CN104567708A - Tunnel full-section high-speed dynamic health detection device and method based on active panoramic vision - Google Patents

Abstract

The invention discloses a tunnel full-section high-speed dynamic health detection device based on an active panoramic vision. Hardware of the device comprises a tunnel detection trolley, an active panoramic vision sensor, an RFID reader and a processor. System software comprises an omnibearing face laser information analysis and point cloud data obtaining unit, a tunnel center axis extracting unit, a noise reducing and adjustment processing unit, a three-dimensional modeling and deformation analyzing unit, a full tunnel all-section cross section benchmark database, a full tunnel longitudinal section splicing unit, a displacement monitoring and sedimentation monitoring unit, a full tunnel health examination result database and a tunnel longitudinal section variable shift amount three-dimensional visualization unit. Transverse and longitudinal deformation of a tunnel are analyzed and recognized by carrying out machine vision processing on laser scanning fracture surface slice images of the inner wall of the tunnel. The invention further provides a tunnel full-section high-speed dynamic health detection method based on the active panoramic vision, and effective technology support is provided for daily maintenance of tunnels of metros and high-speed rails.

Description

Based on full section of tunnel high speed dynamical health pick-up unit and the method for active panoramic vision

Technical field

The present invention relates to the application in the automatic detection and three-dimensional modeling in tunnel of panorama LASER Light Source, omnibearing vision sensor, Digital Image Processing, Three Dimensional Reconfiguration and computer vision technique, particularly relate to a kind of full section of tunnel high speed dynamical health pick-up unit based on active panoramic vision and method, be mainly used in the automatic health monitoring during subway and the operation of high ferro tunnel.

Background technology

The health monitoring in subway and high ferro tunnel can be divided into two stages: construction stage and operation stage.In the operation process of railway tunnel, the safety problem in tunnel is mainly subject to the impact of the following aspects: one is the railway roadbed bulk settling that the medium-term and long-term oscillatory load of train operation process causes; Two is because most of major long tunnel is in complicated geologic condition, the circuit that train load causes non-uniform settling in the axial direction; It is exactly finally the impact of tunnel perimeter buildings.And these impacts can cause the cracking of tunnel contour, distortion, a series of safety problem such as even to come off of leaking.

At present, both at home and abroad for the own comparative maturity of monitoring technology in constructing tunnel stage, but the monitoring attention degree during tunnel operation is nowhere near.In fact, operation stage, because time span is large, influence factor is complicated, disaster social influence large, more should be paid attention to fully to the health monitoring in subway and high ferro tunnel.The detection of tunnel contour can easily grasp tunnel clearance, the distortion in monitoring tunnel, and therefore tunnel structure detects and occupies an important position in railway transport course.

In the O&M stage in tunnel, in order to not affect circuit operation, can only utilize the blank time, and the blank time is certain, this just needs to detect rapidly, and the mode of traditional some standing posture Measure section distortion obviously can not meet the demands.According to the demand that high ferro and city underground develop, the monitoring of high speed dynamical health is carried out to full section of tunnel, and require that observation process can be synchronous with train operation, be convenient to managerial personnel and grasp tunnel change in real time, continuously, rapidly, the hidden danger of transportation safety of eliminating the effects of the act in time.Require that section detects the collection of data, analysis, process, transmission in the process, and provided the three-dimensional model in corresponding profiled outline and tunnel fast, so Real-Time Monitoring and analyze various potential safety hazard.

Tunnel testing has gradually changeable and the chronicity of tunnel deformation, the feature such as measuring point is many, circuit is long, amount of dynamic data is large, data analysis is complicated.Contactless vehicle-mounted tunnel cross-section detects and has become main flow detection technique.At present, contactless vehicle-mounted tunnel cross-section detection scheme mainly have employed laser technology and computer image processing technology etc.And be widely used on track detection vehicle advanced both at home and abroad, detection efficiency is greatly improved.More typically have: the infrastructure high-speed detection train MGV in order to improve high-speed line maintenance that France comes into operation for 2005, use photogrammetric mode can detect contact net, track etc. under the speed of 300km; The GeoRail-Xpress synthetic detection vehicle of Germany can carry out total digitalization and measure, gathers and analyze to the visible and invisible part on circuit; It is equipped with the circuit and environmental detection set that are made up of 4 dimension orbital environment cameras, the track detection device be made up of 6 laser sensors and the digital line scan camera of 2 framves, detection speed is 100km/h; Japan's track synthetic detection vehicle is representative with the East-I type most come into operation in March, 2002, this train adopts 700 to be electric EMU, 6 marshallings, 88 test items can be implemented, the camera head wherein applying train top adopts image processing techniques to measure the situation of railway roadbed and peripheral structure thing, highest running speed 275km/h; Austrian Plasser company speed per hour is the EM250 type high speed track checking car of 250km/h, and Italy's " Archimedes " number synthetic detection vehicle etc. is all equipped with the velocity tunnel checkout equipment using photogrammetric technology or laser assisted image processing techniques.

In sum, the main relative merits detected at vehicle-mounted tunnel cross-section about laser technology and computer image processing technology are: laser scanning measurement does not disturb by extraneous light, but due to the restriction by scanning spot effect weakening, precision when measuring compound section is low, problem is that scanning distance measuring method is advanced by restriction scanning spot in traveling process of mechanical gyro unit twist, sweep frequency does not allow the get Tai Gao that is easy to do, and measured some position is not in an xsect in high speed traveling process, this belongs to original reason error, measuring accuracy must be improved by repeatedly back and forth measuring, be unfavorable for the measurement to long distance tunnel.Simultaneously because repeatedly back and forth kinetic measurement is high to positioning accuracy request, implement also comparatively difficulty.Therefore, some researcher proposes and adopts 30 °, interval installation 9 laser scanners to scan tunnel cross-section simultaneously, and this also improves input and O&M cost undoubtedly, and such scanning is also only only limitted to several point, cannot realize full face real time scan.

Although computer picture triangle is photogrammetric have higher dynamic measurement precision, is subject in sunlight, tunnel the extraneous light interference such as light when measuring and produces deviation, even losing efficacy time serious.Therefore complicated image processing algorithm must be adopted to carry out making up to a certain degree, and such as, by two-dimensional filtering and mode identification technology, filtering light image, identifies the laser the position of optic strip etc. on Tunnel wall.And common kinetic measurement requires the sampling rate of the even thousands of frame of hundreds of per second, current technology is difficult to the real-time storage realizing all images data, if by image processing techniques pick up and store laser the position of optic strip, then image processing step can increase the processing time again greatly, system for restricting sample rate, the real-time of influential system, is not suitable for the safety monitoring in high speed traveling process.

Therefore, under high speed kinetic measurement condition, how to ensure that high measurement accuracy, high sampling rate and quick three-dimensional modeling are that subway and high ferro tunnel automatic monitoring survey important topic simultaneously.

Chinese invention patent application number is 87101789 disclose one laser measurement buildings or cave profile section, it have employed tuning fork slit oscillator, in kinematic train, adopts three grades of optical disc step-by-step countings and reset circuit in angle-measuring equipment, measuring accuracy reaches ± and 2.5%, measuring pole coordinate parameter, maximum clear height and the numerical value such as clear span, section net area that can obtain section through MICROCOMPUTER PROCESSING immediately simultaneously.The main Problems existing of this technology is not suitable for the monitoring in high speed traveling process.

Chinese invention patent application number discloses a kind of tunnel safety tool car and detection method for 201410121950.X, and examination and repair system comprises switch board, display screen, front distance sensor, rear range sensor, swivel bearing mechanism and pick-up unit; Information acquisition system comprises Universal rotary support, 3D holographic scanners and information storage module.This invention claims automatically and fast to detect the region that the problem such as crack or surface peeling appears in inner surface of tunnel, while achieving rapid-maintenance, zonal 3D information acquisition is carried out to tunnel, for the later stage provides foundation, thus promote the efficiency of maintenance and the effect of information acquisition and maintenance.But this technology to analyze point ten minutes is limited, does not carry out any process, still need manual observation holoscan image to identify to 3D holoscan result.

Chinese invention patent application number is 201410009353.8 disclose a kind of intelligent inspection device of railway tunnel and using method thereof, and inspection device comprises and controls mainboard, cabinet, secondary light source, digital camera and alarm; Control, in mainboard, special software is installed; The digital photograph of special software logarithmic code camera shooting carries out analyzing, process and preserving; When result exceeds setting range, alarm work; But the image that there is no in this patent captured by logarithmic code camera carries out the content of Automatic analysis and carries out three-dimensional measurement to full section of tunnel.

Tunnel health monitoring comprises tunnel structure and corrodes monitoring, structural deformation monitoring, structural internal force measure and ambient conditions monitoring, wherein especially structural deformation monitoring is extremely important, and its Contents for Monitoring is mainly the convergent deformation of the Longitudinal Settlement (longitudinal axis distortion) in tunnel, transversal displacement and section.

Several performance requirements of full section of tunnel high speed dynamical health detection method: the 1. requirement of image acquisition speed; System is installed on Tunnel testing car, and under the detection speed of 120km/h, frequency acquisition will meet at least every 0.1m and gather a comprehensive section; 2. the requirement of measuring accuracy; 5mm must be less than to the static measurement error in the space of section; 3. the requirement of reliability; The necessary stability of selected hardware device is high, and system must can meet the requirement that long-time continuous stable is run; 4. the requirement of robotization Gernral Check-up assessment; The deformation that automatic analysis tunnel occurs, sets up tunnel deformation Three-Dimensional Dynamic model, and forms corresponding early warning mechanism; 5. operation and maintenance simplifies requirement; Carry out the monitoring of high speed dynamical health to full section of tunnel, the collection of data, analysis, process, transmission are and automatically complete, and do not need artificial intervention.

Summary of the invention

In order to the robotization and intelligent level that overcome existing tunnel health detecting method are low, be difficult to realization, the deficiencies such as high-acruracy survey, high sampling rate and quick three-dimensional modeling are carried out to tunnel, the invention provides a kind of for carrying out the monitoring of high speed dynamical health to full section of tunnel, tunnel health detection robotization and intelligent level can be improved, high-precision automatic analysis and assessment are carried out to the malformation in tunnel, realizes the modeling of tunnel deformation Three-Dimensional Dynamic.

Realize foregoing invention content, five key problems must be solved: (1) realizes a kind of panorama LASER Light Source; (2) the active panoramic vision sensor that a kind of energy quick high accuracy obtains actual object depth information is realized; (3) longitudinal mileage coordinate of orbital direction is accurately estimated; (4) adopt computer vision technique to the three-dimensional rebuilding method in tunnel; (5) set up spatial data library storage and issue tunnel deformation data. set up tunnel deformation Three-Dimensional Dynamic model, and form early warning mechanism.

The technical solution adopted for the present invention to solve the technical problems is:

Based on full section of tunnel high speed dynamical health pick-up unit and the method for active panoramic vision, its hardware comprises: Tunnel testing car, active panoramic vision sensor, RFID reader, wireless interface transmitting element, controller, level of standing communication system or central monitoring center server; Described central monitoring center server forms full tunnel safety checking network by described station level communication system, the tunnel cross section laser scanning image that described station level communication system reception sends over from the wireless interface transmitting element that described Tunnel testing car configures, and send tunnel cross section laser scanning image to described central monitoring center server by full tunnel safety checking network in time.

Described Tunnel testing car is configured with active panoramic vision sensor, RFID reader, measuring wheel, wireless interface transmitting element and controller, described active panoramic vision sensor is arranged on the central front of described Tunnel testing car, described RFID reader reads and is configured in the RFID information that tunnel inner wall is settled, a measuring wheel is installed in the bottom of described Tunnel testing car, and described controller reads the travel distance Z of the pulse equivalency of photoelectric encoder in measuring wheel the Tunnel testing car described in estimation _i; Described controller reads tunnel cross section laser scanning image that active panoramic vision sensor obtains and with the travel distance Z of described Tunnel testing car _ibe that filename is kept in the storage unit of described controller with present moment; When described Tunnel testing car reaches next website, the tunnel cross section laser scanning image in the storage unit of described controller is sent to described station level communication system by described wireless interface transmitting element by described controller.

Described active panoramic vision sensor, its hardware mainly comprises: omnibearing vision sensor, panorama laser projection light source; Described omnibearing vision sensor is coaxially fixedly connected with described panorama laser projection light source.

Described omnibearing vision sensor comprises hyperboloid minute surface, upper cover, transparent semicircle outer cover, lower fixed seat, image unit holder, image unit, linkage unit and upper cover; Described hyperboloid minute surface be fixed on described on cover, described lower fixed seat and transparent semicircle outer cover link into an integrated entity by described linkage unit, described transparent semicircle outer cover and described upper cover and described upper cover are fixed together, described image unit is fixed on described image unit holder, described image unit holder is fixed on described lower fixed seat, and the output of the described image unit in described omnibearing vision sensor is connected with described controller by kilomega network data-interface.

Described panorama laser projection light source comprises light source upper cover, conical minute surface, transparent housing, ring shape generating laser and base; Ring shape generating laser is fixed on base, the utilizing emitted light axial line of ring shape generating laser is consistent with base axial line, conical minute surface is fixed on light source upper cover for reflection circle cast laser transmitter projects circle laser out for tunnel inner wall provides tunnel to break cross section panoramic scanning light, the axial line of conical minute surface is consistent with light source upper cover axial line, and the base securing ring shape generating laser is integrated into panorama laser projection light source with the light source upper cover securing conical minute surface by transparent housing; The central shaft of ring shape generating laser and the central shaft of conical minute surface overlap.

When described active panoramic vision sensor assembles, by the central shaft arrangement of the central shaft of the central shaft of described ring shape generating laser, described conical minute surface, the central shaft of described hyperboloid minute surface and described image unit on same axial line.

Before Tunnel testing car is about to enter tunnel, positioning system can provide the positional information of Tunnel testing car, when Tunnel testing truck position and tunnel portal mileage are within the scope of 10 meters, controller starts measuring system, active panoramic vision sensor enters Information Monitoring state, simultaneity factor clock system starts, the temporal information that record each position information is corresponding.Keep at a certain distance away at tunnel mouth, Nei Hechu hole, hole installation electronic tag, utilizes RFID to assist and revise the accurate location that mileage has located section.

When Tunnel testing car travels in a detector segments, except needing to obtain the X of tunnel cross-section relative to measuring system, outside Y-coordinate, also should obtain longitudinal mileage coordinate Z of direction along ng a path, be used for determining the position of each Measure section in tunnel, namely this section is apart from the accurate distance measuring reference position.Need to detect the distance travelled in real time of Tunnel testing car for this reason.

Bottom Tunnel testing car, install measuring wheel, during advance, measuring wheel does pure rolling in orbital plane, and the wheel shaft of steamboat is equipped with photoelectric encoder, can read according to scrambler the distance that steamboat passes by rail level.

First described central monitoring center server reads the travel distance Z of described Tunnel testing car after receiving the tunnel cross section laser scanning image of Tunnel testing car by full tunnel safety checking network _iwith the tunnel cross section laser scanning image that present moment is filename; Then from tunnel cross section laser scanning image, the spatial coordinate location value of panorama laser projection point on tunnel cross section is parsed, the three dimensional point cloud namely on tunnel cross section; Then with the travel distance Z of Tunnel testing car _itunnel cross section on three dimensional point cloud reconstruct full tunnel inner wall three-dimensional model; Finally contrast with the full tunnel inner wall three-dimensional model built at first according to the full tunnel inner wall three-dimensional model of up-to-date structure, analyze tunnel-type variable.

The acquisition of tunnel cross-section cloud data, the acquisition mode of tunneling data is centered by the viewpoint of active panoramic vision sensor, parses the travel distance Z of Tunnel testing car _itunnel cross section on face, tunnel on (x, y) planar two dimensional coordinate of impact point, then according to travel distance Z _ito calculate on face, tunnel have (x, y, z) three-dimensional coordinate of a cloud.

Omnibearing vision sensor demarcating module, for determining the parameter of mapping relations between the X-Y scheme picture point in three dimensions point and video camera imaging plane, calibrated parameter leaves in described storage unit.

Comprehensive laser intelligence is resolved and cloud data acquiring unit, for processing laser scanning panoramic image data, laser scanning panoramic picture parse laser projection information and calculates spatial positional information, finally obtaining tunnel inner wall edge cloud data.

Tunnel axis extraction unit, for drawing tunnel linear deformation figure, to carry out three-dimensional modeling according to tunnel axis; Here will along tunnel axis direction continuous print 10 scanning cross-section as a joint, in laser scanning above, each adjacent scanning cross-section is spaced apart 100mm, so just the cloud data unit of 1000mm in a joint is extracted axis, the method extracting axis be to segmentation after tunnel point cloud carry out face of cylinder matching.

Noise reduction and adjustment processing unit, the length for obtaining described tunnel axis extraction module process is further that 1000mm tunnel cloud data carries out noise reduction and adjustment processing.

Three-dimensional modeling and deformation analysis unit, for tunnel deformation visualization processing; Mainly through comparing the deformation quantity in tunnel in same position, to reflect the convergent deformation situation in local, tunnel or a certain section.

Full tunnel profile concatenation unit, for splicing the segmented three-dimensional reconstruction result in full tunnel.

Displacement monitoring and settlement monitoring unit, have the tolerance of quantification to the global displacement in tunnel and sedimentation.

Tunnel profile becomes the amount of moving three-dimensional visualization unit, for having quantification, visual expression to the global displacement in tunnel and sedimentation; Three-dimensional visualization process measure the numerical value, the deflection information that obtain become intuitively, represent with Figure and Image, present to tunnel safety monitoring personnel with the physical quantity of spatial variations in time; Here represent positive deformation by redness, represent negative deformation by blueness, color is more deeply felt and is shown that deformation is larger; To RGB color space, GB color component is set to zero, R component aligns thermomechanical processing from 0 ~ 255 and maps, corresponding to the positive deformation of 0 ~ 255mm; RG color component is set to zero, G component map from 0 ~ 255 negative thermomechanical processing, corresponding to the negative deformation of 0 ~-255mm; Then color-coded on the joint corresponding to tunnel to corresponding to the vertical section deformation quantity of each joint on full tunnel central axis; Broken line mode of saving each on central axis couples together by the coordinate figure on last basis full tunnel central axis, shows the deformation situation of whole tunnel profile by visual means.

Beneficial effect of the present invention is mainly manifested in:

1) a kind of brand-new robotization tunnel health examination mode is provided;

2) in the process doing health check-up to subterranean tunnel, gather the three-dimensional spatial information in tunnel in time, descend basic spatial database primitively for Urban underground Tunnel three-dimensional modeling provides;

3) the various defects judging existence in tunnel are detected automatically, for the maintenance of subterranean tunnel, final acceptance of construction provide effective technical support.

Accompanying drawing explanation

Fig. 1 is a kind of structural drawing of omnibearing vision sensor;

Fig. 2 is single view catadioptric omnibearing vision sensor imaging model, Fig. 2 (a) perspective imaging process, Fig. 2 (b) sensor plane, Fig. 2 (c) plane of delineation;

Fig. 3 is that initiatively panoramic vision sensor carries out the schematic diagram of tunnel inner wall range observation;

Fig. 4 is the structural drawing of panorama laser projection light source;

Fig. 5 is a kind of structural drawing of active panoramic vision sensor;

Fig. 6 adopts initiatively panoramic vision sensor tunnel inner wall to be carried out to the schematic diagram of laser scanning inspection;

Fig. 7 is a kind of overall macroscopical schematic diagram adopting initiatively panoramic vision sensor subterranean tunnel to detect;

Fig. 8 is that one in tunnel inner wall saves in some position angles situation of omnibearing vision sensor, the distribution relation figure of radial-deformation and cloud data;

Fig. 9 is tunnel longitudinal deformation schematic diagram;

Figure 10 is tunnel lateral direction deformation schematic diagram;

Figure 11 is the tunnel three-dimensional plot with the reconstruct of some cloud;

Figure 12 is the processing flow chart that Tunnel testing car carries out tunnel health examination;

Figure 13 is the spliced one section of tunnel figure of Surface Reconstruction from Data Cloud gathered from tunnel;

Figure 14 is the several important geometry variable schematic diagram on tunnel inner wall surface;

Figure 15 is tunnel cross sectional profile diagram;

Figure 16 is deformation processing flow chart in analyzing and processing tunnel in central monitoring center server.

Embodiment

Embodiment 1

With reference to Fig. 1 ~ 16, based on full section of tunnel high speed dynamical health pick-up unit and the method for active panoramic vision, its hardware comprises: Tunnel testing car, active panoramic vision sensor, RFID reader, wireless interface transmitting element, controller, level of standing communication system or central monitoring center server.Central monitoring center server forms full tunnel safety checking network by station level communication system, the tunnel cross section laser scanning image that the reception of level of standing communication system sends over from the wireless interface transmitting element that Tunnel testing car configures, and send tunnel cross section laser scanning image to central monitoring center server by full tunnel safety checking network in time.

Tunnel testing car is configured with active panoramic vision sensor, RFID reader, measuring wheel, wireless interface transmitting element and controller, active panoramic vision sensor is arranged on the central front of Tunnel testing car, RFID reader reads and is configured in the RFID information that tunnel inner wall is settled, a measuring wheel is installed in the bottom of Tunnel testing car, and controller reads the pulse equivalency of photoelectric encoder in measuring wheel and estimates the travel distance Z of Tunnel testing car _i; Controller reads tunnel cross section laser scanning image that active panoramic vision sensor obtains and with the travel distance Z of Tunnel testing car _ibe that filename is kept in the storage unit of controller with present moment; When Tunnel testing car reaches next website, the tunnel cross section laser scanning image in the storage unit of controller is sent to station level communication system by wireless interface transmitting element by controller.

Active panoramic vision sensor, its hardware mainly comprises: omnibearing vision sensor, panorama laser projection light source; Omnibearing vision sensor is coaxially fixedly connected with panorama laser projection light source.

Omnibearing vision sensor, as shown in Figure 1, comprises hyperboloid minute surface 2, upper cover 1, transparent semicircle outer cover 3, lower fixed seat 4, image unit holder 5, image unit 6, linkage unit 7 and upper cover 8.Hyperboloid minute surface 2 is fixed on upper cover 1, lower fixed seat 4 and transparent semicircle outer cover 3 link into an integrated entity by linkage unit 7, together with transparent semicircle outer cover 3 is fixed by screws in upper cover 1 and upper cover 8, image unit 6 is screwed on image unit holder 5, image unit holder 5 is screwed on lower fixed seat 4, and the output of the image unit in omnibearing vision sensor is connected with controller by kilomega network data-interface.

The sample frequency of image unit needs under the detection speed of 120km/h, and frequency acquisition will meet at least every 0.1m and gather a tunnel cross section, calculates sample frequency and meets and be greater than 333.3fps condition.The sampling resolution of image unit, according to the requirement of measuring accuracy, must be less than 5mm to the static measurement error in the space of section; The inconocenter that is detected as of omnibearing vision sensor is 3m from the longest distance of tunnel edge, corresponding to the half scope of minor axis in the imager chip of video camera, if do not consider, interpolation puies forward high-resolution words, and the sampling resolution of image unit needs more than 1200 pixels.Comprehensive above-mentioned situation, image unit selects CR3000 × 2 high-speed camera, and resolution is 1696 × 1710, sample frequency 540fps, high-speed internal memory 16GB.

Panorama laser projection light source comprises light source upper cover, conical minute surface, transparent housing, ring shape generating laser and base.Ring shape generating laser is fixed on base, the utilizing emitted light axial line of ring shape generating laser is consistent with base axial line, conical minute surface is fixed on light source upper cover for reflection circle cast laser transmitter projects circle laser out for tunnel inner wall provides tunnel to break cross section panoramic scanning light, the axial line of conical minute surface is consistent with light source upper cover axial line, the base securing ring shape generating laser is integrated into panorama laser projection light source with the light source upper cover securing conical minute surface by transparent housing, the central shaft of ring shape generating laser and the central shaft of conical minute surface overlap.During active panoramic vision sensor assembling, by the central shaft of the central shaft of ring shape generating laser, conical minute surface, the central shaft of hyperboloid minute surface and the central shaft arrangement of image unit on same axial line.

Before Tunnel testing car is about to enter tunnel, positioning system can provide the positional information of Tunnel testing car, when Tunnel testing truck position and tunnel portal mileage are within the scope of 10 meters, controller starts measuring system, active panoramic vision sensor enters Information Monitoring state, simultaneity factor clock system starts, the temporal information that record each position information is corresponding.Keep at a certain distance away at tunnel mouth, Nei Hechu hole, hole installation electronic tag, i.e. RFID, utilizes RFID to assist and revise the accurate location that mileage has located section.

On RFID is fixed on tunnel mouth, Nei Hechu hole, hole keeps at a certain distance away, here using the tunnel cross-section at fixing RFID place and the intersection point of axis, tunnel as the reference point measured, for uniform coordinate benchmark B is set up in tunnel _i(x, y, z); The spatial positional information B of tunnel point of fixity is stored in the storage unit of RFID _i(x, y, z); The spatial positional information B of tunnel point of fixity _i(x, y, z) obtains through fixed point high-acruracy survey after tunnel builds up; The spatial positional information B of tunnel point of fixity _i(x, y, z) needs fixed cycle to carry out safeguarding and correcting in tunnel operation process, to ensure these spatial positional informations B _i(x, y, z) can as the absolute coordinates benchmark in tunnel.

Controller comprises: RFID data reading unit, for reading the spatial positional information being fixed on RFID in tunnel wall and storing; Running distance evaluation unit, by reading the umber of pulse of photoelectric encoder and utilizing formula (3) to estimate the running distance of Tunnel testing car; Tunnel cross section laser scanning image reads, storage unit, for reading the tunnel cross section laser scanning image that active panoramic vision sensor obtains, and with the travel distance Z of Tunnel testing car _ibe that filename is kept in the storage unit of controller with present moment; Tunnel cross section laser scanning image data transmission unit, for sending to station level communication system by the tunnel cross section laser scanning image in the storage unit of controller; Accompanying drawing 12 carries out the processing flow chart of tunnel health examination for controller.

When Tunnel testing car travels in a detector segments, except needing to obtain the X of tunnel cross-section relative to measuring system, outside Y-coordinate, also should obtain longitudinal mileage coordinate Z of direction along ng a path, be used for determining the position of each Measure section in tunnel, namely this section is apart from the accurate distance measuring reference position, needs to detect the distance travelled in real time of Tunnel testing car for this reason.

The Kinematic Positioning of Tunnel testing car mainly relies on photoelectric encoder, installs a measuring wheel in the bottom of Tunnel testing car, in conjunction with track-circuit signalling determination Tunnel testing car reference position and eliminate longitudinal cumulative errors.Photoelectric encoder can export 1000 ~ 2000 pulses/turn, the distance of travelled by vehicle can be calculated according to the number of pulses collected and measuring wheel wheel diameter, the many algorithms such as FT method, slide system skidding algorithm, Multi-sensor Fusion algorithm in practical application, can be adopted to improve the precision of location.Meanwhile, mileage positioning system can also calibrate the cumulative errors in Tunnel testing car initial position, targeted elimination kinetic measurement process according to track-circuit signalling.

Bottom Tunnel testing car, install measuring wheel, during advance, measuring wheel does pure rolling in orbital plane, and the wheel shaft of steamboat is equipped with photoelectric encoder, can read according to scrambler the distance that steamboat passes by rail level.If the diameter of steamboat is D, P is elected in the graduation of dress photoelectric encoder thereon as, is calculated pulse equivalency (every individual pulse is equivalent to the air line distance that the steamboat is passed by) δ of scrambler by formula (1),

$\mathrm{\δ}=\frac{\mathrm{\πD}}{P}---\left(1\right)$

The diameter of measuring wheel is Φ 58, and the photoelectric encoder on measuring wheel divides and divides 2000 into, and the pulse equivalency formula (2) of scrambler calculates,

$\mathrm{\δ}=\frac{58\mathrm{\π}}{2000}=0.0911\mathrm{mm}---\left(2\right)$

The travel distance Z of Tunnel testing car _isend Z pulse with photoelectric encoder to calculate, computing method as shown in formula (3),

Z _i＝Zδ＝0.0911Z (3)

In order to the tunnel cross section laser scanning image allowing the active panoramic vision sensor be arranged on Tunnel testing car obtain associates with locus during captured image, adopt the travel distance Z with Tunnel testing car here _ifor filename preserves tunnel cross section laser scanning image data; Wirelessly tunnel cross section laser scanning image data are sent to central monitoring center server through station level communication system when station level in Tunnel testing car process of passing through tunnel.

The tunnel health detection flow process of Tunnel testing car as shown in Figure 12, before Tunnel testing car is about to enter tunnel, controller on Tunnel testing car reads the spatial positional information being fixed on RFID in tunnel wall and storing, controller starts measuring system, active panoramic vision sensor enters Information Monitoring state, simultaneity factor clock system starts, and the temporal information that record each position information is corresponding also calibrates the initial position of Tunnel testing car; Controller reads the pulse equivalency of photoelectric encoder in measuring wheel and estimates the travel distance Z of Tunnel testing car _i; Controller reads tunnel cross section laser scanning image that active panoramic vision sensor obtains and with the travel distance Z of Tunnel testing car _ibe that filename is kept in the storage unit of controller with present moment; When Tunnel testing car reaches next website, the tunnel cross section laser scanning image in the storage unit of controller is sent to station level communication system by wireless interface transmitting element by controller; Along with Tunnel testing car moves ahead with the speed of 120km/h, controller constantly reads tunnel cross section laser scanning image that active panoramic vision sensor obtains and reads the pulse equivalency of photoelectric encoder in measuring wheel and estimate the travel distance Z of Tunnel testing car _i, and with the travel distance Z of Tunnel testing car _ibe that filename is kept in the storage unit of controller with present moment, until be tunnel exit when the controller on Tunnel testing car reads the information be fixed in tunnel wall in RFID; Now, controller stops obtaining view data to active panoramic vision sensor, close the panorama laser projection light source in active panoramic vision sensor, and the tunnel cross section laser scanning image be fixed at tunnel exit place in the spatial positional information and storage unit that RFID in tunnel wall stores sends to station level communication system; Such Tunnel testing car is to the health examination end of scan in tunnel, and the transversal section laser scanning image obtained in the health examination scanning process in tunnel sends in the full tunnel cross section laser scanning image storehouse in central monitoring center server through station level communication system.

After central monitoring center server receives the tunnel cross section laser scanning image of Tunnel testing car by full tunnel safety checking network, first read the travel distance Z of Tunnel testing car _iwith the tunnel cross section laser scanning image that present moment is filename; Then from tunnel cross section laser scanning image, the spatial coordinate location value of panorama laser projection point on tunnel cross section is parsed, the three dimensional point cloud namely on tunnel cross section; Then with the travel distance Z of Tunnel testing car _itunnel cross section on three dimensional point cloud reconstruct full tunnel inner wall three-dimensional model; Finally contrast with the full tunnel inner wall three-dimensional model built at first according to the full tunnel inner wall three-dimensional model of up-to-date structure, analyze tunnel-type variable.

Mainly comprise in central monitoring center server: omnibearing vision sensor demarcates unit, active panoramic vision sensor nominal data storehouse, J saves tunnel cross-section view data reading unit, full tunnel cross section laser scanning image storehouse, comprehensive laser intelligence is resolved and cloud data acquiring unit, tunnel axis detection unit, noise reduction and adjustment processing unit, three-dimensional modeling and deformation analysis unit, Ge Jie transversal section, full tunnel benchmark database, full tunnel profile concatenation unit, displacement monitoring and settlement monitoring unit, coordinate reference data storehouse, Ge Jie axis, full tunnel, full tunnel health examination result database and tunnel profile become the amount of moving three-dimensional visualization unit, treatment scheme as shown in Figure 16.

Omnibearing vision sensor demarcates unit, for determining the parameter of mapping relations between the X-Y scheme picture point in three dimensions point and video camera imaging plane, have employed the omnibearing vision sensor of single view in the present invention, the omnibearing vision sensor be made up of hyperboloid catadioptric mirror image principle has single view imaging characteristic; Its image-forming principle as shown in Figure 3.In order to set up the mapping relations in three dimensions point and imaging plane picture point, here the perspective projection imaging model of Micus í k is adopted, as shown in Figure 2, in this imaging model, consider two different reference planes, the plane of delineation (u', v') and sensor plane (u "; v "), the plane of delineation is relevant with the CCD of video camera, represents with pixel coordinate system.Sensor plane is the plane orthogonal with minute surface optical axis of a hypothesis, and its center origin is the intersection point of optical axis and this plane; With the focus of hyperboloid minute surface, i.e. single view O _mfor initial point sets up coordinate system, z " axle and minute surface optical axis alignment; If X=[X, Y, Z] ^tfor in space a bit, u "=[u ", v "] ^tthe projection of X in sensor plane, u'=[u', v'] ^tit is the pixel of the plane of delineation of its correspondence; Volume coordinate point X first projects A point place on minute surface by projective transform matrix, and A point focuses on camera optics central point C by mirror-reflection, and hands over u in sensor plane "=[u ", v "] ^tpoint, u " puts by affined transformation to the plane of delineation being put u'=[u', v'] ^t; What whole single view catadioptric camera imaging model described is by spatial point to catadioptric mirror point, and catadioptric mirror point is to the point on imaging plane, and the point on imaging plane forms the process of the pixel in image again to plane of delineation point.

Conversion between catadioptric minute surface to sensor plane formula (17) represents;

$\mathrm{\λ}{p}^{\′\′}=\mathrm{\λ}\left(\begin{array}{c}{X}^{\′\′T}\\ Z^{\′\′}\end{array}\right)=\mathrm{\λ}\left(\begin{array}{c}h\left(\right|\left|u^{\′\′}\right|\left|\right)u^{\′\′}\\ g\left|\right|u^{\′\′}\left|\right|\end{array}\right)=P\·X,\mathrm{\λ}>0---\left(17\right)$

In formula, X ∈ R ⁴the secondary coordinate of representation space point X, P=[R|T] ∈ R ^{3 × 4}for projective transform matrix, R ∈ R ^{3 × 3}for spatial point is to the rotation matrix of catadioptric mirror point, T ∈ R ^{3 × 1}for spatial point is to the translation matrix of catadioptric mirror point.

Represented by the formula (18) of the conversion sensor plane to the plane of delineation:

u″＝Au′+t (18)

In formula, A ∈ R ^{2 × 2}, t ∈ R ^{2 × 1}.

Scaramuzza is on the basis of Micusik perspective projection model, replace the function g in formula (17) with a function f=g/h, h, namely characterize the relation between three dimensions point and two dimensional surface point by function f, obtain formula (19)

$\mathrm{\λ}{p}^{\′\′}=\mathrm{\λ}\left[\begin{array}{c}{u}^{\′\′}\\ f\left(\right|\left|u^{\′\′}\right|\left|\right)\end{array}\right]=\mathrm{\λ}\left[\begin{array}{c}{\mathrm{Au}}^{\′}+t\\ f\left(\right||{\mathrm{Au}}^{\′}+t|\left|\right)\end{array}\right]=P\·X,\mathrm{\λ}>0---\left(19\right)$

Due to bi-curved rotational symmetry, Scaramuzza Taylor launches polynomial expression and carrys out described function f, represents with formula (20):

f(||u″||)＝a ₀+a ₁||u″||+a ₂||u″|| ²+...+a _n||u″|| ^N(20)

In formula, || u " || for the point on imaging plane is to the distance of this planar central point.

The prerequisite of the model of Scaramuzza and Micusik is all desirable catadioptric camera model, owing to can introduce some errors when reality processing assembling omnibearing vision sensor; Here suppose that the omnibearing vision sensor demarcated meets ideal model, the non-ideal model that there is certain error is substituted into the simplified model conversion formula that Scaramuzza proposes, obtain formula (21);

$\mathrm{\λ}{p}^{\′\′}=\mathrm{\λ}\left[\begin{array}{c}{u}^{\′\′}\\ f\left(\right|\left|u^{\′\′}\right|\left|\right)\end{array}\right]=\mathrm{\λ}\left[\begin{array}{c}{\mathrm{Au}}^{\′}+t\\ f\left(\right||{\mathrm{Au}}^{\′}+t|\left|\right)\end{array}\right]=P\·R\·X,\mathrm{\λ}>0---\left(21\right)$

Concrete calibration process is around omnibearing vision sensor one week by scaling board, take some groups of panoramic pictures, set up some equatioies of pixel in spatial point and imaging plane, optimization algorithm is used to obtain optimum solution, result of calculation is as shown in table 1, is the calibrating parameters of the omnibearing vision sensor used in the present invention;

The calibration result of table 1 ODVS

After calibrating the inside and outside parameter of omnibearing vision sensor, just can set up picture point and the incident ray of an imaging plane, the corresponding relation namely between incident angle, as formula (5) represents;

${\mathrm{tam\α}}_{\mathrm{\β}}=\frac{\left|\right|u^{\′\′}\left|\right|}{f\left(\right|\left|u^{\′\′}\right|\left|\right)}=\frac{\left|\right|u^{\′\′}\left|\right|}{a_0+a_1\left|\right|u^{\′\′}\left|\right|+a_2{\left|\right|u^{\′\′}\left|\right|}^2+...+a_N{\left|\right|u^{\′\′}\left|\right|}^N}---\left(5\right)$

In formula, α _βrepresent any incident angle of tunnel inner wall, || u " || for this point on imaging plane is to the distance of plane of delineation central point, a ₀, a ₁, a ₂, a _nfor the inside and outside parameter of the omnibearing vision sensor of demarcation, set up the mapping table between an arbitrary pixel of imaging plane and incident angle by formula (5); To see reference document about the concrete derivation of calibration formula and implementation method, Yi-ping Tang, QingWang, Ming-li Zong, Jun Jiang, and Yi-huaZhu, Design of Vertically Aligned Binocular Omnistereo Vision Sensor, EURASIP Journalon Image and Video Processing, 2010, P1 ~ 24; Calibrated result can set up the mapping relations between image coordinate and locus, as shown in Figure 3; Calibration result is stored in active panoramic vision sensor nominal data storehouse.

J saves tunnel cross-section view data reading unit, for full tunnel cross section laser scanning image is carried out piecewise analysis; From the entrance in tunnel to outlet, full tunnel is divided into some joints, with the variable of h (j) as joint, h (j) and Tunnel testing car are along the displacement Z on tunnel longitudinal direction _ibetween relation represent with formula (16),

h(j)＝INT(Z _i/10)+1 (16)

After have chosen a certain joint h (j), obtain the displacement Z comprised in this joint according to formula (16) _i, then with displacement Z all in this joint _ifrom full tunnel cross section laser scanning image storehouse, tunnel cross section laser scanning image is read as filename.

Comprehensive laser intelligence is resolved and cloud data acquiring unit, for saving interior all displacement Z to h (j) _itunnel cross section laser scanning image process, obtain tunnel inner wall edge cloud data; The method of resolving the red laser incident point on laser scanning panoramic picture is the mean flow rate be greater than according to the brightness of the pixel in red laser incident point on imaging plane, first be that the RGB color space conversion of panorama sketch is become HIS color space, then using 1.2 of the mean flow rate on imaging plane times as extracting the threshold value in red laser incident point, in order to obtain the accurate location of laser projection line, the present invention adopts Gaussian approximation method to extract the center of laser projection line, and specific implementation algorithm is:

Step1: β=0, initial orientation angle is set;

Step2: retrieve red laser incident point with azimuthal angle beta from the central point of laser scanning panoramic picture on laser scanning panoramic picture, for pixel azimuthal angle beta also existing the projection of several continuous print red lasers, here select the I component in HIS color space, namely brightness value estimates the center of laser projection line by Gaussian approximation method close to three contiguous pixels of mxm.; Circular is provided by formula (7),

$d=\frac{\ln \left(f(i-1)\right)-\ln \left(f(i-1)\right)}{2\×[\ln \left(f(i-1)\right)-2\ln \left(f\left(i\right)\right)+\ln \left(f(i+1)\right)]}---\left(7\right)$

In formula, f (i-1), f (i) and f (i+1) are respectively the brightness value of three neighbors close to highest brightness value, and d is modified value, and i represents i-th pixel from image center; Therefore estimate that the center of the laser projection line obtained is for (i+d), this value corresponds in formula (5) || u " ||;

Step3: the incident angle α calculating this laser projection point with formula (5) _β, and the information of filename according to laser scanning panoramic image data, namely with displacement Z _ifor the form of filename, obtain Tunnel testing car along the displacement Z on tunnel longitudinal direction _i, then use formula (4) to calculate Tunnel testing car along the displacement Z on tunnel longitudinal direction _iand azimuthal angle beta ' laser projection point in=β situation on tunnel inner wall to active panoramic vision sensor central axis between distance d _{(z, β)}; Formula (8) is finally used to calculate the spatial coordinate location value of this laser projection point;

$\left\{\begin{array}{c}z=z_i-H\\ y=d_{(z,\mathrm{\β})}\×\sin \mathrm{\β}\\ x=d_{(z,\mathrm{\β})}\×\cos \mathrm{\β}\end{array}\right.---\left(8\right)$

In formula, Z _ifor Tunnel testing car is along the displacement on tunnel longitudinal direction, H panoramic scanning light is to the single view O of omnibearing vision sensor _mair line distance, d _{(z, β)}for along the position Z on tunnel longitudinal direction _iand azimuthal angle beta ' laser projection point in=β situation on tunnel inner wall to active panoramic vision sensor central axis between distance, x, y, z are respectively laser projection point relative to the single view O of omnibearing vision sensor _mcoordinate figure, β is position angle;

Step4: change position angle and continue retrieval laser projection point, i.e. β=β+Δ β, Δ β=1;

Step5: judge azimuthal angle beta=360, if set up, retrieval terminates; Otherwise forward Step2 to.

The contour edge cloud data in tunnel lateral direction cross section is obtained by above-mentioned process.

Tunnel axis detection unit, for drawing tunnel linear deformation figure, to carry out three-dimensional modeling according to tunnel axis; Here will along tunnel axis direction continuous print 10 scanning cross-section as a joint, in laser scanning above, each adjacent scanning cross-section is spaced apart 100mm, so just the cloud data unit of 1000mm in a joint is extracted axis, extract the method for axis be to segmentation after tunnel point cloud carry out face of cylinder matching, the step of face of cylinder fitting algorithm is as follows:

STEP1: read the cloud data along continuous print 10 cross sections, tunnel axis direction that comprehensive laser intelligence is resolved and obtained after the process of cloud data acquiring unit, form three-dimensional coordinate matrix P=(X, Y, Z);

STEP2: choose representational some cloud 130 points, as shown in the point in accompanying drawing 11 and accompanying drawing 15, forms three-dimensional coordinate matrix P ⁽¹⁾=(X ⁽¹⁾, Y ⁽¹⁾, Z ⁽¹⁾);

STEP3: with u ⁽⁰⁾=(X ⁽¹⁾, Y ⁽¹⁾, π, 0) and be initial value, carry out solving obtaining according to formula (9), (10), (11) $u^{\left(1\right)}=(u_1^{\left(1\right)},u_2^{\left(1\right)},u_3^{\left(1\right)},u_4^{\left(1\right)});$

$\min f\left(u\right)=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(e_i-R_{(h\left(j\right),\mathrm{\β})})}^2---\left(9\right)$

In formula, R _{(h (j), β)}for the radius of fitting circle cylinder, e _ifor the distance between tunnel inner wall marginal point cloud to axis;

Constraint condition $\left\{\begin{array}{c}{g}_1\left(u\right)=u-{\{\min \left(x_i\right),\min \left(y_i\right),0,-\frac{\mathrm{\π}}{2}\}}^T\≥0\\ g_2\left(u\right)={\{\max \left(x_i\right),\max \left(y_i\right),0,\frac{\mathrm{\π}}{2}\}}^T-u\≥0\end{array}\right.---\left(10\right)$

u＝(x ₀,y ₀,λ,φ) ^T(11)

In formula, x ₀, y ₀for the X-coordinate on a point of fixity on axis, tunnel to be asked and Y-coordinate value, λ is the angle of axis, tunnel to be asked between the projection line and Z axis of XOZ plane, and φ is the angle between axis, tunnel to be asked and XOZ plane; Relation in formula between each variable as shown in Figure 14;

STEP4: suitably increase cloud data to 13000 points, form three-dimensional coordinate matrix P ⁽²⁾=(X ⁽²⁾, Y ⁽²⁾, Z ⁽²⁾);

STEP5: with for initial value, carry out solving obtaining according to formula (9), (10), (11) $u^{\left(2\right)}=(u_1^{\left(2\right)},u_2^{\left(2\right)},u_3^{\left(2\right)},u_4^{\left(2\right)});$

STEP6: order:

P _i＝(X ⁽²⁾,Y ⁽²⁾,Z ⁽²⁾)，

$c=({u_1}^{\left(2\right)},{u_2}^{\left(2\right)},\stackrel{\‾}{Z^{\left(2\right)}}),$

$\stackrel{\→}{a}={({{\cos u}_3}^{\left(2\right)}{{\cos u}_4}^{\left(2\right)},{{\sin u}_3}^{\left(2\right)}{{\cos u}_4}^{\left(2\right)},{{\sin u}_4}^{\left(2\right)})}^T,$

Formula (12) is utilized to calculate e _i;

$e_i=\left|\right|(p_i-c)\×\stackrel{\‾}{a}\left|\right|---\left(12\right)$

In formula: p _i=(x _i, y _i, z _i) be any one measurement point coordinate in original point cloud, c=(x ₀, y ₀, z ₀) be the point of fixity coordinate of on the axis of the face of cylinder, for the axis direction vector of unit length on the face of cylinder.

STEP7: delete e _ibe greater than the point of certain critical value, form three-dimensional coordinate matrix P ⁽³⁾=(X ⁽³⁾, Y ⁽³⁾, Z ⁽³⁾);

STEP8: with for initial value, carry out solving obtaining according to formula (9), (10), (11) $u^{\left(3\right)}=(u_1^{\left(3\right)},u_2^{\left(3\right)},u_3^{\left(3\right)},u_4^{\left(3\right)});$

STEP9: calculate the coordinate of a point of fixity on the axis of the face of cylinder and the direction vector of axis:

$c=({u_1}^{\left(3\right)},{u_2}^{\left(3\right)},\stackrel{\‾}{Z^{\left(3\right)}}),$ $\stackrel{\→}{a}={({{\cos u}_3}^{\left(3\right)}{{\cos u}_4}^{\left(3\right)},{{\sin u}_3}^{\left(3\right)}{{\cos u}_4}^{\left(3\right)},{{\sin u}_4}^{\left(3\right)})}^T.$

In above-mentioned calculating, face of cylinder axis is exactly axis, tunnel, u ₁ ⁽ⁱ⁾in axis detection process, value is approached, u for axis, tunnel corresponding to X-coordinate ₂ ⁽ⁱ⁾in axis detection process, value is approached, u for axis, tunnel corresponding to Y-coordinate ₃ ⁽ⁱ⁾for the included angle X of axis, tunnel between the projection line and Z axis of XOZ plane approaches value, u in axis ₄ ⁽ⁱ⁾for the included angle between axis, tunnel and XOZ plane approaches value in axis, i is the number of times of the approximation computation in axis detection process, as shown in Figure 14.A point on central axis that length is 1000mm tunnel and its direction vector is obtained by above-mentioned algorithm process.

Noise reduction and adjustment processing unit, the length for obtaining the process of tunnel axis extraction module is further that 1000mm tunnel cloud data carries out noise reduction and adjustment processing; According to plane strain condition, put cloud in theory at same angle place, be namely azimuthal angle beta for omnibearing vision sensor, in same azimuthal angle beta situation, the radial coordinate in tunnel is equal; According to the back propagation net of surveying, due to the accidental error measured, the tunnel radial coordinate of same azimuthal angle beta will in normal distribution, namely when time, there is following relational expression;

${\mathrm{\ρ}}_1^{\left(\mathrm{cs}2\right)}~N(\mathrm{\μ},{\mathrm{\σ}}^2)---\left(13\right)$

A cloud is divided into 360 groups, even β ₀=0 °, and 1 °, 2 ° ..., 359 ° }, Statistical Radius coordinate and the relation putting cloud quantity; Conveniently calculate and observe, with radial displacement ρ in accompanying drawing 8 ^(cs2)-R, as horizontal ordinate, is divided into the interval that some 1mm are wide, and ordinate represents the some cloud quantity in each interval; Radial displacement deviation average is rough error point more than the point of 3 times of standard deviations, is deleted; The each group of some cloud traveling through all azimuthal angle beta carries out noise reduction according to this method.

After noise reduction, the radial coordinate of same azimuthal angle beta is still unequal, for the fitting precision improving elliptic cylinder needs to carry out adjustment, the measured value of the radial coordinate of same azimuthal angle beta is adjusted to mean value, namely obtains the mean value ρ of the radial coordinate of same azimuthal angle beta ^*; Again form three-dimensional coordinate matrix after adjustment, use cylindrical coordinate P ^*=(Z ^(cs2), β ^*, ρ ^*) represent; Here ${\mathrm{\β}}^{*}=\mathrm{\β},Z^{\left(\mathrm{cs}2\right)}=\stackrel{\‾}{Z^{\left(3\right)}}.$

Three-dimensional modeling and deformation analysis unit, for tunnel deformation visualization processing; Mainly through comparing the deformation quantity in tunnel in same position, to reflect the convergent deformation situation in local, tunnel or a certain section; As shown in Figure 10, its algorithm mainly processes as follows:

STEP1: the three-dimensional point cloud coordinates matrix P that importing represents with cylindrical coordinate ^*=(Z ^(cs2), β ^*, ρ ^*), initialization process, j=1;

STEP2: read jth joint three dimensional point cloud, adopt ellipse fitting algorithm EFA, the tunnel contour line point-cloud fitting of two dimension is become oval:

(β ^e,ρ ^e)＝EFA(β ^*,ρ ^*)(14)

About the realization of ellipse fitting algorithm EFA see paper DELALOYE D.Development of a newmethodology for measuring deformation in tunnels and shafts with terrestrial laserscanning (LIDAR) using elliptical fitting algorithms [M.S.Thesis] [D] .Kingston:Queen ' sUniversity, 2012;

STEP3: with (β ^e, ρ ^e) be directrix, the line segment of parallel Z axis is bus, generates the elliptic cylinder (h (j) herein=1000mm) that length is h (j):

$\left\{\begin{array}{c}{\mathrm{EX}}^{\left(\mathrm{cs}2\right)}={\mathrm{\ρ}}^e\cos {\mathrm{\β}}^e\\ {\mathrm{EY}}^{\left(\mathrm{cs}2\right)}={\mathrm{\ρ}}^e\sin {\mathrm{\β}}^e\\ {\mathrm{EZ}}^{\left(\mathrm{cs}2\right)}=\left\{z\right|z\∈[0,h\left(j\right)]\}\end{array}\right.$

STEP4: the radial displacement ρ of elliptic cylinder each point ^e-R _{(h (j), β)}as deflection, generate three-dimensional radial displacement cloud atlas; Wherein, ρ ^ecalculate with formula (14), R _{(h (j), β)}for measure first full tunnel cross section time formula (14) calculate, be kept in the benchmark database of Ge Jie transversal section, full tunnel; At calculating ρ ^e-R _{(h (j), β)}time obtain R using h (j) and azimuthal angle beta as index _{(h (j), β)}; Here represent positive deformation by redness, represent negative deformation by blueness, color is more deeply felt and is shown that deformation is larger; To RGB color space, GB color component is set to zero, R component maps, corresponding to the positive deformation of 0 ~ 10mm from 0 ~ 255 pair of positive thermomechanical processing of radial displacement; RG color component is set to zero, G component bear thermomechanical processing from 0 ~ 255 pair of radial displacement and map, corresponding to the negative deformation of 0 ~-10mm.

When the full tunnel cross section measured first, perform tunnel axis extraction unit, noise reduction and adjustment processing unit, then utilize formula (14) ellipse fitting algorithm EFA to calculate (β ^e, ρ ^e), by R _{(h (j), β)}=ρ ^e, wherein: h (j) is jth joint tunnel, and β is position angle, this R _{(h (j), β)}value is kept in the benchmark database of Ge Jie transversal section, full tunnel as the benchmark measured; Simultaneously by face of cylinder fitting algorithm execution result, the coordinate of a point of fixity on axis x=u ₁ ⁽³⁾, y=u ₂ ⁽³⁾, as kernel of section coordinate basis P _{h (j)}(x, y, z) is kept in coordinate reference data storehouse, Ge Jie axis, full tunnel.

Full tunnel profile concatenation unit, for splicing the segmented three-dimensional reconstruction result in full tunnel; With regard to same tunnel cross-section, its SECTION EQUATION under absolute coordinate system only has center horizontal ordinate different with center ordinate from its SECTION EQUATION under relative coordinate system; The effect of axis, tunnel can express attitude and the tendency information in tunnel, and the axis in jth joint tunnel and the axis in jth+1 joint tunnel are continuous print under normal circumstances, and on the other hand, the present invention is also continuous print at tunnel contour marginal point cloud data acquisition; Be divided into several to save in whole tunnel due to during full tunnel deformation monitoring, the length of each joint is h (j); The centre coordinate in jth joint tunnel is obtained in the process of tunnel axis extraction unit on the axis in i.e. tunnel certain a bit, and direction vector $\stackrel{\→}{a}={({{\cos u}_3}^{\left(3\right)}{{\cos u}_4}^{\left(3\right)},{{\sin u}_3}^{\left(3\right)}{{\cos u}_4}^{\left(3\right)},{{\sin u}_4}^{\left(3\right)})}^T;$ For the narrow structure feature in tunnel, here with the centre coordinate in jth joint tunnel and the centre coordinate in normal vector estimation jth+1 joint tunnel, then the centre coordinate in tunnel and the centre coordinate in normal vector estimation jth+2 joint tunnel is saved by jth+1 ..., complete the splicing of the axis in full tunnel like this; When the full tunnel profile measured first, preserve each coordinate figure P of axis, full tunnel in a database _{h (j)}(x, y, z), as the reference data compared in measuring as subsequent tunnel linear deformation; Accompanying drawing 13 is the spliced design sketchs in some joint tunnels.

Displacement monitoring and settlement monitoring unit, have the tolerance of quantification to the global displacement in tunnel and sedimentation; Mainly be reflected in the central axis displacement distortion in tunnel as the global displacement in tunnel and sedimentation, according to the centre coordinate and the direction vector that obtain jth joint tunnel in the process of tunnel axis extraction unit $\stackrel{\→}{a}={({{\cos u}_3}^{\left(3\right)}{{\cos u}_4}^{\left(3\right)},{{\sin u}_3}^{\left(3\right)}{{\cos u}_4}^{\left(3\right)},{{\sin u}_4}^{\left(3\right)})}^T;$ With the centre coordinate in jth joint tunnel and the centre coordinate in normal vector estimation jth+1 joint tunnel, then the centre coordinate in tunnel and the centre coordinate in normal vector estimation jth+2 joint tunnel is saved by jth+1 ..., calculated each coordinate figure P' of this axis, full tunnel successively _{h (j)}(x', y', z'); Displacement monitoring is mainly reflected in X-direction, and settlement monitoring is mainly reflected in Y direction, as shown in Figure 9; With the reference value P of the centre coordinate in jth joint tunnel _{h (j)}x in (x, y, z) and this measure the centre coordinate P' that the jth obtained saves tunnel _{h (j)}x' in (x', y', z') compares, and obtains it is exactly the shift offset in jth joint tunnel; With the reference value P of the centre coordinate in jth joint tunnel _{h (j)}y in (x, y, z) and this measure the centre coordinate P' that the jth obtained saves tunnel _{h (j)}y' in (x', y', z') compares, and obtains Δ y _{h (j)}=y'-y, Δ y _{h (j)}it is exactly the sedimentation side-play amount in jth joint tunnel; By the above-mentioned displacement distortion calculating each joint on full tunnel central axis; Finally by the displacement of each joint distortion Δ x _{h (j)}with Δ y _{h (j)}be kept in Test database; What calculate is relative displacement and the settling data amount of tunnel profile by above-mentioned; Utilize the spatial positional information reading and be fixed on RFID in tunnel wall and store, and obtain absolute displacement and the settling data amount of tunnel profile in this, as absolute coordinates benchmark.

Tunnel profile becomes the amount of moving three-dimensional visualization unit, for having quantification, visual expression to the global displacement in tunnel and sedimentation; Three-dimensional visualization process measure the numerical value, the deflection information that obtain become intuitively, represent with Figure and Image, present to tunnel safety monitoring personnel with the physical quantity of spatial variations in time; Here represent positive deformation by redness, represent negative deformation by blueness, color is more deeply felt and is shown that deformation is larger; To RGB color space, GB color component is set to zero, R component aligns thermomechanical processing from 0 ~ 255 and maps, corresponding to the positive deformation of 0 ~ 255mm; RG color component is set to zero, G component map from 0 ~ 255 negative thermomechanical processing, corresponding to the negative deformation of 0 ~-255mm; Then color-coded on the joint corresponding to tunnel to corresponding to the vertical section deformation quantity of each joint on full tunnel central axis; Broken line mode of saving each on central axis couples together by the coordinate figure on last basis full tunnel central axis, shows the deformation situation of whole tunnel profile by visual means.

Embodiment 2

In the present embodiment, roughly the same, difference is that the controller on Tunnel testing car increases memory capacity to all the other implementations, after Tunnel testing car test survey terminates, tunnel scan image is sent to central monitoring center server together by network.

Embodiment 3

In the present embodiment, all the other implementations roughly the same, difference is different according to tunnel cross section, method tunnel cross section being asked to barycenter is adopted in tunnel axis detection unit, using the axis of the barycenter line of each transversal section as tunnel, with tunnel inner wall marginal point to the distance of axis as tunnel deformation detected parameters.

Embodiment 4

In the present embodiment, roughly the same, difference is that the detection speed of Tunnel testing car is different to all the other implementations, for than embodiment 1 more at a slow speed or faster detection speed, adopts the shooting speed camera adapted with it.

Embodiment 5

In the present embodiment, roughly the same, difference is the position of Tunnel testing car being installed active panoramic vision sensor to all the other implementations, and active panoramic vision sensor is configured in Tunnel testing tailstock portion.

Embodiment 6

In the present embodiment, roughly the same, to all the other implementations difference configure active panoramic vision sensor, RFID reader, measuring wheel, wireless interface transmitting element and controller on the subway train and bullet train of normal operation.

Embodiment 7

In the present embodiment, all the other implementations roughly the same, difference is that comprehensive laser intelligence is resolved and cloud data acquiring unit is configured in the controller on Tunnel testing car, controller processes online to tunnel cross section laser scanning image, then the contour edge cloud data in tunnel lateral direction cross section is sent to central monitoring center server by network; In order to process tunnel cross section laser scanning image in the controller, need active panoramic vision sensor nominal data storehouse to be configured in the storage unit of controller; Process with regard to the direct cloud data to tunnel in central monitoring center server in addition, avoid the transmission of mass image data.

Claims (10)

1. the full section of tunnel high speed dynamical health pick-up unit based on active panoramic vision, comprise Tunnel testing car and processor, it is characterized in that, described Tunnel testing car is included in the inspection vehicle body that tunnel track is walked, and is arranged on active panoramic vision sensor, RFID reader and measuring wheel on inspection vehicle body;

Described active panoramic vision sensor comprises co-axially fixed panorama laser projection light source and omnibearing vision sensor, described panorama laser projection light source is used for for tunnel inner wall provides tunnel to break cross section panoramic scanning light, and described omnibearing vision sensor is for gathering the panoramic picture of tunnel inner wall;

Described reader, for reading the tunnel reference data being fixed on and tunnel inner wall stores in RFID;

Described measuring wheel, for calculating the travel distance of Tunnel testing car;

Described processor, for parsing the three dimensional point cloud on tunnel cross section from described panoramic picture, and the travel distance of integrating tunnel inspection vehicle reconstructs full tunnel inner wall three-dimensional model, again the full tunnel inner wall three-dimensional model built and initial full tunnel inner wall three-dimensional model are contrasted, analyze tunnel-type variable.

2. as claimed in claim 1 based on the full section of tunnel high speed dynamical health pick-up unit of active panoramic vision, it is characterized in that, described panorama laser projection light source comprises base, be fixed on the ring shape generating laser on base, and for conical minute surface that reflection circle cast laser transmitter projects circle laser out provides tunnel to break cross section panoramic scanning light for tunnel inner wall.

3., as claimed in claim 1 based on the full section of tunnel high speed dynamical health pick-up unit of active panoramic vision, it is characterized in that, described processor comprises:

Omnibearing vision sensor demarcates unit, for determining the parameter of mapping relations between the X-Y scheme picture point in three dimensions point and video camera imaging plane;

J saves tunnel cross-section view data reading unit, for full tunnel cross section laser scanning image segmentation joint is read in order and analyzed;

Comprehensive laser intelligence is resolved and cloud data acquiring unit, for saving interior all displacement Z to h (j) _itunnel cross section laser scanning image process, obtain tunnel inner wall edge cloud data;

Tunnel axis detection unit, for drawing tunnel linear deformation figure, to carry out three-dimensional modeling according to tunnel axis;

Noise reduction and adjustment processing unit, for carrying out noise reduction and adjustment processing to the tunnel cloud data that described tunnel axis extraction module process obtains further;

Three-dimensional modeling and deformation analysis unit, for tunnel deformation visualization processing;

Full tunnel profile concatenation unit, for splicing the segmented three-dimensional reconstruction result in full tunnel;

Displacement monitoring and settlement monitoring unit, for there being the tolerance of quantification to the global displacement in tunnel and sedimentation;

Tunnel profile becomes the amount of moving three-dimensional visualization unit, for having quantification, visual expression to the global displacement in tunnel and sedimentation.

4., based on a method for the full section of tunnel high speed dynamical health pick-up unit described in any one of claims 1 to 3, it is characterized in that, comprise step:

1) control Tunnel testing car to walk on tunnel track, panorama laser projection light source is adopted to provide tunnel to break cross section panoramic scanning light for tunnel inner wall, and gather tunnel cross section laser scanning image by omnibearing vision sensor, utilize RFID reader to read the tunnel reference data being fixed on and tunnel inner wall stores in RFID simultaneously;

2) estimate the running distance of Tunnel testing car, and tunnel cross section laser scanning image is associated with running distance store;

3) from tunnel cross section laser scanning image, the spatial coordinate location value of panorama laser projection point on tunnel cross section is parsed, the three dimensional point cloud namely on tunnel cross section;

4) according to described running distance and corresponding three dimensional point cloud, full tunnel inner wall three-dimensional model is reconstructed;

5) the full tunnel inner wall three-dimensional model built and the full tunnel inner wall three-dimensional model built at first are contrasted, analyze tunnel-type variable.

5. full section of tunnel high speed dynamical health detection method as claimed in claim 4, it is characterized in that, full tunnel is divided into some joints, and with the variable of h (j) as joint, h (j) and Tunnel testing car are along the displacement Z on tunnel longitudinal direction _ibetween relation represent with formula (16),

h(j)＝INT(Z _i/10)+1 (16)

After have chosen a certain joint h (j), obtain the displacement Z comprised in this joint according to formula (16) _i, then with displacement Z all in this joint _ifrom described full tunnel cross section laser scanning image storehouse, tunnel cross section laser scanning image is read as filename.

6. full section of tunnel high speed dynamical health detection method as claimed in claim 5, is characterized in that, for saving interior all displacement Z to h (j) _itunnel cross section laser scanning image process, obtain tunnel inner wall edge cloud data; Specific implementation algorithm is:

Step1: β=0, initial orientation angle is set;

Step2: retrieve laser projection point with azimuthal angle beta from the central point of laser scanning panoramic picture on laser scanning panoramic picture, for pixel azimuthal angle beta also existing several continuous print laser projections, select the I component in HIS color space, namely brightness value estimates the center of laser projection line by Gaussian approximation method close to three contiguous pixels of mxm.; Circular is provided by formula (7),

$d=\frac{\ln \left(f(i-1)\right)-\ln \left(f(i+1)\right)}{2\×[\ln \left(f(i-1)\right)-2\ln \left(f\left(i\right)\right)+\ln \left(f(i+1)\right)]}---\left(7\right)$

In formula, f (i-1), f (i) and f (i+1) are respectively the brightness value of three neighbors close to highest brightness value, and d is modified value, and i represents i-th pixel from image center; Therefore estimate that the center of the laser projection line obtained is for (i+d);

Step3: the incident angle α calculating this laser projection point _β, and according to Tunnel testing car along the displacement Z on tunnel longitudinal direction _i, calculate Tunnel testing car along the displacement Z on tunnel longitudinal direction _iand azimuthal angle beta ' laser projection point in=β situation on tunnel inner wall to active panoramic vision sensor central axis between distance d _{(z, β)}; Formula (8) is finally used to calculate the spatial coordinate location value of this laser projection point;

$\left\{\begin{array}{c}z=Z_i-H\\ y=d_{(z,\mathrm{\β})}\×\sin \mathrm{\β}\\ x=d_{(z,\mathrm{\β})}\×\cos \mathrm{\β}\end{array}\right.---\left(8\right)$

In formula, Z _ifor Tunnel testing car is along the displacement on tunnel longitudinal direction, H panoramic scanning light is to the single view O of omnibearing vision sensor _mair line distance, d _{(z, β)}for along the position Z on tunnel longitudinal direction _iand azimuthal angle beta ' laser projection point in=β situation on tunnel inner wall to active panoramic vision sensor central axis between distance, x, y, z are respectively laser projection point relative to the single view O of omnibearing vision sensor _mcoordinate figure, β is position angle;

Step4: change position angle and continue retrieval laser projection point, i.e. β=β+Δ β, Δ β=1;

Step5: judge azimuthal angle beta=360, if set up, retrieval terminates; Otherwise forward Step2 to.

7. full section of tunnel high speed dynamical health detection method as claimed in claim 6, it is characterized in that, when building three-dimensional model, tunnel linear deformation figure is drawn according to tunnel axis, extract the method for axis be to segmentation after tunnel point cloud carry out face of cylinder matching, the step of face of cylinder fitting algorithm is as follows:

STEP1: read the cloud data along several cross sections of continuous print, tunnel axis direction that comprehensive laser intelligence is resolved and obtained after the process of cloud data acquiring unit, form three-dimensional coordinate matrix P=(X, Y, Z);

STEP2: choose representational some cloud, forms three-dimensional coordinate matrix P ⁽¹⁾=(X ⁽¹⁾, Y ⁽¹⁾, Z ⁽¹⁾);

STEP3: with u ⁽⁰⁾=(X ⁽¹⁾, Y ⁽¹⁾, π, 0) and be initial value, carry out solving obtaining according to formula (9), (10), (11) $u^{\left(1\right)}=(u_1^{\left(1\right)},u_2^{\left(1\right)},u_3^{\left(1\right)},u_4^{\left(1\right)});$

$\min f\left(u\right)=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{(e_i-R_{(h\left(j\right),\mathrm{\β})})}^2---\left(9\right)$

In formula, R _{(h (j), β)}for the radius of fitting circle cylinder, e _ifor the distance between tunnel inner wall marginal point cloud to axis;

u＝(x ₀，y ₀，λ，φ) ^T

In formula, x ₀, y ₀for the X-coordinate on a point of fixity on axis, tunnel to be asked and Y-coordinate value, λ is the angle of axis, tunnel to be asked between the projection line and Z axis of XOZ plane, and φ is the angle between axis, tunnel to be asked and XOZ plane;

STEP4: suitably increase cloud data, forms three-dimensional coordinate matrix P ⁽²⁾=(X ⁽²⁾, Y ⁽²⁾, Z ⁽²⁾);

STEP5: with for initial value, carry out solving obtaining according to formula (9), (10), (11) $u^{\left(2\right)}=(u_1^{\left(2\right)},u_2^{\left(2\right)},u_3^{\left(2\right)},u_4^{\left(2\right)});$

STEP6: order:

P _i＝(X ⁽²⁾,Y ⁽²⁾,Z ⁽²⁾)，

$c=({u_1}^{\left(2\right)},{u_2}^{\left(2\right)},\stackrel{\‾}{Z^{\left(2\right)}}),$

$\stackrel{\→}{a}={({{\cos u}_3}^{\left(2\right)},{{\cos u}_4}^{\left(2\right)},{{\sin u}_3}^{\left(2\right)},{{\cos u}_4}^{\left(2\right)},{{\sin u}_4}^{\left(2\right)})}^T,$

Formula (12) is utilized to calculate e _i;

$e_i=\left|\right|(p_i-c)\×\stackrel{\→}{a}\left|\right|---\left(12\right)$

In formula: p _i=(x _i, y _i, z _i) be any one measurement point coordinate in original point cloud, c=(x ₀, y ₀, z ₀) be the point of fixity coordinate of on the axis of the face of cylinder, for the axis direction vector of unit length on the face of cylinder;

STEP7: delete e _ibe greater than the point of certain critical value, form three-dimensional coordinate matrix P ⁽³⁾=(X ⁽³⁾, Y ⁽³⁾, Z ⁽³⁾);

STEP8: with for initial value, carry out solving obtaining according to formula (9), (10), (11) $u^{\left(3\right)}=(u_1^{\left(3\right)},u_2^{\left(3\right)},u_3^{\left(3\right)},u_4^{\left(3\right)});$

STEP9: calculate the coordinate of a point of fixity on the axis of the face of cylinder and the direction vector of axis: $c=({u_1}^{\left(3\right)},{u_2}^{\left(3\right)},\stackrel{\‾}{Z^{\left(3\right)}}),$ $\stackrel{\→}{a}={({{\cos u}_3}^{\left(3\right)},{{\cos u}_4}^{\left(3\right)},{{\sin u}_3}^{\left(3\right)},{{\cos u}_4}^{\left(3\right)},{{\sin u}_4}^{\left(3\right)})}^T,$

In above-mentioned calculating, face of cylinder axis is exactly axis, tunnel, u ₁ ⁽ⁱ⁾in axis detection process, value is approached, u for axis, tunnel corresponding to X-coordinate ₂ ⁽ⁱ⁾in axis detection process, value is approached, u for axis, tunnel corresponding to Y-coordinate ₃ ⁽ⁱ⁾for the included angle X of axis, tunnel between the projection line and Z axis of XOZ plane approaches value, u in axis ₄ ⁽ⁱ⁾for the included angle between axis, tunnel and XOZ plane approaches value in axis, i is the number of times of the approximation computation in axis detection process.

8. full section of tunnel high speed dynamical health detection method as claimed in claim 4, is characterized in that, in step 5) in, the algorithm of the deformation quantity in tunnel in same position is as follows:

STEP1: the three-dimensional point cloud coordinates matrix P that importing represents with cylindrical coordinate ^*=(Z ^(cs2), β ^*, ρ ^*), initialization process, j=1;

STEP2: read jth joint three dimensional point cloud, adopt ellipse fitting algorithm EFA, the tunnel contour line point-cloud fitting of two dimension is become oval:

(β ^e,ρ ^e)＝EFA(β ^*,ρ ^*) (14)

STEP3: with (β ^e, ρ ^e) be directrix, the line segment of parallel Z axis is bus, generates the elliptic cylinder (h (j) herein=1000mm) that length is h (j):

$\left\{\begin{array}{c}{\mathrm{EX}}^{\left(\mathrm{cs}2\right)}={\mathrm{\ρ}}^e\cos {\mathrm{\β}}^e\\ {\mathrm{EY}}^{\left(\mathrm{cs}2\right)}={\mathrm{\ρ}}^e\sin {\mathrm{\β}}^e\\ {\mathrm{EZ}}^{\left(\mathrm{cs}2\right)}=\left\{z\right|z\∈[0,h\left(j\right)]\}\end{array}\right.---\left(15\right)$

STEP4: the radial displacement ρ of elliptic cylinder each point ^e-R _{(h (j), β)}as deflection, generate three-dimensional radial displacement cloud atlas; Wherein, ρ ^ecalculate with formula (14), R _{(h (j), β)}for measure first full tunnel cross section time formula (14) calculate, at calculating ρ ^e-R _{(h (j), β)}time obtain R using h (j) and azimuthal angle beta as index _{(h (j), β)}; To RGB color space, GB color component is set to zero, R component maps, corresponding to the positive deformation of 0 ~ 10mm from 0 ~ 255 pair of positive thermomechanical processing of radial displacement; RG color component is set to zero, G component bear thermomechanical processing from 0 ~ 255 pair of radial displacement and map, corresponding to the negative deformation of 0 ~-10mm.

9. full section of tunnel high speed dynamical health detection method as claimed in claim 8, is characterized in that, the segmented three-dimensional reconstruction result in full tunnel spliced; Be divided into several to save in whole tunnel due to during full tunnel deformation monitoring, the length of each joint is h (j); The centre coordinate in jth joint tunnel is obtained in the process of tunnel axis extraction unit on the axis in i.e. tunnel certain a bit and direction vector $\stackrel{\→}{a}={({{\cos u}_3}^{\left(3\right)},{{\cos u}_4}^{\left(3\right)},{{\sin u}_3}^{\left(3\right)},{{\cos u}_4}^{\left(3\right)},{{\sin u}_4}^{\left(3\right)})}^T,$ For the narrow structure feature in tunnel, with the centre coordinate in jth joint tunnel and the centre coordinate in normal vector estimation jth+1 joint tunnel, then save the centre coordinate in tunnel and the centre coordinate in normal vector estimation jth+2 joint tunnel by jth+1, complete the splicing of the axis in full tunnel by that analogy.

10. full section of tunnel high speed dynamical health detection method as claimed in claim 9, is characterized in that, described method also comprises the tolerance global displacement in tunnel and sedimentation being had to quantification; Mainly be reflected in the central axis displacement distortion in tunnel as the global displacement in tunnel and sedimentation, according to the centre coordinate and the direction vector that obtain jth joint tunnel in the process of tunnel axis extraction unit $\stackrel{\→}{a}={({{\cos u}_3}^{\left(3\right)},{{\cos u}_4}^{\left(3\right)},{{\sin u}_3}^{\left(3\right)},{{\cos u}_4}^{\left(3\right)},{{\sin u}_4}^{\left(3\right)})}^T;$ With the centre coordinate in jth joint tunnel and the centre coordinate in normal vector estimation jth+1 joint tunnel, then the centre coordinate in tunnel and the centre coordinate in normal vector estimation jth+2 joint tunnel is saved by jth+1 ..., calculated each coordinate figure P' of this axis, full tunnel successively _{h (j)}(x', y', z'); Displacement monitoring is mainly reflected in X-direction, and settlement monitoring is mainly reflected in Y direction; With the reference value P of the centre coordinate in jth joint tunnel _{h (j)}x in (x, y, z) and this measure the centre coordinate P' that the jth obtained saves tunnel _{h (j)}x' in (x', y', z') compares, and obtains Δ x _{h (j)}=x'-x, Δ x _{h (j)}it is exactly the shift offset in jth joint tunnel; With the reference value P of the centre coordinate in jth joint tunnel _{h (j)}y in (x, y, z) and this measure the centre coordinate P' that the jth obtained saves tunnel _{h (j)}y' in (x', y', z') compares, and obtains Δ y _{h (j)}=y'-y, Δ y _{h (j)}it is exactly the sedimentation side-play amount in jth joint tunnel; By the above-mentioned displacement distortion calculating each joint on full tunnel central axis; Finally by the displacement of each joint distortion Δ x _{h (j)}with Δ y _{h (j)}be kept in Test database; What calculate is relative displacement and the settling data amount of tunnel profile by above-mentioned.