | 2019-09-13 | 泰尔茂株式会社 | Pressure-detecting device and extracorporeal circulation apparatus |
| CN110332903A (en) * | 2019-07-16 | 2019-10-15 | 中国二十冶集团有限公司 | The method of contactless monitoring of structures deformation based on Digital Image Processing |
| CN110596578A (en) * | 2019-08-30 | 2019-12-20 | 国网浙江省电力有限公司电力科学研究院 | Non-contact measuring device for deformation of GIS (gas insulated switchgear) |
| CN111121947A (en) * | 2019-12-18 | 2020-05-08 | 广电计量检测（沈阳）有限公司 | Object vibration measuring method |
| CN111157200A (en) * | 2017-01-25 | 2020-05-15 | 松下知识产权经营株式会社 | Rigidity measuring device and rigidity measuring method |
| CN112233076A (en) * | 2020-09-30 | 2021-01-15 | 石家庄铁道大学 | Structural vibration displacement measurement method and device based on red round target image processing |
| CN112229500A (en) * | 2020-09-30 | 2021-01-15 | 石家庄铁道大学 | Structural vibration displacement monitoring method and terminal equipment |
| CN112254663A (en) * | 2020-10-20 | 2021-01-22 | 常州市建筑科学研究院集团股份有限公司 | Plane deformation monitoring and measuring method and system based on image recognition |
| CN112508982A (en) * | 2020-12-04 | 2021-03-16 | 杭州鲁尔物联科技有限公司 | Method for monitoring displacement of dam in hillside pond based on image recognition |
| CN112697050A (en) * | 2021-01-13 | 2021-04-23 | 南京宥安传感科技有限公司 | Night side slope displacement monitoring system based on luminous body |
| CN112964351A (en) * | 2021-03-01 | 2021-06-15 | 东北林业大学 | Transmission tower vibration displacement measurement method based on center point frame difference method |
| CN113203638A (en) * | 2021-04-20 | 2021-08-03 | 上海交通大学 | Method and system for testing shearing large strain of metal sheet under microscopic scale |
| CN114067533A (en) * | 2021-11-27 | 2022-02-18 | 四川大学 | Geological disaster photographing monitoring and early warning method |
| CN114235239A (en) * | 2021-12-20 | 2022-03-25 | 北京有竹居网络技术有限公司 | Pressure detection method and device and electronic equipment |
| CN118329341A (en) * | 2024-06-12 | 2024-07-12 | 河北交规院瑞志交通技术咨询有限公司 | Experimental device based on road surface bridge design detects |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101566465A (en) * | 2009-05-18 | 2009-10-28 | 西安交通大学 | Method for measuring object deformation in real time |
| CN201364143Y (en) * | 2008-12-30 | 2009-12-16 | 中铁大桥局集团武汉桥梁科学研究院有限公司 | Bridge moving displacement measuring device based on machine vision |
| CN101694373A (en) * | 2009-10-23 | 2010-04-14 | 北京航空航天大学 | Antenna deformation measuring method |
| WO2013060218A1 (en) * | 2011-10-28 | 2013-05-02 | 中联重科股份有限公司 | Arm frame vibration displacement measurement method, system and concrete pump truck with the same |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103994724B (en) * | 2014-05-13 | 2019-07-02 | 滕军 | Structure two-dimension displacement and strain monitoring method based on digital image processing techniques |
-
2014
- 2014-05-13 CN CN201410201870.5A patent/CN103994724B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201364143Y (en) * | 2008-12-30 | 2009-12-16 | 中铁大桥局集团武汉桥梁科学研究院有限公司 | Bridge moving displacement measuring device based on machine vision |
| CN101566465A (en) * | 2009-05-18 | 2009-10-28 | 西安交通大学 | Method for measuring object deformation in real time |
| CN101694373A (en) * | 2009-10-23 | 2010-04-14 | 北京航空航天大学 | Antenna deformation measuring method |
| WO2013060218A1 (en) * | 2011-10-28 | 2013-05-02 | 中联重科股份有限公司 | Arm frame vibration displacement measurement method, system and concrete pump truck with the same |
Non-Patent Citations (1)
| Title |
|---|
| 田国伟等: "基于视频图像处理技术的振动台试验动态位移测量方法", 《世界地震工程》 * |
Cited By (40)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103994724B (en) * | 2014-05-13 | 2019-07-02 | 滕军 | Structure two-dimension displacement and strain monitoring method based on digital image processing techniques |
| CN104463888A (en) * | 2014-12-19 | 2015-03-25 | 中国科学院武汉岩土力学研究所 | Crack width change monitoring device and method |
| CN104374538A (en) * | 2014-12-19 | 2015-02-25 | 广州大学 | Multi-dimensional vibration test system |
| CN105371769B (en) * | 2015-11-04 | 2018-06-12 | 西北工业大学 | A kind of method and device for measuring Dynamic Crack length |
| CN105371769A (en) * | 2015-11-04 | 2016-03-02 | 西北工业大学 | Method and device of measuring dynamic crack lengths |
| CN107044844A (en) * | 2015-11-16 | 2017-08-15 | 通用电气公司 | method for monitoring component |
| CN105486244B (en) * | 2016-01-28 | 2018-11-09 | 广州大学 | The Two-dimensional strain measuring system measured based on superhigh precision digital picture |
| CN105486244A (en) * | 2016-01-28 | 2016-04-13 | 广州大学 | Two-dimensional strain measuring system based on ultrahigh precision digital image measurement |
| CN106123785B (en) * | 2016-07-25 | 2019-06-21 | 海南省水利水电勘测设计研究院 | A kind of arch dam monitoring system for hydraulic and hydroelectric engineering |
| CN106123785A (en) * | 2016-07-25 | 2016-11-16 | 肖锐 | A kind of arch dam monitoring system for Hydraulic and Hydro-Power Engineering |
| CN106441138B (en) * | 2016-10-12 | 2019-02-26 | 中南大学 | The deformation monitoring method of view-based access control model measurement |
| CN106441138A (en) * | 2016-10-12 | 2017-02-22 | 中南大学 | Deformation monitoring method based on vision measurement |
| CN106739050A (en) * | 2016-12-30 | 2017-05-31 | 西安航天精密机电研究所 | The detection method of tire tag position in a kind of Productive Process of Tyre Vulcanizing |
| CN106595497A (en) * | 2017-01-23 | 2017-04-26 | 中车建设工程有限公司 | Building surface crack real-time monitoring and early warning system and early warning method thereof |
| CN111157200A (en) * | 2017-01-25 | 2020-05-15 | 松下知识产权经营株式会社 | Rigidity measuring device and rigidity measuring method |
| CN107121126A (en) * | 2017-05-16 | 2017-09-01 | 华中科技大学 | A kind of roadbed settlement monitoring method and system based on image procossing |
| CN107121126B (en) * | 2017-05-16 | 2020-08-14 | 华中科技大学 | Roadbed settlement monitoring method and system based on image processing |
| CN110234971A (en) * | 2017-07-10 | 2019-09-13 | 泰尔茂株式会社 | Pressure-detecting device and extracorporeal circulation apparatus |
| US11278656B2 (en) | 2017-07-10 | 2022-03-22 | Terumo Kabushiki Kaisha | Pressure measuring device and extracorporeal circulator |
| CN109459127A (en) * | 2018-11-27 | 2019-03-12 | 华南农业大学 | One kind being based on the contactless blade wind shake measurement method of MATLAB image procossing |
| CN109990936A (en) * | 2019-03-12 | 2019-07-09 | 高新兴创联科技有限公司 | High speed railway track stress automated watch-keeping facility and method |
| CN110044289A (en) * | 2019-04-17 | 2019-07-23 | 上海同岩土木工程科技股份有限公司 | The multibreak face convergent deformation video monitoring device in tunnel and method based on autozoom |
| CN110220635A (en) * | 2019-07-12 | 2019-09-10 | 南昌航空大学 | Fluorescence type fluid pressure measurement method |
| CN110332903A (en) * | 2019-07-16 | 2019-10-15 | 中国二十冶集团有限公司 | The method of contactless monitoring of structures deformation based on Digital Image Processing |
| CN110596578A (en) * | 2019-08-30 | 2019-12-20 | 国网浙江省电力有限公司电力科学研究院 | Non-contact measuring device for deformation of GIS (gas insulated switchgear) |
| CN111121947A (en) * | 2019-12-18 | 2020-05-08 | 广电计量检测（沈阳）有限公司 | Object vibration measuring method |
| CN112233076A (en) * | 2020-09-30 | 2021-01-15 | 石家庄铁道大学 | Structural vibration displacement measurement method and device based on red round target image processing |
| CN112229500A (en) * | 2020-09-30 | 2021-01-15 | 石家庄铁道大学 | Structural vibration displacement monitoring method and terminal equipment |
| CN112233076B (en) * | 2020-09-30 | 2022-12-13 | 石家庄铁道大学 | Structural vibration displacement measurement method and device based on red round target image processing |
| CN112229500B (en) * | 2020-09-30 | 2022-05-20 | 石家庄铁道大学 | Structural vibration displacement monitoring method and terminal equipment |
| CN112254663B (en) * | 2020-10-20 | 2022-03-18 | 常州市建筑科学研究院集团股份有限公司 | Plane deformation monitoring and measuring method and system based on image recognition |
| CN112254663A (en) * | 2020-10-20 | 2021-01-22 | 常州市建筑科学研究院集团股份有限公司 | Plane deformation monitoring and measuring method and system based on image recognition |
| CN112508982A (en) * | 2020-12-04 | 2021-03-16 | 杭州鲁尔物联科技有限公司 | Method for monitoring displacement of dam in hillside pond based on image recognition |
| CN112697050A (en) * | 2021-01-13 | 2021-04-23 | 南京宥安传感科技有限公司 | Night side slope displacement monitoring system based on luminous body |
| CN112964351A (en) * | 2021-03-01 | 2021-06-15 | 东北林业大学 | Transmission tower vibration displacement measurement method based on center point frame difference method |
| CN113203638A (en) * | 2021-04-20 | 2021-08-03 | 上海交通大学 | Method and system for testing shearing large strain of metal sheet under microscopic scale |
| CN114067533A (en) * | 2021-11-27 | 2022-02-18 | 四川大学 | Geological disaster photographing monitoring and early warning method |
| CN114235239A (en) * | 2021-12-20 | 2022-03-25 | 北京有竹居网络技术有限公司 | Pressure detection method and device and electronic equipment |
| CN118329341A (en) * | 2024-06-12 | 2024-07-12 | 河北交规院瑞志交通技术咨询有限公司 | Experimental device based on road surface bridge design detects |
| CN118329341B (en) * | 2024-06-12 | 2024-08-30 | 河北交规院瑞志交通技术咨询有限公司 | Experimental device based on road surface bridge design detects |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103994724B (en) | 2019-07-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103994724A (en) | Method for monitoring two-dimensional displacement and strain of structure based on digital image processing technology | |
| US10832478B2 (en) | Method and system for virtual sensor data generation with depth ground truth annotation | |
| CN103971406B (en) | Submarine target three-dimensional rebuilding method based on line-structured light | |
| CN105335966B (en) | Multiscale morphology image division method based on local homogeney index | |
| CN110599537A (en) | Mask R-CNN-based unmanned aerial vehicle image building area calculation method and system | |
| CN102768022B (en) | Tunnel surrounding rock deformation detection method adopting digital camera technique | |
| CN103617426B (en) | Pedestrian target detection method under interference by natural environment and shelter | |
| CN109472822A (en) | Dimension of object measurement method based on depth image processing | |
| CN106156744A (en) | SAR target detection method based on CFAR detection with degree of depth study | |
| CN104463249B (en) | A kind of remote sensing images airfield detection method based on Weakly supervised learning framework | |
| CN104794737B (en) | A kind of depth information Auxiliary Particle Filter tracking | |
| CN101094413A (en) | Real time movement detection method in use for video monitoring | |
| CN101398886A (en) | Rapid three-dimensional face identification method based on bi-eye passiveness stereo vision | |
| CN108764082A (en) | A kind of Aircraft Targets detection method, electronic equipment, storage medium and system | |
| US20220315243A1 (en) | Method for identification and recognition of aircraft take-off and landing runway based on pspnet network | |
| CN102565062B (en) | Method for testing turbidity of liquid based on detection of image gray | |
| CN108492298A (en) | Based on the multispectral image change detecting method for generating confrontation network | |
| CN108805016A (en) | A kind of head and shoulder method for detecting area and device | |
| CN103268496A (en) | Target identification method of SAR (synthetic aperture radar) images | |
| CN104318559A (en) | Quick feature point detecting method for video image matching | |
| CN112902869B (en) | Method and device for adjusting laser plane of rail profile measuring system | |
| CN109410248A (en) | A kind of flotation froth motion feature extracting method based on r-K algorithm | |
| CN106339677A (en) | Video-based railway wagon dropped object automatic detection method | |
| CN116824335A (en) | YOLOv5 improved algorithm-based fire disaster early warning method and system | |
| CN113884011A (en) | Non-contact concrete surface crack measuring equipment and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| 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: 20190702 Termination date: 20210513 |