CN102519965B - Online roadbed compactness detection method based on machine vision - Google Patents

Abstract

The invention discloses an online roadbed compactness detection method based on machine vision and belongs to the field of roadbed compactness detection method. The online roadbed compactness detection method comprises the following steps of: (1) establishing a vision measurement system: a measurement system for one compactness sampling point consists of a CCD ( Charge Coupled Device ) camera, a sampling mark feature point and a mark pile; (2) image processing: the CCD camera collects the image of a compaction site and obtains three crossed shapes of the mark pile through fitting, and the image coordinate of the center of sampling mark feature point and the long axis and short axis parameters of an ellipse are extracted; (3) settlement calculation: the vertical distance between the center of the sampling mark feature point and a selected reference point of the mark pile is obtained through calculation, and the settlement of the compaction operation is obtained; and (4) roadbed compactness calculation: the settlement is substituted into the settlement-compactness mathematical model. The online roadbed compactness detection method takes images as the carrier for detection and information transfer; and as image data is stored in a computer, construction documents can be well saved, and the quality and efficiency of compaction operation are improved.

Description

A kind of Subgrade Compaction online test method based on machine vision

Technical field

The present invention relates to a kind of Subgrade Compaction detection method, more particularly, relate to a kind of Subgrade Compaction online test method based on machine vision.

Background technology

Compactness refers to the ratio of dry density after soil or other road-making material compactings and standard maximum dry density, represents with percent.Compactness is one of key index of subgrade and pavement detection of construction quality, characterizes the density situation after on-the-spot compacting, and compactness is higher, and density is larger, and material monolithic performance is better.Compactness detects the method that conventionally adopts random sampling, on the material rolling, samples, and survey density, static modulus of elasticity reflect compaction quality.In engineering, the method for normally used random sampling detection compactness has core cutter method, sand replacement method, nucleon compactness instrument method, fills with oil process, water bag method etc.In said method, except nucleon compactness instrument method, be destructive test assessment method, and result order of accuarcy depends on manual skill level, be can not continuous sampling assessment method.It is fast that nucleon compactness instrument method has measuring speed, do not destroy the advantages such as soil layer construction, can be used for continuous sampling.But the method price is high, environment for use condition harshness, radiomaterial is harmful in addition, once generation problem also can cause environmental pollution.

Start the beginning of the seventies in last century, earthwork construction amount is increasing, adopt traditional detection methods of compaction degree to be difficult to meet construction demand guiding construction operation in time, the begin one's study online measuring technique of compactness of more external experts, compactness online detection and control technology is the new direction of modern compaction technology, and it is along with the development of hydraulic technique, electron controls technology occurs.On street roller, configure compactness on-line detecting system, immediately detect the compaction of compacting material, thereby improved compacting operation quality and operating efficiency.Compactness on-line detecting system is undoubtedly the indispensable parts of following intellectualized vibratory roller.

20 century 70s, the begin one's study online detection of compactness of the compacting expert of Sweden, they install a three axis accelerometer in vibroll vibrating wheels, after vibrating wheels, hang a micro-wheels and be used for detecting the transmission of vibration from vibrating wheels to micro-wheels, then a seismoreceiver is got off to test the vibration of ground with being embedded in.The signal of receiving by each sensor draws, along with the increase of soil compaction, the signal intensity that the accelerometer in vibrating wheels collects is the most obvious.And carry out experimental study, by every all over the in-site measurement digital proof after compacting be feasible with vibrating wheels acceleration signal as measuring object.Under the enlightenment of this test, the Geodynamics company of Sweden has developed first compactness meter in the world in 1975-1977, and the compacting meeting that early 1980s holds in Paris, enters on compactness meter first as international subject under discussion.After this, domestic and international many scientific research institutions and famous compacting machinary manufacturer participate in this research work in succession, have obtained certain achievement, have developed some products.But domestic and international existing compactness online measuring technique has a common ground, be all that applying vibration wheel acceleration signal carries out mathematical modeling, realize compactness and obtain online.

Machine vision, refers to and utilizes computing machine and some utility appliance to realize people's visual performance, thereby realizes by the extraneous things of two dimensional image perception and objective three-dimensional world, is an emerging subject, is one of interesting research frontier.Along with the development of machine vision technique, occurred a kind of image be used as detect and the means of transmission of information or the new detection method that carrier is used, i.e. the image detecting technique based on machine vision.It is taking contemporary optics as basis, melts the modern detecting that the science and technology such as optoelectronics, computer graphics, information processing, machine vision are integrated.The instrument and equipment of the image detecting technique based on machine vision can be realized intellectuality, digitizing, miniaturization, networking and multifunction, possess online detection, real-time analysis, the real-time ability of controlling, obtain extensive concern and application in fields such as military affairs, industry, medical science.

On compacting equipment, adopt vision measurement technology to detect online Subgrade Compaction, it is a kind of road quality Non-Destructive Testing new technology, can make staff in compacting process, detect in real time compacting situation, control compaction quality, thereby ensure that subgrade and pavement obtains abundant compacting under minimum number of rolling, avoid the phenomenons such as the not enough or undue compacting of compacting, there is obvious Social benefit and economic benefit.Increasing measuring point, can also record two operation quality parameters of irregularity degree and overlapping width, is that other are incomparable at compactness line detecting method.In addition, can also use the image information of acquisition, instruct safe construction.Therefore, use and combine the multi-disciplinary visible sensation method acquisition such as mechanics, mechanics, electronics and digital signal processing compacting operation mass parameter, there is suitable theoretical research and be worth and wide application prospect.

1. prior art one related to the present invention

The technical scheme of prior art one: People's Republic of China's industry standard " highway subgrade road surface on-the-spot test code " (JTJ059-95) specifies, Subgrade Compaction can adopt the sand replacement method of digging pit (T0921-95), Nuclear Gauge method (T0922-95) and core cutter method (T0923-95) to measure.People's Republic of China's standard " vibroll method for testing performance " (GB/T4478-1995) specifies, selects core cutter method to measure compactness, the consolidation effect of checking street roller.Lower mask body narration core cutter method is measured compactness step:

(1) compaction test.According to relevant test methods, test samples is carried out to compaction test by same material, obtain maximum dry density ρ _cand optimum moisture content;

(2) clean and change to, take cutting ring mass M ₂, be accurate to 0.1g;

(3) by clean sampling spot cleaning, and compacted lift scalped to surface and float and irregular part, reach certain depth, after cutting ring is laid, can meet the requirements of the degree of depth that fetches earth, but must not be by lower floor's disturbance;

(4) cutting ring mouth down is placed in sampling spot, covers cutting ring lid, and hammering cutting ring lid makes cutting ring reach sampling depth, notes keeping cutting ring vertically downward in process;

(5) remove cutting ring lid, cutting ring and sample are dug out with pick;

(6) with repair native cutter from limit to soil more than middle cancellation cutting ring two ends, detect until equating with ruler;

(7) clean cutting ring outer wall, weigh in the balance and take out cutting ring and sample total mass M ₁, be accurate to 0.1g;

(8) in cutting ring, take out sample, get representative sample, measure its water cut w.

(9) measuring point compactness is calculated:

$\mathrm{\ρ}=\frac{4\×(M_1-M_2)}{\mathrm{\π}\·d^2\·h}$

${\mathrm{\ρ}}_d=\frac{\mathrm{\ρ}}{1+0.01w}$

$k=\frac{{\mathrm{\ρ}}_d}{{\mathrm{\ρ}}_c}\×100$

In formula: ρ---sample wet density, g/cm ³;

ρ _d---sample dry density, g/cm ³;

M ₁---cutting ring and sample add up to weight, g;

M ₂---cutting ring weight, g;

D---cutting ring diameter, cm;

H---cutting ring height, cm;

W---water cut, %;

ρ _c---maximum dry density, g/cm ³;

K---measuring point compactness, %.

The shortcoming of prior art one: above-mentioned detection method exists following shortcoming,

(1) this assay method need to destroy soil layer construction, and too much sampling affects roadbed performance, and the fewer data confidence level of sampled point is lower;

(2) in compacting process, can not measure and assess compaction state, can only after compacting finishes, detect;

(3) test findings accuracy depends on operating personnel's skills involved in the labour;

(4) workman's labour intensity is larger;

(5) detection speed is slow, cannot realize online detection.

The sand replacement method method step of digging pit that " highway subgrade road surface on-the-spot test code " mentioned in (JTJ059-95) is similar to above-mentioned core cutter method, also has above-mentioned weak point.Nuclear Gauge method can realize online detection, but the method may cause nuclear leakage, and threat personnel are healthy, can cause environmental pollution.

2. prior art two related to the present invention

The technical scheme of prior art two: Fig. 1 is the vibroll-soil mathematical model obtaining based on a series of hypothesis, the paper of writing from Li Xishan etc. " research of vibroll compactness instrument ".Reach a conclusion by Numerical Simulation Analysis, vibrating wheels acceleration amplitude is with soil stiffness K ₂increase and become large; On the contrary, vibrating wheels acceleration amplitude is with soil damping C ₂increase and diminish.Be fruitful and point out, in actual compacting process, along with the carrying out of compacting, compactness increases gradually, and it is large that the rigidity of soil also becomes gradually, and damping diminishes gradually.Accordingly, be not difficult to draw that acceleration amplitude increases and becomes large conclusion with compactness, the two has positive correlation.The method that adopts this principle to detect compactness is called amplitude ratioing technigue.Concrete steps:

(1) arrange acceleration transducer in vibrating wheels appropriate position, and configure signal acquiring system;

(2) in selected certain bid section of engineering certain limit as test section, carry out the subgrade compaction operation of 8-12 time according to relevant code, and while end, adopt the compactness of core cutter method determination test section the 2nd, 6,8 times (carrying out 12 times compactings also should comprise the 10th, 12 times) compactings;

(3) acceleration signal of compacting operation process is analyzed, obtained (carrying out 12 times compactings also should comprise the 10th, 12 times) acceleration signal amplitude the 2nd, 6,8 times;

(4) to carrying out data fitting by compactness data and acceleration amplitude data that core cutter method obtains, set up the mathematical model that characterizes compactness with acceleration amplitude;

(5) this project subgrade engineering forecasts construction quality by this mathematical model.

The shortcoming of prior art two: at present, the positive correlation of acceleration amplitude and compactness is too controversial, and applicant has organized and implemented repeatedly vibrating compacting test during majoring in master, find the carrying out along with compacting, compactness is increasing gradually, but but there is Novel presentation in acceleration signal, different the causing of the parameter such as grating, water cut that may be native.That is to say, along with the difference of laminated material character, such as the sandy soil of northern Shensi, the amplitude of acceleration not necessarily increases and becomes large with compactness, thereby has affected the compactness meter of developing according to this principle.In addition, this compactness meter in use must be before each project starts chosen in advance test section, carry out mathematical modeling (some achievement is called demarcation), this has just affected its broad applicability.

3. prior art three related to the present invention

The technical scheme of prior art three: in compacting operation, the compacting initial stage, vibration frequency was mainly fundamental frequency because soil is softer, can not produce higher hamonic wave; But along with the carrying out of compacting, soil compaction rate increases, and can produce higher hamonic wave.Be that high-order harmonic generation and signal distortion degree and compactness have close contacting, in order to disclose the objective law of the two, it is upper that sight has been gathered harmonic ratio HVR by domestic and international many scholars, i.e. fundamental frequency amplitude and this parameter of secondary harmonic amplitude ratio.Fig. 2 quotes from Deutsche Bundespatent (DE3308476/AI).Can be reached a conclusion by Fig. 2, harmonic ratio HVR changes along with compacting counting is increased according to certain rule.The method of carrying out compactness detection according to this principle is called harmonic ratio method.Concrete steps are as follows:

(1) arrange acceleration transducer in vibrating wheels appropriate position, and configure signal acquiring system;

(2) in selected certain bid section of engineering certain limit as test section, carry out the subgrade compaction operation of 8-12 time according to relevant code, and while end, adopt the compactness of core cutter method determination test section the 2nd, 6,8 times (carrying out 12 times compactings also should comprise the 10th, 12 times) compactings;

(3) acceleration signal of compacting operation process is analyzed, obtained (carrying out 12 times compactings also should comprise the 10th, 12 times) acceleration signal harmonic ratio HVR the 2nd, 6,8 times;

(4) to carrying out data fitting by compactness data and acceleration amplitude data that core cutter method obtains, set up the mathematical model that characterizes compactness with acceleration amplitude;

(5) this project subgrade engineering forecasts construction quality by this mathematical model.

The shortcoming of prior art three: first, the applicability of this compactness meter in some new debulking methods need checking, as the good proposition chaos of China Agricultural University's dragon cloud compact technique, vibrating wheels acceleration signal has distribution in wide range, and harmonic components is not fairly obvious.The graceful protecting against shock vibrating compaction method of proposition and the impact shock compound rolling method that the poplar people of Chang An University phoenix proposes of waiting of the Liu Xiao of Chang An University, acceleration signal is to be also distributed in very wide frequency range.

Summary of the invention

1. the technical matters that invention will solve

The invention reside in and overcome deficiency of the prior art, a kind of Subgrade Compaction detection method based on machine vision is proposed, settle the marker peg of some along the line at construction section, the mark of arranging regular geometry at road surfaces tested point is as unique point, obtain fast section high precision image, adopt simple image processing algorithm to obtain the positional information of unique point in the vertical direction, relatively obtain this compaction and subsidence amount with the position data of measuring gained last time again, bring calculated with mathematical model into and obtain compactness.

2. technical scheme

For achieving the above object, technical scheme provided by the invention is:

A kind of Subgrade Compaction online test method based on machine vision of the present invention, the steps include:

(1) set up vision measurement system:

The compactness of a compactness sampled point is measured, its measuring system is made up of 1 CCD camera, 1 sampling flag sign point and 1 marker peg, when multiple compactness sampled point sampling, corresponding increase sampling flag sign point and marker peg quantity, wherein, marker peg amounts to and divides three layers, and every layer is " ten " font, is designated as respectively from left to right face face and face the angle point of these three faces is labeled as 1,2,3 successively ..., 12;

(2) image processing:

CCD collected by camera compacting image scene, then extracts the image coordinate that three layers of marker pegs amount to 36 points, and matching obtains three " ten " fonts of marker peg; Extract image coordinate and circular oval major semi-axis and the oval minor semi-axis Pixel Dimensions size parameter of sampling after the imaging of flag sign point of sampling flag sign dot center;

(3) settling amount calculates:

According to full-size(d) and the picture size of " ten " font full-size(d) and picture size and sampling flag sign point, calculate sampling flag sign dot center to the vertical range between the selected reference point of marker peg, previous vertical range is compared with it, obtains the settling amount of this compacting operation;

(4) Subgrade Compaction is calculated:

Settling amount substitution settling amount-compactness mathematical model that step (3) is obtained, can calculate the compactness that roadbed is new.

Further, the marker peg in step (1) is arranged near compactness sampled point, and this compactness sampled point and marker peg are within the visual field of CCD camera, and wherein, marker peg is arranged in construction section and goes up along the line, face face and face be parallel to each other, face perpendicular with construction road surface; Compactness sampling flag sign point is the mode by spray paint, at road surfaces compactness sample point spraying circle marker as unique point.

Further, in the imaging model of step (2) image processing, establish O _c-X _cy _cz _cfor CCD camera coordinates system, O _cfor CCD camera photocentre, Z _caxle and CCD camera optical axis coincidence, by world coordinate system (O _w-X _wy _wz _w) overlap with CCD camera coordinates system, I is picture plane, and O-XY is that image physical coordinates is, and o-uv is image pixel coordinate system, P _ijfor face (i=1,2,3) upper j angle point (j=1,2 ..., 12), corresponding picture point P _ij' represent P ₀be artificial target's compactness sampling flag sign point, be the measuring point of settling amount, picture point is designated as P ₀'; Set following parameter or geometric condition:

A) P _ijat O _c-X _cy _cz _con (X for coordinate _c ^ij, Y _c ^ij, Z _c ⁱ) represent;

B) P _ij' (X for coordinate on image coordinate system O-XY ^ij, Y ^ij) represent;

C) P _ij' be (the u for coordinate on o-uv at pixel coordinate ^ij, v ^ij) represent;

D) focal length of camera represents with f;

E) pixel is respectively dx and dy at imaging plane X and Y-direction size;

F) coordinate of the initial point of coordinate system O-XY on image pixel coordinate system o-uv is (u ₀, v ₀);

If g) line segment p _ijp _ij+1in camera coordinates system, be vertical direction, p so _ij+1p _ij+2must, in horizontal direction, might as well suppose p _ijp _ij+1for the line segment of vertical direction in camera coordinates system, wherein: P _ijp _ij+1angle point P _ijwith angle point P _ij+1line, P _ij+1p _ij+2angle point P _ij+1with angle point P _ij+2line;

Get P _ij, P _ij+1and P _ij+2three points, in conjunction with imaging model, its imaging process is as follows:

According to pinhole imaging system principle, obtain following equation and set up:

$\left\{\begin{array}{c}{X}^{\mathrm{ij}}=f\·({X_c}^{\mathrm{ij}}/{Z_c}^i)\\ Y^{\mathrm{ij}}=f\·({Y_c}^{\mathrm{ij}}/{Z_c}^i)\end{array}\right.---\left(0.1\right)$

$\left\{\begin{array}{c}{u}^{\mathrm{ij}}=X^{\mathrm{ij}}/\mathrm{dx}+u_0\\ v^{\mathrm{ij}}=Y^{\mathrm{ij}}/\mathrm{dy}+v_0\end{array}\right.---\left(0.2\right)$

Note point P _ijwith a P _ij+1distance between two points is l ₁, some P _ij' and put P _ij+1' two pel spacing is from being l ₁', have:

$\left\{\begin{array}{c}{l}_1={Y_c}^{\mathrm{ij}}-{Y_c}^{\mathrm{ij}+1}\\ {l_1}^{\′}=v^{\mathrm{ij}}-v^{\mathrm{ij}+1}=\frac{f}{\mathrm{dy}}\·\frac{{Y_c}^{\mathrm{ij}}-{Y_c}^{\mathrm{ij}+1}}{{Z_c}^i}\end{array}\right.---\left(0.3\right)$

β is compared in the measurement of definition vertical direction _⊥:

${\mathrm{\β}}_{\⊥}=\frac{l_1}{{l_1}^{\′}}=\frac{{Z_c}^i}{f/\mathrm{dy}}---\left(0.4\right)$

β is compared in the measurement that can define equally horizontal direction _-:

${\mathrm{\β}}_{-}=\frac{{Z_c}^i}{f/\mathrm{dx}}---\left(0.5\right)$

Above-mentioned measurement is than being physical quantity for characterization of visual measuring system relation between given distance testee physical dimension and Pixel Dimensions;

Face there are six to be parallel to X _cthe limit of axle, be horizontal sides, after imaging, can obtain the measurement ratio of six horizontal directions, Chen Xiangwei points out in its doctorate paper " research of Computer Vision Inspection of Mechanical Part gordian technique ", in visual field, the imaging of diverse location place obtains and measures ratio, can eliminate the distortion of vision system, therefore, using the mean value of the measurement ratio of six horizontal directions as face the measurement ratio of position horizontal direction, uses symbol represent, same, can obtain face the measurement ratio of position horizontal direction face the measurement ratio of position horizontal direction face there are six to be parallel to Y _cthe limit of axle, is vertical edge, can obtain the measurement ratio of six vertical directions after imaging, and by the method for averaging, calculating is appeared too the measurement ratio of position vertical direction in like manner, obtain face the measurement ratio of position vertical direction face vertical square of position to measurement ratio Δ ₁₂and Δ ₁₃be two process variable, represent respectively the distance between first cross and second cross, and distance between first cross and the 3rd cross; C1 represents variable Δ ₁₂value; C2 represents variable Δ ₁₃value, have:

$\left\{\begin{array}{c}{{\stackrel{\‾}{\mathrm{\β}}}_{-}}^1=\frac{{Z_c}^1}{f/\mathrm{dx}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{-}}^2=\frac{{Z_c}^2}{f/\mathrm{dx}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{-}}^3=\frac{{Z_c}^3}{f/\mathrm{dx}}\\ {Z_c}^1={Z_c}^2+{\mathrm{\Δ}}_{12}\\ {z_c}^1={z_c}^3+{\mathrm{\Δ}}_{13}\\ {\mathrm{\Δ}}_{12}=C_1\\ {\mathrm{\Δ}}_{13}=C_2\end{array}\right.---\left(0.6\right)$

$\left\{\begin{array}{c}{{\stackrel{\‾}{\mathrm{\β}}}_{\⊥}}^1=\frac{{Z_c}^1}{f/\mathrm{dy}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{\⊥}}^2=\frac{{Z_c}^2}{f/\mathrm{dy}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{\⊥}}^3=\frac{{Z_c}^3}{f/\mathrm{dy}}\\ {Z_c}^1={Z_c}^2+{\mathrm{\Δ}}_{12}\\ {z_c}^1={z_c}^3+{\mathrm{\Δ}}_{13}\\ {\mathrm{\Δ}}_{12}=C_1\\ {\mathrm{\Δ}}_{13}=C_2\end{array}\right.---(0.7)$

Solve the approximate solution that overdetermined equation group (0.6) can obtain parameter f/dx, solve overdetermined equation group (0.7) and can obtain the approximate solution of parameter f/dy, circular feature point P ₀after imaging, be an ellipse, choose the measurement that is parallel to the diameter calculated characteristics point place horizontal direction of imaging plane on circle and compare β _- ⁰, according to formula (0.5) can obtain the sampling coordinate Z at flag sign point place _c ⁰.

Further, in step (3), the implementation procedure that settling amount visible sensation method is measured is as follows: selected P ₁₁for reference point, utilize pixel coordinate and measure than calculating P ₀distance between point and reference point on vertical direction, i.e. P ₁₁point and P ₀point Yc coordinate difference, just can obtain the sedimentation value of measuring point after each compacting;

Set up following system of equations according to formula (0.1) and formula (0.2):

$\left\{\begin{array}{c}{v}^{11}=\frac{{Y_c}^{11}}{{Z_c}^1}\·f/\mathrm{dy}+v_0\\ v^{12}=\frac{{Y_c}^{12}}{{Z_c}^1}\·f/\mathrm{dy}+v_0\\ {Y_c}^{12}=f\left({Y_c}^{11}\right)\\ \·\·\·\·\·\·\end{array}\right.---\left(0.8\right)$

Solving equations (0.8) can obtain parameter Y _c ^1j(j=1,2 ... 12) and parameter v ₀, substitution formula (0.9) can be to parameter Y _c ⁰solve,

$v^0=\frac{{Y_c}^0}{{Z_c}^0}\·f/\mathrm{dy}+v_0---\left(0.9\right)$

So, subsidence survey point, flag sign point P namely samples ₀with reference point P ₁₁at Y _con axle, coordinate difference is Δ=Y _c ⁰-Y _c ¹¹, be the distance of in the vertical direction between measuring point and reference point, roadbed again after a compacting, by obtain a new distance value Δ ', after twice compacting, subgrade settlement is Δ '-Δ.

Further, in step (4), in compactness measuring process, because water environment, the particle voids of soil change, before and after compacting operation, dry density and the wet density of soil all change, but the nt wt net weight of laying soil meets the mass conservation, before and after compacting operation, do not change, have following formula to set up:

h'·s·ρ _d'＝h·s·ρ _d

In formula:

Laying thickness before h---compacting operation, cm;

ρ _d---dry density before compacting operation, kg/m ³

H'---laying thickness after a compacting, cm;

ρ _d'---laying dry density after a compacting, kg/m ³

S---laying thickness area, m ²;

, can calculate compactness according to following formula:

$K^{\′}=K\·\frac{h}{h-\mathrm{\Δ}}---\left(0.10\right)$

In formula:

Laying thickness before h---compacting operation, cm;

Δ---the settling amount after a compacting, cm;

The compactness of laying before K---compacting operation;

K'---the compactness after a compacting;

When being roadbed, laying thickness controls parameter for aerial one, initial compactness is measured by traditional method, and the settling amount that initial compactness, laying thickness and step (3) are calculated is brought formula (0.10) into can calculate the new compactness of roadbed after a compacting operation.

3. beneficial effect

Adopt technical scheme provided by the invention, compared with existing known technology, there is following remarkable result:

(1) a kind of Subgrade Compaction online test method based on machine vision of the present invention, against existing technologies one, the present invention is the carrier as detection and transmission of information with image, be stored in computing machine, can realize online detection, the error of having avoided manual operation to bring, has reduced labour intensity, can utilize in addition video information to instruct safe construction;

(2) a kind of Subgrade Compaction online test method based on machine vision of the present invention, against existing technologies two, the present invention meets people's law of cognition, realize people's visual performance by computing machine and utility appliance, thereby realize the numerical value with respect to sedimentation after compacting last time by laying after two dimensional image perception compacting each time, the increase of sedimentation means the increase of soil layer compaction rate, avoid the deficiency of prior art two, disclose at present Vibration Compaction Mechanism not yet completely, people are not comprehensive to the understanding of vibrating wheels acceleration signal, still can not explain the abnormal conditions of acceleration signal,

(3) a kind of Subgrade Compaction online test method based on machine vision of the present invention, against existing technologies three, detection limit of the present invention be after compacting each time laying with respect to compaction and subsidence value last time, and sedimentation increases not polysemy of sign soil layer compaction rate increase, therefore the compact technique adopting in detection method and construction is separate, the subgrade compaction operation that this detection method goes for new method detects, and is applicable to vibroll unlike prior art three;

(4) the present invention, taking image as detecting and transmission of information carrier, can obtain the compactness with data accurate description, and data are stored in to be calculated above, contributes to preserve construction information;

(5) the present invention has realized Subgrade Compaction and has detected online, thereby has improved compacting operation quality and operating efficiency;

(6) accuracy of detection of the present invention is mainly relevant with the model accuracy of setting up, and does not need movement and location to camera to do special processing;

(7) the present invention suitably increases marker peg quantity, can realize long distance construction section and detect online.

Brief description of the drawings

Fig. 1 is the vibroll-soil mathematical model obtaining based on a series of hypothesis in the paper write such as prosperous mountain;

Fig. 2 is the curve map that in prior art three, disclosed harmonic ratio HVR changes along with compacting counting increase;

Fig. 3 is the schematic diagram of vision testing system of the present invention;

Fig. 4 is marker peg three-decker schematic diagram of the present invention;

Fig. 5 is the present invention program's imaging model.

Label declaration in schematic diagram:

1-CCD camera; 2-sampling flag sign point; 3-marker peg.

Embodiment

Below in conjunction with accompanying drawing, the invention will be further described.

As shown in Figure 3, vision measurement system is made up of 1 CCD camera, 1 marker peg and 1 compactness sampling flag sign point, need to measure multiple compactness sampled points, the quantity of corresponding increase marker peg and sampling flag sign point quantity.Marker peg is three-decker, and every layer is " ten " font, but size difference, as shown in Figure 4.Marker peg make in need to ensure three layers parallel, marker peg install time need to ensure that the face of " ten " word is vertical with road surface.Sampling flag sign point is for circular, and size is not done special requirement, and the mode of available paint spraying realizes.

Shown in Fig. 5, be the mathematical model of this measuring system, wherein O _c-X _cy _cz _cfor camera coordinates system, O _cfor camera photocentre, Z _caxle and camera optical axis coincidence, get world coordinate system (O _w-X _wy _wz _w) overlap with camera coordinates system, I is picture plane, and O-XY is that image physical coordinates is, and o-uv is image pixel coordinate system, P _ijfor face (i=1,2,3) upper j angle point (j=1,2 ..., 12), corresponding picture point P _ij' represent P ₀be artificial target's unique point, be the measuring point of settling amount, picture point is designated as P ₀'.

Below in conjunction with embodiment, the invention will be further described.

Embodiment

A kind of Subgrade Compaction online test method based on machine vision of the present embodiment, the steps include:

(1) set up vision measurement system:

The compactness of a compactness sampled point is measured, and its measuring system is made up of 1 CCD camera, 1 sampling flag sign point and 1 marker peg, and wherein, marker peg amounts to and divides three layers, and every layer is " ten " font, is designated as respectively from left to right face face and face the angle point of these three faces is labeled as 1,2,3 successively ..., 12;

Marker peg in step (1) is arranged near compactness sampled point, and this compactness sampled point and marker peg are within the visual field of CCD camera, and wherein, marker peg is arranged in construction section and goes up along the line, face face and face be parallel to each other, face perpendicular with construction road surface; Compactness sampling flag sign point is the mode by spray paint, at road surfaces compactness sample point spraying circle marker as unique point.

(2) image processing:

CCD collected by camera compacting image scene, then extracts the image coordinate that three layers of marker pegs amount to 36 points, and matching obtains three " ten " fonts of marker peg; Extract image coordinate and circular oval major semi-axis and the oval minor semi-axis Pixel Dimensions size parameter of sampling after the imaging of flag sign point of sampling flag sign dot center;

In the imaging model of step (2) image processing, establish O _c-X _cy _cz _cfor CCD camera coordinates system, O _cfor CCD camera photocentre, Z _caxle and CCD camera optical axis coincidence, by world coordinate system (O _w-X _wy _wz _w) overlap with CCD camera coordinates system, I is picture plane, and O-XY is that image physical coordinates is, and o-uv is image pixel coordinate system, P _ijfor face (i=1,2,3) upper j angle point (j=1,2 ..., 12), corresponding picture point P _ij' represent P ₀be artificial target's compactness sampling flag sign point, be the measuring point of settling amount, picture point is designated as P ₀'; Set following parameter or geometric condition:

A) P _ijat O _c-X _cy _cz _con (X for coordinate _c ^ij, Y _c ^ij, Z _c ⁱ) represent;

B) P _ij' (X for coordinate on image coordinate system O-XY ^ij, Y ^ij) represent;

C) P _ij' be (the u for coordinate on o-uv at pixel coordinate ^ij, v ^ij) represent;

D) focal length of camera represents with f;

E) pixel is respectively dx and dy at imaging plane X and Y-direction size;

F) coordinate of the initial point of coordinate system O-XY on image pixel coordinate system o-uv is (u ₀, v ₀);

If g) line segment p _ijp _ij+1in camera coordinates system, be vertical direction, p so _ij+1p _ij+2must, in horizontal direction, might as well suppose p _ijp _ij+1for the line segment of vertical direction in camera coordinates system, wherein: P _ijp _ij+1angle point P _ijwith angle point P _ij+1line, P _ij+1p _ij+2angle point P _ij+1with angle point P _ij+2line;

Get P _ij, P _ij+1and P _ij+2three points, in conjunction with imaging model, its imaging process is as follows:

According to pinhole imaging system principle, obtain following equation and set up:

$\left\{\begin{array}{c}{X}^{\mathrm{ij}}=f\·({X_c}^{\mathrm{ij}}/{Z_c}^i)\\ Y^{\mathrm{ij}}=f\·({Y_c}^{\mathrm{ij}}/{Z_c}^i)\end{array}\right.---\left(0.1\right)$

$\left\{\begin{array}{c}{u}^{\mathrm{ij}}=X^{\mathrm{ij}}/\mathrm{dx}+u_0\\ v^{\mathrm{ij}}=Y^{\mathrm{ij}}/\mathrm{dy}+v_0\end{array}\right.---\left(0.2\right)$

Note point P _ijwith a P _ij+1distance between two points is l ₁, some P _ij' and put P _ij+1' two pel spacing is from being l ₁', have:

$\left\{\begin{array}{c}{l}_1={Y_c}^{\mathrm{ij}}-{Y_c}^{\mathrm{ij}+1}\\ {l_1}^{\′}=v^{\mathrm{ij}}-v^{\mathrm{ij}+1}=\frac{f}{\mathrm{dy}}\·\frac{{Y_c}^{\mathrm{ij}}-{Y_c}^{\mathrm{ij}+1}}{{Z_c}^i}\end{array}\right.---\left(0.3\right)$

β is compared in the measurement of definition vertical direction _⊥:

${\mathrm{\β}}_{\⊥}=\frac{l_1}{{l_1}^{\′}}=\frac{{Z_c}^i}{f/\mathrm{dy}}---\left(0.4\right)$

β is compared in the measurement that can define equally horizontal direction _-:

${\mathrm{\β}}_{-}=\frac{{Z_c}^i}{f/\mathrm{dx}}---\left(0.5\right)$

Above-mentioned measurement is than being physical quantity for characterization of visual measuring system relation between given distance testee physical dimension and Pixel Dimensions;

Face article six, be parallel to X _cthe limit of axle, is horizontal sides, can obtain the measurement ratio of six horizontal directions after imaging, using the mean value of the measurement ratio of six horizontal directions as face the measurement ratio of position horizontal direction, uses symbol represent, same, can obtain face the measurement ratio of position horizontal direction face the measurement ratio of position horizontal direction face there are six to be parallel to Y _cthe limit of axle, is vertical edge, can obtain the measurement ratio of six vertical directions after imaging, and by the method for averaging, calculating is appeared too the measurement ratio of position vertical direction in like manner, obtain face the measurement ratio of position vertical direction face vertical square of position to measurement ratio

When image is processed: measure face with vernier caliper face and face upper " ten " length on all limits of font, and carry out image processing and obtain face face and face it is upper that " ten " length in pixels on all limits of font, calculates face the measurement ratio of position horizontal direction face the measurement ratio of position horizontal direction face the measurement ratio of position horizontal direction; Face the measurement ratio of position vertical direction face the measurement ratio of position vertical direction face vertical square of position to measurement ratio

Note Δ ₁₂and Δ ₁₃be two process variable, represent respectively the distance between first cross and second cross, and distance between first cross and the 3rd cross; C1 represents variable Δ ₁₂value; C2 represents variable Δ ₁₃value, have:

$\left\{\begin{array}{c}{{\stackrel{\‾}{\mathrm{\β}}}_{-}}^1=\frac{{Z_c}^1}{f/\mathrm{dx}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{-}}^2=\frac{{Z_c}^2}{f/\mathrm{dx}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{-}}^3=\frac{{Z_c}^3}{f/\mathrm{dx}}\\ {Z_c}^1={Z_c}^2+{\mathrm{\Δ}}_{12}\\ {z_c}^1={z_c}^3+{\mathrm{\Δ}}_{13}\\ {\mathrm{\Δ}}_{12}=C_1\\ {\mathrm{\Δ}}_{13}=C_2\end{array}\right.---\left(0.6\right)$

$\left\{\begin{array}{c}{{\stackrel{\‾}{\mathrm{\β}}}_{\⊥}}^1=\frac{{Z_c}^1}{f/\mathrm{dy}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{\⊥}}^2=\frac{{Z_c}^2}{f/\mathrm{dy}}\\ {{\stackrel{\‾}{\mathrm{\β}}}_{\⊥}}^3=\frac{{Z_c}^3}{f/\mathrm{dy}}\\ {Z_c}^1={Z_c}^2+{\mathrm{\Δ}}_{12}\\ {z_c}^1={z_c}^3+{\mathrm{\Δ}}_{13}\\ {\mathrm{\Δ}}_{12}=C_1\\ {\mathrm{\Δ}}_{13}=C_2\end{array}\right.---(0.7)$

Measure face with vernier caliper face face between distance, in conjunction with the measurement ratio calculating, solve overdetermined equation group (0.6) and can obtain the approximate solution of parameter f/dx, solve overdetermined equation group (0.7) and can obtain the approximate solution of parameter f/dy, the circular feature point P of known diameter ₀after imaging, be an ellipse, process and obtain oval major semi-axis image length, minor semi-axis image length and oval heart image coordinate, P by image ₀diameter compare β with the measurement that oval major semi-axis ratio is unique point place horizontal direction _- ⁰, according to formula (0.5) can obtain the sampling coordinate Z at flag sign point place _c ⁰.

(3) settling amount calculates:

According to full-size(d) and the picture size of " ten " font full-size(d) and picture size and sampling flag sign point, calculate sampling flag sign dot center to the vertical range between the selected reference point of marker peg, previous vertical range is compared with it, obtains the settling amount of this compacting operation;

In step (3), the implementation procedure that settling amount visible sensation method is measured is as follows: selected P ₁₁for reference point, utilize pixel coordinate and measure than calculating P ₀distance between point and reference point on vertical direction, i.e. P ₁₁point and P ₀point Yc coordinate difference, just can obtain the sedimentation value of measuring point after each compacting;

Set up following system of equations according to formula (0.1) and formula (0.2):

$\left\{\begin{array}{c}{v}^{11}=\frac{{Y_c}^{11}}{{Z_c}^1}\·f/\mathrm{dy}+v_0\\ v^{12}=\frac{{Y_c}^{12}}{{Z_c}^1}\·f/\mathrm{dy}+v_0\\ {Y_c}^{12}=f\left({Y_c}^{11}\right)\\ \·\·\·\·\·\·\end{array}\right.---\left(0.8\right)$

Solving equations (0.8) can obtain parameter Y _c ^1j(j=1,2 ... 12) and parameter v ₀, substitution formula (0.9) can be to parameter Y _c ⁰solve,

$v^0=\frac{{Y_c}^0}{{Z_c}^0}\·f/\mathrm{dy}+v_0---\left(0.9\right)$

If selected P ₁₁for reference point, unique point P ₀with P ₁₁at Y _con axle, coordinate difference is Delta1=Y _c ⁰-Y _c ¹¹, be the distance of in the vertical direction between measuring point and reference point.Roadbed after a compacting, will obtain a new distance value Delta2 again, and after twice compacting, subgrade settlement is Δ=Delta2-Delta1.

(4) Subgrade Compaction is calculated:

Settling amount substitution settling amount-compactness mathematical model that step (3) is obtained, can calculate the compactness that roadbed is new; In step (4), calculate compactness according to following formula:

$K^{\′}=K\·\frac{h}{h-\mathrm{\Δ}}---\left(0.10\right)$

In formula:

Laying thickness before h---compacting operation, cm;

Δ---the settling amount after a compacting, cm;

The compactness of laying before K---compacting operation;

K'---the compactness after a compacting;

Measure new settlement values, under laying thickness before known this compacting operation and compactness two Parameter Conditions, can calculate the new compactness of roadbed after this compacting operation according to settling amount-compactness mathematical modulo pattern (0.10), initial compactness is measured by traditional method.

A kind of Subgrade Compaction online test method based on machine vision of the present invention, it is a kind of Subgrade Compaction detection method completely newly, based on machine vision, online, its unique distinction is: taking image as detecting and transmission of information carrier, in obtaining compactness, view data is stored in to be calculated above, contributes to preserve construction information.Separate between detection method and compact technique, the compacting operation that this method can be applicable to different debulking methods or different compacting equipment detects.

Claims (1)

1. the Subgrade Compaction online test method based on machine vision, the steps include:

(1) set up vision measurement system:

The compactness of a compactness sampled point is measured, its measuring system is made up of 1 CCD camera, 1 sampling flag sign point and 1 marker peg, when multiple compactness sampled point sampling, corresponding increase sampling flag sign point and marker peg quantity, wherein, marker peg amounts to and divides three layers, and every layer is " ten " font, is designated as respectively face ₁, face ₂with face ₃, the angle point of these three faces is labeled as 1,2,3 successively ..., 12, wherein: face ₁, face ₂with face ₃set gradually, and face ₁, face ₂with face ₃be parallel to each other, face ₁perpendicular with construction road surface;

(2) image processing:

CCD collected by camera compacting image scene, then extracts the image coordinate that three layers of marker pegs amount to 36 points, and matching obtains three " ten " fonts of marker peg; Extract image coordinate and circular oval major semi-axis and the oval minor semi-axis Pixel Dimensions size parameter of sampling after the imaging of flag sign point of sampling flag sign dot center;

(3) settling amount calculates:

According to full-size(d) and the picture size of " ten " font full-size(d) and picture size and sampling flag sign point, calculate sampling flag sign dot center to the vertical range between the selected reference point of marker peg, previous vertical range is compared with it, obtains the settling amount of this compacting operation;

(4) Subgrade Compaction is calculated:

Settling amount substitution settling amount-compactness mathematical model that step (3) is obtained, can calculate the compactness that roadbed is new;

Wherein: the marker peg in step (1) is arranged near compactness sampled point, this compactness sampled point and marker peg are within the visual field of CCD camera, and wherein, marker peg is arranged in construction section and goes up along the line, face ₁, face ₂with face ₃be parallel to each other, face ₁perpendicular with construction road surface; Compactness sampling flag sign point is the mode by spray paint, at road surfaces compactness sample point spraying circle marker as unique point;

In the imaging model of step (2) image processing, establish O _c-X _cy _cz _cfor CCD camera coordinates system, O _cfor CCD camera photocentre, Z _caxle and CCD camera optical axis coincidence, by world coordinate system O _w-X _wy _wz _wsystem overlaps with CCD camera coordinates, and I is picture plane, and O-XY is that image physical coordinates is, o-uv is image pixel coordinate system, P _ijfor face _iupper j angle point, wherein: i=1,2,3, j=1,2 ..., 12, corresponding picture point P _ij' represent P ₀be artificial target's compactness sampling flag sign point, be the measuring point of settling amount, picture point is designated as P ₀'; Set following parameter or geometric condition:

A) P _ijat O _c-X _cy _cz _con (X for coordinate _c ^ij, Y _c ^ij, Z _c ⁱ) represent;

B) P _ij' (X for coordinate on image coordinate system O-XY ^ij, Y ^ij) represent;

C) P _ij' be (the u for coordinate on o-uv at pixel coordinate ^ij, v ^ij) represent;

D) focal length of camera represents with f;

E) pixel is respectively dx and dy at imaging plane X and Y-direction size;

F) coordinate of the initial point of coordinate system O-XY on image pixel coordinate system o-uv is (u ₀, v ₀);

If g) line segment P _ijp _ij+1in camera coordinates system, be vertical direction, P so _ij+1p _ij+2must, in horizontal direction, might as well suppose P _ijp _ij+1for the line segment of vertical direction in camera coordinates system, wherein: P _ijp _ij+1angle point P _ijwith angle point P _ij+1line, P _ij+1p _ij+2angle point P _ij+1with angle point P _ij+2line;

Get P _ij, P _ij+1and P _ij+2three points, in conjunction with imaging model, its imaging process is as follows:

According to pinhole imaging system principle, obtain following equation and set up:

Note point P _ijwith a P _ij+1distance between two points is l ₁, some P _ij' and put P _ij+1' two pel spacing is from being l ₁', have:

β is compared in the measurement of definition vertical direction _⊥:

β is compared in the measurement that can define equally horizontal direction _-:

Above-mentioned measurement is than being physical quantity for characterization of visual measuring system relation between given distance testee physical dimension and Pixel Dimensions;

Face ₁there are six to be parallel to X _cthe limit of axle, is horizontal sides, can obtain the measurement ratio of six horizontal directions after imaging, using the mean value of the measurement ratio of six horizontal directions as face ₁the measurement ratio of position horizontal direction, uses symbol represent, same, can obtain face ₂the measurement ratio of position horizontal direction face ₃the measurement ratio of position horizontal direction face ₁there are six to be parallel to Y _cthe limit of axle, is vertical edge, can obtain the measurement ratio of six vertical directions after imaging, by the method for averaging, calculates the that appears too ₁the measurement ratio of position vertical direction in like manner, obtain face ₂the measurement ratio of position vertical direction face ₃the measurement ratio of position vertical direction △ ₁₂and △ ₁₃be two process variable, represent respectively the distance between first cross and second cross, and distance between first cross and the 3rd cross; C ₁represent variable △ ₁₂value; C ₂represent variable △ ₁₃value, have:

Solve the approximate solution that overdetermined equation group (0.6) can obtain parameter f/dx, solve overdetermined equation group (0.7) and can obtain the approximate solution of parameter f/dy, circular feature point P ₀after imaging, be an ellipse, choose the measurement that is parallel to the diameter calculated characteristics point place horizontal direction of imaging plane on circle and compare β _- ⁰, according to formula (0.5) can obtain the sampling coordinate Z at flag sign point place _c ⁰;

In step (3), the implementation procedure that settling amount visible sensation method is measured is as follows: selected P ₁₁for reference point, utilize pixel coordinate and measure than calculating P ₀distance between point and reference point on vertical direction, i.e. P ₁₁point and P ₀point Y _ccoordinate difference, just can obtain the sedimentation value of measuring point after each compacting;

Set up following system of equations according to formula (0.1) and formula (0.2):

Solving equations (0.8) can obtain parameter Y _c ^1jwith parameter v ₀, wherein: j=1,2 ... 12, substitution formula (0.9) can be to parameter Y _c ⁰solve,

So, subsidence survey point, flag sign point P namely samples ₀with reference point P ₁₁at Y _con axle, coordinate difference is Δ=Y _c ⁰-Y _c ¹¹, be the distance of in the vertical direction between measuring point and reference point, roadbed again after a compacting, by obtain a new distance value Δ ', after twice compacting, subgrade settlement is Δ '-Δ;

In step (4), calculate compactness according to following formula:

In formula:

Laying thickness before h---compacting operation, cm;

△---the settling amount after a compacting, cm;

The compactness of laying before K---compacting operation;

K'---the compactness after a compacting;

When being roadbed, laying thickness controls parameter for aerial one, initial compactness is measured by traditional method, and the settling amount that initial compactness, laying thickness and step (3) are calculated is brought formula (0.10) into can calculate the new compactness of roadbed after a compacting operation.