CN101458072A - Three-dimensional contour outline measuring set based on multi sensors and measuring method thereof - Google Patents

Abstract

The invention relates to a multisensor-based three-dimensional outline measuring device and a measuring method thereof. The measuring device comprises a base plate, a vertical movement platform, a working table surface and light measurement sensor assemblies. The working table surface is provided with four or more than four light measurement sensor assemblies which are evenly distributed around a measuring hole; each light measurement sensor assembly comprises a laser, a camera and a camera mount; the lower end of each camera mount is buckled on an assembly guide way which is fixed on the working table surface; the cameras are arranged on the camera mounts; and the lasers are fixed on the working table surface. The invention utilizes the light measurement sensor assemblies to improve the measurement speed and provides an effective technical measure for the batch detection of industrial site components; and the invention adjusts field coverage and amplification ratio according to the surface shape of a tested component and is adapted to the measurement of a complex curve surface of a blade or a conical gear, and the like, which cannot be effectively measured by using conventional methods.

Description

A kind of three-dimensional contour outline measuring set and measuring method thereof based on multisensor

Technical field

The invention belongs to the optical detection field, relate to a kind of optical detecting method of three-D profile, particularly a kind of complex-curved 360 degree three-dimensional contour outline measuring set and measuring methods thereof based on multisensor.

Background technology

In recent years, the application of object three-dimensional contour outline detection technique is universal day by day, is widely used in fields such as product design, quality control, reverse engineering emulation, machine vision and biologic medical.In commercial production, significance is arranged to measuring such as the profile of complicated precision components such as blade of aviation engine, turbine blade, crowngear, cross helical gear for the control product quality.

Existing object measuring three-dimensional profile method has many kinds, is divided into three kinds substantially, is respectively three coordinate measuring engine measurement, stereoscopic vision mensuration and structural light measurement method, wherein:

1) three coordinate measuring engine measurement efficient is low, is not suitable for the batch detection of part;

2) stereoscopic vision mensuration precision is limited, is not suitable for the detection of precision component;

3) the structural light measurement method can be divided into the dot structure photo measure again, area-structure light is measured and line-structured light is measured three classes.Wherein, the slow efficient of dot structure photo measure speed is low; The area-structure light measurement is subject to the interference of piece surface reflectivity, and precision is lower; So dot structure photo measure and area-structure light measuring method all are not suitable for the batch profile measurement of industry spot to precision workpiece.

The line-structured light mensuration, claiming light cross-section method again, is to reappear the object dimensional pattern with one or more light (finishing tool) image, promptly extracts the finishing tool center from the finishing tool image, utilize principle of triangulation that the pointwise of finishing tool center is found the solution then, obtain shape face three-dimensional data.This method is the line source projection owing to what adopt, can survey the data of a section at every turn, and measuring speed is fast than the point-to-point measurement method, and, only need the sub-scanning telecontrol equipment of one dimension just can realize the full form planar survey.Signal Processing is simple and reliable, need not complicated fringe analysis just can unique elevation information of determining each measurement point, it is concavo-convex to differentiate object automatically, even there is breakpoint to make the pattern can not be continuous on the object, also can not influence measurement.In many measuring methods, the line-structured light three-dimensional measurement with its wide range, big visual field, degree of precision, finishing tool image information be easy to extract, characteristics such as real-time and initiatively controlled, in industrial environment, obtained in recent years using widely.But traditional line-structured light measuring method is based on single-sensor, can only measure a contour of object on the direction at every turn, and the parts measurement that has 360 degree profiles for blade, gear so then can't be competent at.

Summary of the invention

The objective of the invention is to overcome the shortcoming of above-mentioned prior art, a kind of three-dimensional contour outline measuring set and measuring method thereof based on multisensor is provided, this device cooperates its measuring method can realize the 360 degree measuring three-dimensional profiles that part is complex-curved fast.

The objective of the invention is to solve by the following technical programs:

This three-dimensional contour outline measuring set based on multisensor, comprise base plate, be fixed in the vertical moving platform on the base plate, through the support bar level to being fixed in the work top on the base plate and being located at photo measure sensor module on the work top, it is characterized in that, described vertical moving platform is provided with stepper motor and leading screw, on the described leading screw objective table is installed, objective table is located at the below of work top, described objective table is provided with anchor clamps, the center of described work top offers measured hole, described work top is provided with four or more photo measure sensor module, described photo measure sensor module be uniformly distributed in measured hole around, described photo measure sensor module comprises laser instrument, video camera and camera mount, described camera mount lower end button is located on the package traces, and described package traces is fixed on the work top, video camera is installed on the camera mount, and described laser instrument is fixed on the work top.

Above-mentioned photo measure sensor module is made up of a camera mount, a laser instrument and a video camera, described video camera is more than 1,000,000 pixels, its lens focus is within 8～100mm, described four or more photo measure sensor module is symmetrical distribution, and the angle of adjacent light survey sensor inter-module is 0～90 degree.

Above-mentioned camera mount comprises slide block, camera tracks, manual displacement platform and video camera sleeve, described slide block is installed on the package traces, also be provided with dog screw on the slide block, slide block is fixedlyed connected with the camera tracks that is in tilted layout through connection rib, slidably manual displacement platform is installed on the camera tracks, be fixed with the video camera sleeve on the manual displacement platform, described video camera is installed in the video camera sleeve, and also is provided with the Camera Positioning screw rod in the video camera sleeve.

The output end of image of above-mentioned video camera is connected with computing machine through image card.

A kind of measuring three-dimensional profile method of said apparatus specifically may further comprise the steps:

1) at first with the measured object clamping on the anchor clamps of objective table, regulate the height of objective table on the vertical moving platform, measured object passed by measured hole and be on the work top;

2) open laser instrument, each laser instrument shines the measured object cross section simultaneously from different perspectives, and each video camera is taken the finishing tool image from different perspectives;

3) objective table is done motion in one dimension in vertical direction under the drive of stepper motor and leading screw, and the whole surface of measured object is all arrived by laser illumination;

4) after video camera photographs measured object whole surface image data, the image that each video camera obtains is sent into computing machine respectively through image card, carrying out the finishing tool center successively by data processing software extracts and coordinate conversion, through the wide splicing of the laggard road wheel of coordinate conversion, finally obtain the three-dimensional information on the whole surface of measured object.

The method that extract at above-mentioned finishing tool center is: at first extract the skeleton of finishing tool, ask for the normal direction of each pixel on the skeleton then, the pixel that is specially on the skeleton carries out curve fitting, and the method for curve fitting is utilized y=ax for getting 10 picture elements ²+ bx+c carries out the second order match, establishes (x ₀, y ₀) be the coordinate of pixel, then (x ₀, y ₀) point slope be t=2ax ₀+ b, if t=0, then the weighted mean direction is the y direction, if t ≠ 0, then (x ₀, y ₀) point the method line slope be n=-1/t, the weighted mean direction is a normal direction; After calculating the normal direction of each point on the skeleton of image, ask for the intensity profile of finishing tool on its normal direction; Utilize the grey scale centre of gravity method to obtain the centre of gravity place of this finishing tool at last, be the point position of testee, these point are coupled together the outline line that has just formed the cross section at this place.

Above-mentioned coordinate conversion is meant finishing tool center image planimetric coordinates is converted to finishing tool center image physics physical plane coordinate, and is specific as follows:

If (u v) is a finishing tool center image planimetric coordinates, (X is a finishing tool center image physics physical plane coordinate Y), and the coordinate transformation relation between them is as follows:

$\left\{\begin{array}{c}({\mathrm{<figure-callout\; id="1"\; label="Xh"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">Xh</figure-callout>}}_1+{\mathrm{<figure-callout\; id="2"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_2+h_3)-u({\mathrm{<figure-callout\; id="7"\; label="Xh"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">Xh</figure-callout>}}_7+{\mathrm{<figure-callout\; id="8"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_8+1)=0\\ ({\mathrm{Xh}}_4+{\mathrm{<figure-callout\; id="5"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_5+h_6)-v({\mathrm{<figure-callout\; id="7"\; label="Xh"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">Xh</figure-callout>}}_7+{\mathrm{<figure-callout\; id="8"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_8+1)=0\end{array}\right.$

Wherein, h=[h ₁h ₂h ₃h ₄h ₅h ₆h ₇h ₈] ^TBe the camera parameters vector, demarcate through the camera parameters vector and determine.The scaling method of described camera parameters vector h is as follows:

Make a demarcation target surface that physics physical plane coordinate is known, the physics physical plane coordinate of establishing N actual scene point on the target surface is (x _n, y _n), n=1,2 ... N is by the image coordinate (u of camera acquisition corresponding pixel points _n, v _n), n=1,2 ... N, by following simultaneous equations (N=4):

$\left[\begin{array}{cccccccc}{x}_1& y_1& 1& 0& 0& 0& {-u}_1{x}_1& {-u}_1{y}_1\\ 0& 0& 0& x_1& y_1& 1& {-v}_1{x}_1& {-u}_1{y}_1\\ x_2& y_2& 1& 0& 0& 0& {-u}_2{y}_2& {-u}_2{y}_2\\ 0& 0& 0& x_2& y_2& 1& {-v}_2{x}_2& {-v}_2{y}_2\\ x_3& y_3& 1& 0& 0& 0& {-u}_3{x}_3& {-u}_3{y}_3\\ 0& 0& 0& x_3& y_3& 1& {-v}_3{x}_3& {-v}_3{y}_3\\ x_4& y_4& 1& 0& 0& 0& {-u}_4{x}_4& {-u}_4{x}_4\\ 0& 0& 0& x_4& y_4& 1& {-v}_4{x}_4& {-v}_4{x}_4\end{array}\right]\left[\begin{array}{c}{h}_1\\ h_2\\ h_3\\ h_4\\ h_5\\ h_6\\ h_7\\ {\mathrm{<figure-callout\; id="8"\; label="h"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">h</figure-callout>}}_8\end{array}\right]=\left[\begin{array}{c}{u}_1\\ v_1\\ u_2\\ v_2\\ u_3\\ v_3\\ u_4\\ v_4\end{array}\right]$

Obtain each Camera calibration matrix h=[h successively ₁h ₂h ₃h ₄h ₅h ₆h ₇h ₈] ^T, finish the camera parameters vector and demarcate.

Three-dimensional contour outline measuring set and measuring method thereof based on multisensor of the present invention adopts a plurality of photo measure sensor modules simultaneously measured object to be carried out the line structure measurement, carry out image mosaic by computing machine then, can accurately measure the part with 360 degree profiles, the present invention specifically has following some superiority:

1) configuration by a plurality of photo measure sensor modules has improved measuring speed, for the batch detection of industry spot part provides effective technical means;

2) can regulate visual field and magnification according to the surface configuration characteristics of tested part, be applicable to the complex-curved measurement that effectively to measure as classic methods such as blade, conical gear;

3) three-dimensional contour outline measuring set based on multisensor of the present invention has only the elevating movement of one dimension, has effectively reduced source of error, has improved measuring accuracy;

4) the present invention is easy and simple to handle, ease for use good.Utilize collaborative computer to handle image, realized the robotization of measuring.

Description of drawings

Fig. 1 is the three-dimensional contour outline measuring set structural representation based on multisensor of the present invention;

Fig. 2 is a photo measure sensor module structural representation of the present invention;

Wherein: 1 is package traces; 2,3,4 and 5 be the photo measure sensor module; 6 is work top; 7 is laser instrument; 8 is anchor clamps; 9 is objective table; 10 is base plate; 11 is the vertical moving platform; 12 is dog screw; 13 is camera mount; 14 is the manual displacement platform; 15 is the video camera sleeve; 16 is the Camera Positioning screw rod; 17 is video camera; 18 is camera lens; 19 is support bar; 20 is measured hole; 21 is camera tracks; 22 is slide block; 23 is stepper motor.

Fig. 3 is a profile splicing synoptic diagram of the present invention;

Fig. 4 is the measurement result figure of blade of aviation engine.

Embodiment

Below in conjunction with accompanying drawing the present invention is done and to describe in further detail:

Fig. 1 is the three-dimensional contour outline measuring set structural representation based on multisensor of the present invention, this three-dimensional contour outline measuring set based on multisensor, comprise base plate 10, be fixed in the vertical moving platform 11 on the base plate 10, through support bar 19 levels to being fixed in the work top 6 on the base plate 10 and being located at four or more photo measure sensor module on the work top 6, Fig. 1 is a kind of embodiment of the present invention, it is provided with four photo measure sensor modules 2 on work top 6,3,4 and 5, the center of work top 6 offers measured hole 20, described vertical moving platform 11 is the big little wedge of, which is provided with stepper motor 23 and leading screw, stepper motor 23 is installed in the upper end of vertical moving platform 11, its output shaft is connected with lead screw shaft, leading screw is vertically mounted on the vertical plane of vertical moving platform 11 inboards and objective table 9 is installed on the leading screw, objective table 9 is to be sleeved on the leading screw by movable block, and objective table 9 carries out the one dimension vertical moving thereby stepper motor 23 drives leading screw rotation drive.Objective table 9 is located at the below of work top 6, and objective table 9 is provided with anchor clamps 8, described anchor clamps 8 be in measured hole 20 under, photo measure sensor module 2～5 be distributed in described measured hole 20 around.

Fig. 2 is a photo measure sensor module structural representation, described photo measure sensor module 2～5 comprises laser instrument 7, video camera 17 and camera mount 13, described camera mount 13 lower ends button is located on the package traces 1, and described package traces 1 is fixed on the work top 6.Camera mount 13 comprises slide block 22, camera tracks 21, manual displacement platform 14 and video camera sleeve 15, the slide block 22 of camera mount 13 through its lower end is installed on the package traces 1, can on package traces 1, axially be free to slide, package traces 1 is fixed on the work top 6, also be provided with dog screw 12 on the slide block 22, be used for slide block 22 is fixed in package traces 1, slide block 22 is fixedlyed connected with the camera tracks 21 that is in tilted layout through connection rib, being equipped with on the camera tracks 21 can be along the manual displacement platform 14 of camera tracks 21 oblique slips, be fixed with video camera sleeve 15 on the manual displacement platform 14, video camera 17 is installed in the video camera sleeve 15, and in video camera sleeve 15, also be provided with the Camera Positioning screw rod 16 that is used for positioning shooting machine 17, camera lens 18 is installed on the outer lower portion of video camera sleeve 15, the measured hole direction of camera lens 18 on work top 6.The output end of image of video camera 17 is connected with computing machine through image card.Laser instrument 7 is fixed on the work top 6 on package traces 1 next door, and each photo measure assembly is furnished with a laser instrument 7, measured hole 20 directions of the only directive work top 6 that laser instrument 7 sends.

Described photo measure sensor module is made up of a camera mount 13, a laser instrument 7 and a video camera 17, video camera 17 is more than 1,000,000 pixels, its lens focus is within 8～100mm, the video camera 17 of telephoto lens (focal length〉30mm) is responsible for measuring small size profiles (radius＜3mm), (video camera of focal length＜30mm) 17 is responsible for measuring large scale profiles (radius〉3mm) to the short-focus mirror head, described four or more photo measure sensor module is symmetrical distribution, and the angle of adjacent light survey sensor inter-module is 0～90 degree.2～5 angles of four photo measure sensor modules among Fig. 1 are 90 degree.

Utilize the concrete measuring method of three-dimensional contour outline measuring set based on multisensor shown in Figure 1 as follows:

1) at first with the measured object clamping on the anchor clamps 8 of objective table 9, regulate the height of objective tables 9 on vertical moving platform 11 by stepper motor 23, measured object passed by measured hole 20 and be on the work top 6; Regulate the position of the camera lens 18 of video cameras 17 by manual displacement platform 14, make it can photograph best image;

2) open laser instrument 7, each laser instrument 7 shines the measured object cross section simultaneously from different perspectives, and each video camera 17 is taken the finishing tool image from different perspectives;

3) objective table 9 is done motion in one dimension in vertical direction under the drive of stepper motor 23 and leading screw, and the whole surface of measured object is all shone by laser instrument 7;

4) after video camera 17 photographs measured object whole surface image data, the image that each video camera 17 obtains is sent into computing machine respectively through image card, carrying out the finishing tool center successively by data processing software extracts and coordinate conversion, through the wide splicing of the laggard road wheel of coordinate conversion, the physics physical plane coordinate of the picture point that described profile splicing is converted to each video camera 17 exactly is incorporated into together, finally obtains the three-dimensional information on the whole surface of measured object.

The method that extract at the finishing tool center is: at first extract the skeleton of finishing tool, ask for the normal direction of each pixel on the skeleton then, the pixel that is specially on the skeleton carries out curve fitting, and the method for curve fitting is utilized y=ax for getting 10 picture elements ²+ bx+c carries out the second order match, establishes (x ₀, y ₀) be the coordinate of pixel, then (x ₀, y ₀) point slope be t=2ax ₀+ b, if t=0, then the weighted mean direction is the y direction, if t ≠ 0, then (x ₀, y ₀) point the method line slope be n=-1/t, the weighted mean direction is a normal direction; After calculating the normal direction of each point on the skeleton of image, ask for the intensity profile of finishing tool on its normal direction; Utilize the grey scale centre of gravity method to obtain the centre of gravity place of this finishing tool at last, be the point position of testee, these point are coupled together the outline line that has just formed the cross section at this place.

The grey scale centre of gravity method is direct arrangement according to striation gray-scale value in a certain interval, gets grey scale centre of gravity along horizontal ordinate and represents the finishing tool center, and this method can accurately be located the Gaussian distribution center of striation.Gravity model appoach is exactly to each row in the image, by the center of following formulas Extraction grey scale centre of gravity as finishing tool.Be located in the delegation of image, the u coordinate of finishing tool pixel is u _m(m=0,1,2 ..., N), its gray-scale value is g _m(m=0,1,2 ..., N), wherein N is the number of finishing tool pixel.

Utilize the center of gravity formula can obtain the center u of finishing tool _n:

$u_n=\underset{m=0}{\overset{N}{\mathrm{\&Sigma;}}}{g}_m{u}_m/\underset{m=0}{\overset{N}{\mathrm{\&Sigma;}}}{g}_m$

The arithmetic speed of gravity model appoach is very fast, especially after image is carried out filtering and noise reduction and definite edge pixel, adopts gravity model appoach to ask for the precision that the finishing tool center can obtain inferior pixel again.

The skeleton that below how to obtain finishing tool is a very important step, by it being analyzed the center that the skeleton of image as can be known should be positioned at the finishing tool image, can show the trend of finishing tool fully, and have single connectedness.Generally, the finishing tool image is carried out the lines that refinement obtains and to satisfy these requirements substantially.The thinning process that obtains skeleton is exactly a boundary pixel of repeatedly peeling bianry image off, up to the process of the wide connection line (skeleton) of single pixel of acquisition.But the connectedness that must keep target during the stripping borderline pixel promptly can not change the topological property of original image.Can carry out refinement to bianry image with several different methods.Wherein, it is comparatively commonly used to adopt template to carry out the method for refinement.

The above coordinate conversion is meant finishing tool center image planimetric coordinates is converted to finishing tool center image physics physical plane coordinate, finishing tool center image planimetric coordinates is to extract the coordinate figure of each point under this video camera of handling on the cross section contour that obtains through the finishing tool center behind the camera acquisition image, physics physical plane coordinate is meant chooses a common initial point outside each video camera, with this initial point is the coordinate figure of each pixel of benchmark, and the method for coordinate conversion is specific as follows:

If (u v) is a finishing tool center image planimetric coordinates, (X is a physics physical plane coordinate Y), and the coordinate transformation relation between them is as follows:

$\left\{\begin{array}{c}({\mathrm{<figure-callout\; id="1"\; label="Xh"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">Xh</figure-callout>}}_1+{\mathrm{<figure-callout\; id="2"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_2+h_3)-u({\mathrm{<figure-callout\; id="7"\; label="Xh"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">Xh</figure-callout>}}_7+{\mathrm{<figure-callout\; id="8"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_8+1)=0\\ ({\mathrm{Xh}}_4+{\mathrm{<figure-callout\; id="5"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_5+h_6)-v({\mathrm{<figure-callout\; id="7"\; label="Xh"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">Xh</figure-callout>}}_7+{\mathrm{<figure-callout\; id="8"\; label="Yh"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">Yh</figure-callout>}}_8+1)=0\end{array}\right.$

Wherein, h=[h ₁h ₂h ₃h ₄h ₅h ₆h ₇h ₈] ^TBe the camera parameters vector, demarcate through the camera parameters vector and determine.The scaling method of camera parameters vector h is as follows:

Make a demarcation target surface that physics physical plane coordinate is known, the physics physical plane coordinate of establishing N actual scene point on the target surface is (x _n, y _n), n=1,2 ... N is by the image coordinate (u of camera acquisition corresponding pixel points _n, v _n), n=1,2 ... N, by following simultaneous equations (N=4):

$\left[\begin{array}{cccccccc}{x}_1& {\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">y</figure-callout>}}_1& 1& 0& 0& 0& {-u}_1{x}_1& {-u}_1{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">y</figure-callout>}}_1\\ 0& 0& 0& x_1& {\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="A200910020835E00181.png,A200910020835E00182.png"\; state="\{\{state\}\}">y</figure-callout>}}_1& 1& {-v}_1{x}_1& {-u}_1{y}_1\\ x_2& {\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">y</figure-callout>}}_2& 1& 0& 0& 0& {-u}_2{y}_2& {-u}_2{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">y</figure-callout>}}_2\\ 0& 0& 0& x_2& {\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">y</figure-callout>}}_2& 1& {-v}_2{x}_2& {-v}_2{y}_2\\ x_3& {\mathrm{<figure-callout\; id="3"\; label="y"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">y</figure-callout>}}_3& 1& 0& 0& 0& {-u}_3{x}_3& {-u}_3{\mathrm{<figure-callout\; id="3"\; label="y"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">y</figure-callout>}}_3\\ 0& 0& 0& x_3& {\mathrm{<figure-callout\; id="3"\; label="y"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">y</figure-callout>}}_3& 1& {-v}_3{x}_3& {-v}_3{y}_3\\ x_4& y_4& 1& 0& 0& 0& {-u}_4{x}_4& {-u}_4{x}_4\\ 0& 0& 0& x_4& y_4& 1& {-v}_4{x}_4& {-v}_4{x}_4\end{array}\right]\left[\begin{array}{c}{h}_1\\ h_2\\ h_3\\ h_4\\ h_5\\ h_6\\ h_7\\ {\mathrm{<figure-callout\; id="8"\; label="h"\; filenames="A200910020835E00181.png"\; state="\{\{state\}\}">h</figure-callout>}}_8\end{array}\right]=\left[\begin{array}{c}{u}_1\\ v_1\\ u_2\\ v_2\\ u_3\\ v_3\\ u_4\\ v_4\end{array}\right],$

Obtain each Camera calibration matrix h=[h successively ₁h ₂h ₃h ₄h ₅h ₆h ₇h ₈] ^T, finish the camera parameters vector and demarcate.

Embodiment

Be that example is analyzed feasibility of the present invention with the blade of aviation engine profile measurement below.Employed three-dimensional contour outline measuring set based on multisensor is the device with four photo measure assemblies as shown in Figure 1 in the present embodiment.

1. before measuring beginning, need do following a series of preliminary work:

(1) at first adjusts four laser instruments 7 in same plane, concrete implementation method is to accept any two finishing tools that laser instrument 7 sends with vertical screen, when accepting screen by near and from the close-by examples to those far off moving process, two finishing tools are shown as point-blank all the time, at this moment we think these two finishing tools in same surface level, and promptly laser instrument 7 is also in same surface level;

(2) blade of aviation engine is clamped on the anchor clamps 8;

(3) scope of the tested blade profile of bigness scale, the breadth extreme of blade back and leaf basin is about 50mm, and leading edge and trailing edge are very little arc-shaped surface, can only probably find out its curved surface characteristic.Regulate the position of camera mount 13 on package traces 1, and the focal length of adjustment video camera 17, drive stepping motor makes objective table 9 motions, guarantee in the process of blade of aviation engine lifting, all enough big in the visual field of the video camera 17 of each photo measure sensor module, and the broken string situation does not appear in the finishing tool image that photographs.

2. images acquired

The demarcation masterplate is put after determining well in the position of camera mount 13 and video camera 17, and fine setting Camera Positioning screw rod 14 makes the central cross of camera acquisition image overlap with the central cross of demarcating masterplate, gathers uncalibrated image.Put blade of aviation engine afterwards,, determine that per 200 steps are that 0.05mm gathers the primary measured object image according to the step-length of the width and the stepper motor of finishing tool.

3. multi-sensor data splicing

The image that each video camera obtains is sent into computing machine respectively through image card, carrying out the finishing tool center successively by data processing software extracts and coordinate conversion, after the view data process coordinate conversion that will obtain, carrying out the profile splicing handles, the physics physical plane coordinate of the picture point that described profile splicing is converted to each video camera exactly is incorporated into together, as shown in Figure 3, finally obtain the three-dimensional information on the whole surface of blade of aviation engine, spliced blade of aviation engine profile as shown in Figure 4.

In sum, three-dimensional contour outline measuring set and measurement side thereof based on multisensor of the present invention An advantage of method is that the configuration by multisensor has improved measuring speed, is the industry spot part Batch detection fast and effectively technological means is provided, the present invention can be according to tested part in addition The surface configuration characteristics regulate visual field and magnifying power, be applicable to such as conventional methods such as blade, bevel gears The complex-curved measurement that can't effectively measure removes in addition, and measurement mechanism of the present invention is when measuring Only have the elevating movement of one dimension, reduced source of error, improved the certainty of measurement of system, and This device is easy and simple to handle, easy-to-use.

Claims (8)

1. three-dimensional contour outline measuring set based on multisensor, comprise base plate (10), be fixed in the vertical moving platform (11) on the base plate (10), through support bar (19) level to being fixed in the work top (6) on the base plate (10) and being located at photo measure sensor module on the work top (6), it is characterized in that, described vertical moving platform (11) is provided with stepper motor (23) and leading screw, objective table (9) is installed on the described leading screw, objective table (9) is located at the below of work top (6), described objective table (9) is provided with anchor clamps (8), the center of described work top (6) offers measured hole (20), described work top (6) is provided with four or more photo measure sensor module, described photo measure sensor module be uniformly distributed in measured hole (20) around, described photo measure sensor module comprises laser instrument (7), video camera (17) and camera mount (13), described camera mount (13) lower end button is located on the package traces (1), described package traces (1) is fixed on the work top (6), video camera (17) is installed on the camera mount (13), and described laser instrument (7) is fixed on the work top (6).

2. the three-dimensional contour outline measuring set based on multisensor according to claim 1, it is characterized in that, described photo measure sensor module is made up of a camera mount (13), a laser instrument (7) and a video camera (17), described video camera (17) is more than 1,000,000 pixels, its lens focus is within 8～100mm, described four or more photo measure sensor module is symmetrical distribution, and the angle of adjacent light survey sensor inter-module is 0～90 degree.

3. the three-dimensional contour outline measuring set based on multisensor according to claim 1, it is characterized in that, described camera mount (13) comprises slide block (22), camera tracks (21), manual displacement platform (14) and video camera sleeve (15), described slide block (22) is installed on the package traces (1), also be provided with dog screw (12) on the slide block (22), slide block (22) is fixedlyed connected with the camera tracks that is in tilted layout (21) through connection rib, slidably manual displacement platform (14) is installed on the camera tracks (21), be fixed with video camera sleeve (15) on the manual displacement platform (14), described video camera (17) is installed in the video camera sleeve (15), and also is provided with Camera Positioning screw rod (16) in video camera sleeve (15).

4. the three-dimensional contour outline measuring set based on multisensor according to claim 1 is characterized in that, the output end of image of described video camera (17) is connected with computing machine through image card.

5. the measuring three-dimensional profile method based on the described device of claim 1 is characterized in that, specifically may further comprise the steps:

1) at first with the measured object clamping on the anchor clamps (8) of objective table (9), regulate the height of objective table (9) on vertical moving platform (11), measured object passed by measured hole (20) and be on the work top (6);

2) open laser instrument (7), each laser instrument (7) shines the measured object cross section simultaneously from different perspectives, and each video camera (17) is taken the finishing tool image from different perspectives;

3) objective table (9) is done motion in one dimension in vertical direction under the drive of stepper motor (23) and leading screw, and the whole surface of measured object is all shone by laser instrument (7);

4) after video camera (17) photographs measured object whole surface image data, the image that each video camera (17) obtains is sent into computing machine respectively through image card, carrying out the finishing tool center successively by data processing software extracts and coordinate conversion, through the wide splicing of the laggard road wheel of coordinate conversion, finally obtain the three-dimensional information on the whole surface of measured object.

6. measuring three-dimensional profile method according to claim 5, it is characterized in that, the method that extract at described finishing tool center is: the skeleton that at first extracts finishing tool, ask for the pixel that the normal direction of each pixel on the skeleton is specially on the skeleton then and carry out curve fitting, the method for curve fitting is utilized y=ax for getting 10 picture elements ²+ bx+c carries out the second order match, establishes (x ₀, y ₀) be the coordinate of pixel, then (x ₀, y ₀) point slope be t=2ax ₀+ b, if t=0, then the weighted mean direction is the y direction, if t ≠ 0, then (x ₀, y ₀) point the method line slope be n=-1/t, the weighted mean direction is a normal direction; After calculating the normal direction of each point on the skeleton of image, ask for the intensity profile of finishing tool on its normal direction; Utilize the grey scale centre of gravity method to obtain the centre of gravity place of this finishing tool at last, be the point position of testee, these point are coupled together the outline line that has just formed the cross section at this place.

7. measuring three-dimensional profile method according to claim 5 is characterized in that, described coordinate conversion is meant finishing tool center image planimetric coordinates is converted to finishing tool center image physics physical plane coordinate that concrete grammar is as follows:

If (u v) is a finishing tool center image planimetric coordinates, (X is a finishing tool center image physics physical plane coordinate Y), and the coordinate transformation relation between them is as follows:

$\left\{\begin{array}{c}({\mathrm{Xh}}_1+{\mathrm{Yh}}_2+h_3)-u({\mathrm{Xh}}_7+{\mathrm{Yh}}_8+1)=0\\ ({\mathrm{Xh}}_4+{\mathrm{Yh}}_5+h_6)-v({\mathrm{Xh}}_7+{\mathrm{Yh}}_8+1)=0\end{array}\right.$

Wherein, h=[h ₁h ₂h ₃h ₄h ₅h ₆h ₇h ₈] ^TBe the camera parameters vector, demarcate through the camera parameters vector and determine.

8. measuring three-dimensional profile method according to claim 7 is characterized in that, the scaling method of described camera parameters vector h is as follows:

Make a demarcation target surface that physics physical plane coordinate is known, the physics physical plane coordinate of establishing N actual scene point on the target surface is (x _n, y _n), be (u by the image coordinate of camera acquisition corresponding pixel points _n, v _n), n=1 wherein, 2 ... N, and N=4, according to following simultaneous equations:

$\left[\begin{array}{cccccccc}{x}_1& y_1& 1& 0& 0& 0& {-u}_1{x}_1& {-u}_1{y}_1\\ 0& 0& 0& x_1& y_1& 1& {-v}_1{x}_1& {-u}_1{y}_1\\ x_2& y_2& 1& 0& 0& 0& {-u}_2{y}_2& {-u}_2{y}_2\\ 0& 0& 0& x_2& y_2& 1& {-v}_2{x}_2& {-v}_2{y}_2\\ x_3& y_3& 1& 0& 0& 0& {-u}_3{x}_3& {-u}_3{y}_3\\ 0& 0& 0& x_3& y_3& 1& {-v}_3{x}_3& {-v}_3{y}_3\\ x_4& y_4& 1& 0& 0& 0& {-u}_4{x}_4& {-u}_4{x}_4\\ 0& 0& 0& x_4& y_4& 1& {-v}_4{x}_4& {-y}_4{x}_4\end{array}\right]\left[\begin{array}{c}{h}_1\\ h_2\\ h_3\\ h_4\\ h_5\\ h_6\\ h_7\\ h_8\end{array}\right]=\left[\begin{array}{c}{u}_1\\ v_1\\ u_2\\ v_2\\ u_3\\ v_3\\ u_4\\ v_4\end{array}\right],$

Obtain each Camera calibration matrix h=[h successively ₁h ₂h ₃h ₄h ₅h ₆h ₇h ₈] ^T, finish the camera parameters vector and demarcate.

Images