TW201310004A - Correlation arrangement device of digital images - Google Patents
Correlation arrangement device of digital images Download PDFInfo
- Publication number
- TW201310004A TW201310004A TW100125747A TW100125747A TW201310004A TW 201310004 A TW201310004 A TW 201310004A TW 100125747 A TW100125747 A TW 100125747A TW 100125747 A TW100125747 A TW 100125747A TW 201310004 A TW201310004 A TW 201310004A
- Authority
- TW
- Taiwan
- Prior art keywords
- camera
- digital image
- cameras
- image
- correlation device
- Prior art date
Links
Landscapes
- Length Measuring Devices By Optical Means (AREA)
Abstract
Description
本發明是有關於一種編列數位影像關係裝置,尤指一種可應用於量測大範圍區域位移和應變等相關應用之編列數位影像關係裝置者。The present invention relates to a device for composing digital image relationships, and more particularly to a device for recording digital image relationships that can be applied to related applications such as displacement and strain in a wide range of regions.
按,數位影像關係法早在1980年代出被發展出來,Peters和Ranson於1982年第一次利用數位影像的特性來觀察物件受力之後的行為(如:文獻Peters,W.H.,and Ranson,W.F.,1982,“Digital image techniques on experimental stress analysis”Opt.Eng.,21(3),427-431.),開啟數位影像應用於變形量測先河。之後上述二人又與他人合作,首先對影像的前處理和數值的收斂和求值做探討(如:文獻Sutton,M.A.,Wolters,W.J.,et al,1983,“Determination of displacements using an improved digital image correlation method,”Image Vision Computing,1(3),133-139.、Chu,T.C.,Peters,W.H.,Sutton,M.A.,McNeil,S.R.,1985,“Application of digital image correlation to experimental mechanics,”Image Vision Computing,25(3),232-245.以及Sutton,M.A.,McNeil,S.R.,et al,1988,“Effect of subpixels image restoration on digital correlation error estimates,”Opt.Eng.,27(10),870-877.),後續研究包括對數位影像數值所做插值法的優劣探討,和選取的子影像(Subimage)大小對收斂時間和準確性的影響等等。According to the digital image relationship method, it was developed in the 1980s. In 1982, Peters and Ranson first used the characteristics of digital images to observe the behavior of objects after they were stressed (eg, Peters, WH, and Ranson, WF, 1982, "Digital image techniques on experimental stress analysis" Opt. Eng., 21(3), 427-431.), the digital image is turned on for the first measurement of deformation measurement. Afterwards, the above two people cooperated with others, firstly discussing the pre-processing of images and the convergence and evaluation of numerical values (eg, Sutton, MA, Wolters, WJ, et al, 1983, “Determination of displacements using an improved digital image”. Correlation method," Image Vision Computing, 1(3), 133-139., Chu, TC, Peters, WH, Sutton, MA, McNeil, SR, 1985, "Application of digital image correlation to experimental mechanics," Image Vision Computing , 25(3), 232-245. and Sutton, MA, McNeil, SR, et al, 1988, "Effect of subpixels image restoration on digital correlation error estimates," Opt. Eng., 27(10), 870-877 .), follow-up studies include the merits of interpolation methods for digital image values, and the effect of subimage size on convergence time and accuracy.
而基本概念建立之後相關的數位影像法的應用研究如同雨後春筍般不斷出現,如Sutton(如:文獻Sutton,M.A.,Peters,W.H.,Chao,Y.J.,Luo,P.F.,1993,“Accurate measurement of three-dimensional Deformations in deformable and rigid bodies using computer vision,”Experimental Mechanics,33(2),123-132.)等人於1993年首度將將此方法應用於3D立體變形的量測,接著繼續將三維立體的觀測概念利用於表面輪廓測量研究(如:文獻Sutton,M.A.,McNeil,S.R.,et al,1997,“Measurement of surface profile using digital image correlation,“Experimental Mechanics,37(1),13-20.)。除了Sutton等人之外,許多學者專家如Oulamara(如:文獻Oulamara,A.,Tribillon,G.,Dunvernoy,J.,1988,“Subpixel speckle displacement using a digital processing technique,”Journal of Modern Optics,37,1201-1211.)和James(如:文獻James,M.R.,Morris,W.L.,Cox,B.N.,1990,“A high accuracy automated strain-field mapper,”Experimental Mechanics,30(1),60-67.)等人亦利用數位影像法建立自動變形或應變場量測系統,Septanika(如:文獻Septanika,E.G.,Ernst,L.J.,van den Hoof,L.A.C.M.,1998,“An automatic and interactive large-deformation measurement system based on image processing,”Experimental Mechanics,38(3),181-188.)則發展量測表面大應變率的量測系統。After the establishment of the basic concept, the application research of related digital image method has sprung up like Sutton (such as: Sutton, MA, Peters, WH, Chao, YJ, Luo, PF, 1993, "Accurate measurement of three-dimensional". Deformations in deformable and rigid bodies using computer vision, "Experimental Mechanics, 33(2), 123-132.) et al., for the first time in 1993, will apply this method to the measurement of 3D stereoscopic deformation, and then continue to be three-dimensional The observation concept is utilized in surface profilometry studies (eg, Sutton, MA, McNeil, SR, et al, 1997, "Measurement of surface profile using digital image correlation," "Experimental Mechanics, 37(1), 13-20.). In addition to Sutton et al., many scholars such as Oulamara (eg, Oulamara, A., Tribillon, G., Dunvernoy, J., 1988, "Subpixel speckle displacement using a digital processing technique," Journal of Modern Optics, 37 , 1201-1211.) and James (eg, James, MR, Morris, WL, Cox, BN, 1990, "A high accuracy automated strain-field mapper," Experimental Mechanics, 30 (1), 60-67.) Et al. also used the digital image method to establish an automatic deformation or strain field measurement system, Septanika (eg, Septanika, EG, Ernst, LJ, van den Hoof, LACM, 1998, "An automatic and interactive large-deformation measurement system based on Image processing, "Experimental Mechanics, 38(3), 181-188.) develops a measurement system that measures the large strain rate of the surface.
目前市場上應用影像相關方法為基礎之非接觸式影像量測結構體變形的產品有Correlated Solutions公司開發的VIC 3D產品,如第1圖所示,其係為一支架10結兩部攝影機11、12所構成立體視覺之量測實體架構,可用來量測物體全域表面形變的系統,此技術乃利用物體表面的特徵,做為表面位移判斷與比對的標的,其結合攝影機影像捕捉與影像關係法演算法,使得3D表面形變的量測能在最快的時間內完成,並得到最佳的解析度,任何表面細微的位移,均能在數據與影像上呈現出來,藉由後處理,也可了解單點或表面的位移與應變;At present, the product of the non-contact image measurement structure based on the image-related method on the market is a VIC 3D product developed by Correlated Solutions, as shown in Fig. 1, which is a bracket 10 knot two cameras 11, The 12-dimensional stereoscopic physical structure can be used to measure the surface deformation of an object. This technique uses the surface features of the object as the target for surface displacement judgment and comparison. It combines camera image capture with image relationship. The algorithm makes the measurement of 3D surface deformation can be completed in the fastest time, and the best resolution is obtained. Any slight displacement of the surface can be presented on the data and image, and by post-processing, Can understand the displacement and strain of a single point or surface;
然而,該項設備量測範圍不足以進行戶外場景如岩石山脈走勢觀測,因為傳統式應變計可達到高精度的應變量測,若要進行多點或全域的量測,在應變計的黏貼上是既費工又費時,不但不符合經濟成本,現地量測儀器的線路配置也是一大問題,且位移量測是結構監測中一個重要的項目,也是安全性評估的依據之一,因為結構產生的位移反應與其材料特性和剛度有關,過大的位移量可能意味結構剛度的衰減,內部構件已受到損傷,此時需再以其他儀器檢測,進一步判定損傷之位置與程度,並估算現有結構的剛度或承載力是否符合安全標準,以利後續工程師做維修、加勁或拆除重建等判斷。However, the measurement range of this equipment is not sufficient for outdoor scenes such as rocky mountain trend observation, because the traditional strain gauge can achieve high-precision strain measurement, if multi-point or global measurement is to be carried out, the strain gauge is pasted. It is laborious and time consuming, not only does not meet the economic cost. The line configuration of the local measuring instrument is also a big problem, and displacement measurement is an important project in structural monitoring, and it is also one of the basis for safety evaluation, because the structure is generated. The displacement response is related to its material properties and stiffness. Excessive displacement may mean the attenuation of the structural stiffness, and the internal components have been damaged. At this time, it is necessary to detect with other instruments to further determine the location and extent of the damage and estimate the stiffness of the existing structure. Whether the bearing capacity meets the safety standards, in order to facilitate subsequent engineers to make repairs, stiffening or dismantling and other judgments.
因此,如何快速、便利又有效量測結構體的變形是工程界不斷致力發展與研究的目標。基於上述原因,本發明係多部攝影機結合編列數位影像關係進行變形分析之技術,發展非接觸式量測結構體變形的方法,具備全域、遠距與便利等優點,可用來進行快速之位移量測,以攝影機量測得到之位移場進行應變分析,計算求得結構在外力作用的應變反應。Therefore, how to quickly, conveniently and effectively measure the deformation of the structure is the goal of the engineering community to continuously develop and research. Based on the above reasons, the present invention is a technique for performing deformation analysis by combining digital image relationships with a plurality of cameras, and developing a non-contact measurement method for deformation of a structure, which has the advantages of global domain, distance and convenience, and can be used for rapid displacement. The strain field is measured by the displacement field measured by the camera, and the strain response of the structure under the external force is calculated.
如第1圖所示為習知3D立體變形的量測方法,使用兩部攝影機架構進行量測。當量測範圍超過兩部攝影機量測可視範圍時,只能觀測局部影像,無法量測物體遮敝區域。加上,針對不同物體需要調整攝影機姿態,並重新進行攝影機校正。針對習知技術存在問題,本專利採用共圓方式,解決物體有遮敝問題,且藉由多部攝影機可增加交會區域及對於校正參數互相比對。As shown in Fig. 1, the measurement method of the conventional 3D stereoscopic deformation is performed using two camera architectures. When the equivalent measurement range exceeds the measurement range of the two cameras, only the partial image can be observed, and the concealed area of the object cannot be measured. Plus, you need to adjust the camera pose for different objects and re-calibrate the camera. In view of the problems with the prior art, this patent adopts a co-circle method to solve the concealing problem of the object, and the intersection area can be increased by multiple cameras and the correction parameters are compared with each other.
該量測桿係呈一中空之圓狀,使第一、二及第三攝影機之三點中心光軸共圓交於一點,且使第一、二及第三攝影機利用調整裝置於量測桿上進行適當間距之調整。當現場量測重建三維資訊時,不需要先經過複雜校正處理,系統內部已針對每個角度建立校正參數。當觀測不同大小之待測物時,只要移動調整裝置於量測桿上進行適當間距之調整並取像,即可快速進行重建三維資訊。The measuring rod is in a hollow shape, so that the three-point central optical axes of the first, second and third cameras are rounded at one point, and the first, second and third cameras are used to measure the rod by using the adjusting device. Adjust the appropriate spacing on the top. When the on-site measurement reconstructs the three-dimensional information, it does not need to undergo complicated correction processing first, and the correction parameters are established inside the system for each angle. When observing different sizes of the object to be tested, the three-dimensional information can be quickly reconstructed by simply moving the adjustment device on the measuring rod to adjust and take the appropriate spacing.
本發明之主要目的係在於:The main purpose of the invention is to:
1.量測儀器架構與試驗程序簡單、可行性及準確度高,且裝置具有平衡裝置確保量測結果穩定。1. The measuring instrument architecture and test procedure are simple, feasible and accurate, and the device has a balance device to ensure stable measurement results.
2.量測範圍可任意選取,大面積量測。2. The measurement range can be arbitrarily selected and measured in a large area.
3.藉由多部攝影機內外部參數互相比對,能正確估算位移及應變參數。3. By comparing the internal and external parameters of multiple cameras, the displacement and strain parameters can be correctly estimated.
4.所建立之自動化計算程式,快速將受力前後之數位影像資料進行分析,並計算出位移及應變值。4. The established automatic calculation program quickly analyzes the digital image data before and after the force and calculates the displacement and strain values.
5.攝影機所監測區域受力變形之影像紀錄可永久儲存作為其他方面研究之參考。5. The image record of the force deformation of the area monitored by the camera can be permanently stored as a reference for other research.
為達上述之目的,本發明係一種編列數位影像關係裝置,包含有:一具有基座及量測桿之承載機構;至少三個活動設於量測桿上之調整機構;分別設於各調整裝置上之第一、二及第三攝影機所構成之影像擷取機構,且該第一、二及第三攝影機係以第一及第二校正點作為量測基準點;一設於第一、二及第三攝影機光軸上之校正板;以及一與第一、二及第三攝影機連接之處理機構,其係具有一編列數位影像運算模組,可將第一、二及第三攝影機所擷取之數位影像資料進行分析。For the purpose of the present invention, the present invention is a digital image correlation device comprising: a carrier having a base and a measuring rod; and at least three adjusting mechanisms disposed on the measuring rod; respectively, respectively An image capturing mechanism formed by the first, second and third cameras on the device, wherein the first, second and third cameras use the first and second correction points as the measurement reference points; a calibration plate on the optical axes of the second and third cameras; and a processing mechanism coupled to the first, second and third cameras, which has a digital image computing module for the first, second and third cameras Digital image data captured for analysis.
於本發明之一實施例中,該量測桿係呈一中空之圓弧狀,使第一、二及第三攝影機之三點中心光軸共圓交於一點,且使第一、二及第三攝影機可利用調整裝置於量測桿上進行適當間距之調整。In an embodiment of the invention, the measuring rod is in the shape of a hollow arc, so that the central optical axes of the three points of the first, second and third cameras are intersected at one point, and the first and second The third camera can adjust the appropriate spacing on the measuring rod by using the adjusting device.
於本發明之一實施例中,各調整機構係分別包含有一與量測桿活動結合之滑座、一設於滑座上之調整座、一設於滑座上之平衡裝置、及一設於滑座上之電控裝置,而該第一、二及第三攝影機係分別設於各調整座上。In an embodiment of the present invention, each of the adjusting mechanisms includes a sliding seat coupled to the measuring rod, a adjusting seat disposed on the sliding seat, a balancing device disposed on the sliding seat, and a The electric control device on the sliding seat, and the first, second and third camera systems are respectively arranged on the respective adjusting seats.
於本發明之一實施例中,該滑座之兩端係分別設有與量測桿抵靠之滑輪及靠緊輪,而該靠緊輪係連接有一調整旋鈕。In an embodiment of the present invention, the two ends of the sliding seat are respectively provided with a pulley and a pressing wheel that abut against the measuring rod, and the fastening wheel is connected with an adjusting knob.
於本發明之一實施例中,該第一校正點係設於基座上,用於固定第一、二及第三攝影機之光軸中心位於同一光軸點,做為量測初始點,而該第二校正點係為所在環境之一參考點,且與第二攝影機光軸中心及第一校正點皆位於同一光軸線上,做為初始化量測使用。In an embodiment of the present invention, the first calibration point is disposed on the pedestal for fixing the centers of the optical axes of the first, second, and third cameras at the same optical axis point as the initial point of measurement, and The second calibration point is a reference point of the environment, and is located on the same optical axis as the optical axis center of the second camera and the first calibration point, and is used for initial measurement.
於本發明之一實施例中,該第二校正點係配合有一標示桿,可架設於遠方之待測物位置,以利第一、二及第三攝影機進行後續重覆量測使用。In an embodiment of the invention, the second calibration point is coupled with a pointing rod, which can be placed at a remote object to be tested, so that the first, second and third cameras can be used for subsequent repeated measurement.
於本發明之一實施例中,該標示桿上係具有可進行校正之標示刻度。In an embodiment of the invention, the indicator bar has an indicator scale that can be corrected.
於本發明之一實施例中,該校正板係為具有固定間距之校正格點中心座,作為第一、二及第三攝影機校正時所需之校正點。In an embodiment of the invention, the calibration plate is a correction grid center seat having a fixed pitch as a correction point required for the first, second and third cameras to be calibrated.
於本發明之一實施例中,該校正板之校正格點可為圓形或方形。In an embodiment of the invention, the calibration grid of the calibration plate may be circular or square.
於本發明之一實施例中,當校正板之校正格點與第一、二及第三攝影機之影像平面不平行時,所擷取影像之格點將形成扭曲形狀。In an embodiment of the invention, when the correction grid of the calibration plate is not parallel to the image planes of the first, second and third cameras, the lattice points of the captured image will form a twisted shape.
於本發明之一實施例中,該處理機構更進一步連接有一預警單元。In an embodiment of the invention, the processing mechanism is further connected to an early warning unit.
於本發明之一實施例中,該編列數位影像運算模組至少包含有相互校正參數單元、全景拼接製作單元、全景編列數位影像單元及三維重建單元,可使第一、二及第三攝影機配合校正板進行校正,並將各攝影機所擷取之數位影像資料進行分析,且計算出位移及應變值。In an embodiment of the present invention, the digital video computing module includes at least a mutual calibration parameter unit, a panoramic mosaic processing unit, a panoramic digital image unit, and a three-dimensional reconstruction unit, so that the first, second, and third cameras can cooperate. The calibration plate performs calibration, and the digital image data captured by each camera is analyzed, and the displacement and strain values are calculated.
於本發明之一實施例中,該相互校正參數單元係可調整第一攝影機之光軸垂直於校正板之可視範圍內。In an embodiment of the invention, the mutual correction parameter unit adjusts the optical axis of the first camera to be perpendicular to the visual range of the calibration plate.
於本發明之一實施例中,該相互校正參數單元係可取得第一攝影機之水平及垂直方向誤差值,以第一攝影機之移動取得其影像之水平與垂直方向的Scale Factor(mm/pixel),並產生一世界座標。In an embodiment of the present invention, the mutual correction parameter unit can obtain the horizontal and vertical direction error values of the first camera, and obtain the scale factor (mm/pixel) of the horizontal and vertical directions of the image by the movement of the first camera. And produce a world coordinate.
於本發明之一實施例中,該相互校正參數單元係可調整第二及第三攝影機,使其觀察範圍與第一攝影機相同,並校正第二及第三攝影機所形成之三維重建。In an embodiment of the invention, the mutual correction parameter unit adjusts the second and third cameras to have the same viewing range as the first camera, and corrects the three-dimensional reconstruction formed by the second and third cameras.
於本發明之一實施例中,該相互校正參數單元係可調整攝影機第二攝影機,使其光軸垂直於校正板之可視範圍內。In an embodiment of the invention, the mutual correction parameter unit adjusts the second camera of the camera such that its optical axis is perpendicular to the visual range of the calibration plate.
於本發明之一實施例中,該相互校正參數單元係可取得第二攝影機之水平及垂直方向誤差值,以第二攝影機之移動取得其影像之水平與垂直方向的Scale Factor(mm/pixel),並產生一世界座標。In an embodiment of the present invention, the mutual correction parameter unit can obtain the horizontal and vertical direction error values of the second camera, and obtain the scale factor (mm/pixel) of the horizontal and vertical directions of the image by the movement of the second camera. And produce a world coordinate.
於本發明之一實施例中,該相互校正參數單元係可調整第一及第三攝影機,使其觀察範圍與第二攝影機相同,並校正第一及第三攝影機所形成之三維重建。In an embodiment of the invention, the mutual correction parameter unit adjusts the first and third cameras to have the same viewing range as the second camera, and corrects the three-dimensional reconstruction formed by the first and third cameras.
於本發明之一實施例中,該相互校正參數單元係可調整攝影機第三攝影機,使其光軸垂直於校正板之可視範圍內。In an embodiment of the invention, the mutual correction parameter unit adjusts the third camera of the camera such that its optical axis is perpendicular to the visual range of the calibration plate.
於本發明之一實施例中,該相互校正參數單元係可取得第三攝影機之水平及垂直方向誤差值,以第三攝影機之移動取得其影像之水平與垂直方向的Scale Factor(mm/pixel),並產生一世界座標。In an embodiment of the present invention, the mutual correction parameter unit can obtain the horizontal and vertical direction error values of the third camera, and obtain the scale factor (mm/pixel) of the horizontal and vertical directions of the image by the movement of the third camera. And produce a world coordinate.
於本發明之一實施例中,該相互校正參數單元係可調整第一及第二攝影機,使其觀察範圍與攝影機第三相同,並校正第一及第二攝影機所形成之三維重建。In an embodiment of the invention, the mutual correction parameter unit adjusts the first and second cameras to have the same viewing range as the third camera, and corrects the three-dimensional reconstruction formed by the first and second cameras.
於本發明之一實施例中,該全景拼接製作單元係可選取大面積量測範圍,主要先藉由左、右影像,先行校正求出第一、二及第三攝影機之參數,並根據求得知參數,將左、右兩個影像畫面經過影像匹配投影至同一場景。In an embodiment of the present invention, the panoramic splicing unit can select a large-area measurement range, and firstly obtain the parameters of the first, second, and third cameras by using the left and right images first, and according to the Knowing the parameters, the left and right image frames are projected to the same scene through image matching.
於本發明之一實施例中,該同一場景係為該系列之每張影像都經過同一點光學中心來拍攝,可任意角度取景,相鄰的影像需要適當之重疊。In an embodiment of the present invention, the same scene is that each image of the series is photographed through the same optical center, and the image can be viewed at an arbitrary angle, and adjacent images need to be appropriately overlapped.
於本發明之一實施例中,該同一場景係為扭曲或投影所輸入的影像到一個特殊之數學模型,然後從第一張扭曲過後的影像和第二張扭曲後影像直到最後一張影像計算梯度,根據這些參數來計算影像接合之參數。In an embodiment of the invention, the same scene is to distort or project the input image to a special mathematical model, and then calculate from the first warped image and the second warped image to the last image. Gradient, based on these parameters to calculate the parameters of the image joint.
於本發明之一實施例中,該全景編列數位影像單元係可調整第一與第二攝影機、第二與第三攝影機及第三與第一攝影機,進行全景拼接製作得到各攝影機所形成之局部全景。In an embodiment of the present invention, the panoramic digital video unit can adjust the first and second cameras, the second and third cameras, and the third and first cameras to perform panoramic stitching to obtain a part of each camera. panoramic.
於本發明之一實施例中,該全景編列數位影像單元係結合第一與第二攝影機、第二與第三攝影機及第三與第一攝影機所形成之局部全景,形成全域式三維重建全景影像畫面。In an embodiment of the present invention, the panoramic-array digital image unit combines the partial panoramas formed by the first and second cameras, the second and third cameras, and the third and first cameras to form a global three-dimensional reconstructed panoramic image. Picture.
於本發明之一實施例中,該全景編列數位影像單元係以三維重建後之全景影像畫面,量測剛體位移,以評估其基本量測誤差。In an embodiment of the present invention, the panoramic digital video unit is configured to measure the rigid body displacement by using a three-dimensionally reconstructed panoramic image frame to evaluate the basic measurement error.
於本發明之一實施例中,該三維重建單元係可將影像以具有不同景深(視差圖)之三維型態加以呈現。In an embodiment of the invention, the three-dimensional reconstruction unit can present images in a three-dimensional form having different depth of fields (disparity map).
請參閱『第2~第9圖』所示,係分別為係本發明之外觀示意圖、本發明影像擷取機與處理機構之方塊示意圖、本發明之使用狀態示意圖、本發明編列數位影像之三維應變量測系統示意圖、本發明攝影機校正參數量測之示意圖、本發明之攝影機模型與投影原理示意圖、本發明位於Π'平面上之校正點示意圖、本發明攝影機校正參數量測示意圖及本發明全景拼接製作之示意圖。如圖所示:本發明係一種編列數位影像關係裝置,其至少包含有一承載機構1、至少三個調整機構2、一影像擷取機構3、一校正板4以及一處理機構5所構成。Please refer to FIG. 2 to FIG. 9 for a schematic diagram of the appearance of the present invention, a block diagram of the image capturing machine and the processing mechanism of the present invention, a schematic diagram of the state of use of the present invention, and a three-dimensional image of the digital image of the present invention. a schematic view of the measurement system variables should schematic view camera correction parameter measurements of the present invention, the camera according to the present invention model projection principle diagram of the present invention positioned Π 'correction schematic points on the plane, the camera calibration parameters measuring a schematic view of the present invention and the panorama of the present invention Schematic diagram of splicing production. As shown in the figure, the present invention is a digital image correlation device comprising at least one carrying mechanism 1, at least three adjusting mechanisms 2, an image capturing mechanism 3, a correcting plate 4, and a processing mechanism 5.
上述所提之承載機構1係包含有一基座11、及一與基座11結合之量測桿12,該量測桿12係呈一中空之圓弧狀。The above-mentioned carrier mechanism 1 includes a base 11 and a measuring rod 12 coupled with the base 11, and the measuring rod 12 has a hollow arc shape.
各個調整機構2係活動設於量測桿12上,而各調整機構2係分別包含有一與量測桿12活動結合之滑座21、一設於滑座21上之調整座22、一設於滑座21上之平衡裝置23、及一設於滑座21上之電控裝置24,其中該滑座21之兩端係分別設有與量測桿12抵靠之滑輪211及靠緊輪212,而該靠緊輪212係連接有一調整旋鈕213,使滑座21藉由滑輪211及靠緊輪212於量測桿12上移動,並以調整旋鈕213配合靠緊輪212使滑座21加以定位,且利用平衡裝置23與電控裝置24可調整各攝影機凝視角度,使之具有共圓特性,讓調整機構帶動各攝影機其中心都保持指向圓心。Each of the adjustment mechanisms 2 is disposed on the measuring rod 12, and each of the adjusting mechanisms 2 includes a sliding seat 21 that is movably coupled to the measuring rod 12, an adjusting seat 22 disposed on the sliding seat 21, and a a balancing device 23 on the sliding seat 21 and an electric control device 24 disposed on the sliding seat 21, wherein the two ends of the sliding seat 21 are respectively provided with a pulley 211 and a pressing wheel 212 that abut against the measuring rod 12. And the adjusting wheel 213 is connected to the adjusting wheel 213, so that the sliding seat 21 moves on the measuring rod 12 by the pulley 211 and the abutting wheel 212, and the adjusting knob 213 cooperates with the abutting wheel 212 to make the sliding seat 21 Positioning, and using the balancing device 23 and the electronic control device 24, the gaze angle of each camera can be adjusted to have a co-circular characteristic, so that the adjusting mechanism drives the centers of the cameras to keep pointing toward the center of the circle.
該影像擷取機3構係包含有分別設於各調整裝置2所設調整座上之第一、二及第三攝影機31、32、33,且該第一、二及第三攝影機31、32、33係以第一及第二校正點301、302作為量測基準點,由於該量測桿12係呈一中空之圓弧狀,因此,可使第一、二及第三攝影機31、32、33之三點中心光軸共圓交於一點,且使第一、二及第三攝影機31、32、33可利用調整裝置2於量測桿12上進行適當間距之調整,其中該第一校正點301係設於基座11上,用於固定第一、二及第三攝影機31、32、33之光軸中心位於同一光軸點,做為量測初始點,而該第二校正點302係為所在環境之一參考點,且與第二攝影機32光軸中心及第一校正點301皆位於同一光軸線上,做為初始化量測使用,而該第二校正點302係配合有一標示桿303,可架設於遠方之待測物位置,以利第一、二及第三攝影機31、32、33進行後續重覆量測使用,該標示桿303上係具有可進行校正之標示刻度。The image capturing machine 3 includes first, second and third cameras 31, 32, 33 respectively arranged on the adjusting seats provided by the adjusting devices 2, and the first, second and third cameras 31, 32 The 33 series uses the first and second correction points 301 and 302 as the measurement reference points. Since the measuring rod 12 has a hollow arc shape, the first, second and third cameras 31 and 32 can be made. The center axis of the 33 points is rounded at a point, and the first, second and third cameras 31, 32, 33 can be adjusted by the adjustment device 2 on the measuring rod 12 by an appropriate spacing, wherein the first The calibration point 301 is disposed on the base 11 for fixing the centers of the optical axes of the first, second and third cameras 31, 32, 33 at the same optical axis point as the measurement initial point, and the second correction point The 302 is a reference point of the environment, and is located on the same optical axis as the optical axis center of the second camera 32 and the first calibration point 301, and is used for initial measurement, and the second calibration point 302 is matched with an indication. The rod 303 can be placed at a position of the object to be tested in the distance, so as to facilitate the first, second and third cameras 31, 32, 33 Repeated measurements using the lever 303 having a label based on the marking scale can be corrected.
該校正板4係設於第一、二及第三攝影機31、32、33之光軸上,而該校正板4係為具有固定間距之校正格點41中心座,作為第一、二及第三攝影機31、33、33校正時所需之校正點,且該校正板4之校正格點41可為圓形或方形,另當校正板4之校正格點41與第一、二及第三攝影機31、33、33之影像平面不平行時,所擷取影像之格點將形成扭曲形狀。The calibration plate 4 is disposed on the optical axes of the first, second, and third cameras 31, 32, and 33, and the calibration plate 4 is a center block of the correction grid 41 having a fixed pitch as the first, second, and The correction points required for the correction of the three cameras 31, 33, 33, and the correction grid 41 of the calibration plate 4 may be circular or square, and the correction grid 41 of the calibration plate 4 is first, second and third. When the image planes of the cameras 31, 33, and 33 are not parallel, the lattice points of the captured images will form a twisted shape.
該處理機構5係與第一、二及第三攝影機31、32、33連接,其係具有一編列數位影像運算模組51,可將第一、二及第三攝影機31、33、33所擷取之數位影像資料進行分析,而該處理機構5更進一步連接有一預警單元53,其中該編列數位影像運算模組51至少包含有相互校正參數單元511、全景拼接製作單元512、全景編列數位影像單元513及三維重建單元514,可使第一、二及第三攝影機31、32、33配合校正板4進行校正,並將各攝影機31、32、33所擷取之數位影像資料進行分析,且計算出位移及應變值。The processing mechanism 5 is connected to the first, second and third cameras 31, 32, 33, and has a digital image computing module 51, which can be used by the first, second and third cameras 31, 33, 33. The digital video data is analyzed, and the processing unit 5 is further connected to an early warning unit 53. The digital video computing module 51 includes at least a mutual correction parameter unit 511, a panoramic mosaic processing unit 512, and a panoramic digital video unit. 513 and 3D reconstruction unit 514, the first, second and third cameras 31, 32, 33 can be corrected with the correction plate 4, and the digital image data captured by each camera 31, 32, 33 can be analyzed and calculated. Displacement and strain values.
當使用時係以運用於戶外現場為例,可偵測物件於三維方向的變形,只需在戶外現場前架設承載機構1、三個調整機構2、影像擷取機構3之第一、二及第三攝影機31、32、33及校正板4,而利用處理機構5之編列數位影像運算模組51以自動化機器視覺檢測的概念將戶外現場當成待測樣本,主要負責的工作為連續取像及變形(位移)相關係數的演算,使處理機構5控制各第一、二及第三攝影機31、32、33及進行人眼輔助監視,透過網路傳述方式進行網路分佈,做即時影像的傳輸,可大面積即時應力檢測,並將取得的數位影像自動分別對高壓鐵塔、山脈走勢、建築物、基座及岩石分佈進行連續性的監測;在戶外現場變形或位移達到危殆臨界值時,即控制預警單元52發出警訊。For example, when used in an outdoor field, the deformation of the object in the three-dimensional direction can be detected, and only the first and second of the carrying mechanism 1, the three adjusting mechanisms 2, and the image capturing mechanism 3 are erected in front of the outdoor scene. The third camera 31, 32, 33 and the correction plate 4, and the digital image computing module 51 of the processing mechanism 5 uses the concept of automated machine vision detection to treat the outdoor scene as a sample to be tested, and the main responsible work is continuous image capturing and The calculation of the deformation (displacement) correlation coefficient causes the processing mechanism 5 to control the first, second and third cameras 31, 32, 33 and perform eye-assisted monitoring, and distribute the network through the network to perform real-time image transmission. It can detect large-area real-time stress and automatically monitor the digital image of the high-voltage tower, the mountain trend, the building, the pedestal and the rock distribution separately. When the outdoor field deformation or displacement reaches the critical threshold, The control early warning unit 52 issues a warning.
為增加處理速度,需簡化原始影像資料量,採部分特徵取樣方式以大幅度縮減計算時間,至於取樣特徵的選擇方面,本發明最主要的目的在於開發戶外現地特徵區域辨識,作為物件與物件的相對位移當為最主要的監測目標,針對此點可以高壓鐵塔、山脈走勢、建築物、基座及岩石分佈的相對接合為特徵,可取高壓鐵塔、山脈走勢、建築物、基座及岩石分佈之邊界作為演算的特徵資料來代表位於戶外現場的辨識特徵曲線,而特徵曲線可透過邊界搜尋理論自動得到;然,特徵曲線的意義是在數位影像陣列中被保留的位置資料,位在特徵曲線位置上的資料,可集合成資料簡化的陣列I(t),對I(t)特徵陣列連續進行交叉相關連演算(Cross-correlation),可取得即時特徵曲線陣列與原始特徵陣列之交叉相關連係數(Cross-correlation Coefficients),此係數可視為戶外現場危殆因子,其數值高低與戶外現場危險的程度呈高度的負相關。In order to increase the processing speed, the amount of original image data needs to be simplified, and some feature sampling methods are adopted to greatly reduce the calculation time. As for the selection of sampling features, the most important purpose of the present invention is to develop outdoor feature area identification as objects and objects. Relative displacement is the most important monitoring target. It can be characterized by high-pressure towers, mountain trends, relative joints of buildings, pedestals and rock distribution. High-voltage towers, mountain trends, buildings, pedestals and rock distributions are available. The boundary is used as the characteristic data of the calculus to represent the identification characteristic curve located in the outdoor scene, and the characteristic curve can be automatically obtained through the boundary search theory; however, the meaning of the characteristic curve is the position data retained in the digital image array, which is located at the characteristic curve position. The above data can be assembled into a simplified array of data I(t), and the cross-correlation of the I(t) feature array is continuously performed, and the cross-correlation coefficient between the immediate characteristic curve array and the original feature array can be obtained. (Cross-correlation Coefficients), this coefficient can be regarded as an outdoor site crisis Son, whose value level was highly negatively correlated with the degree of risk of the outdoor scene.
而當影像擷取機構3之第一、二及第三攝影機31、32、33擷取所需之數位影像資料後則可由處理機構5之編列數位影像運算模組51進行分析計算出位移及應變值,如第4圖所示,虛線部份代表第一、二及第三攝影機31、32、33光軸,影像區域R為擬觀測之物體,今依據下列規則以編列數位影像運算模組51之相互校正參數單元511、全景拼接製作單元512、全景編列數位影像單元513及三維重建單元514進行編列數位影像如下:依據本發明之第一、二及第三攝影機31、32、33之擺位架構,可改善習知兩部攝影機取像架構三維影像,分別以第一、二及第三攝影機31、32、33光軸垂直於校正試件,對應於第二攝影機32與第三攝影機33、第一攝影機31與第三攝影機33及第一攝影機31與第二攝影機32,可求出三組三維應變量測之內外部參數;而全景影像分別由攝影機第二攝影機32與第三攝影機33、第一攝影機31與第三攝影機33及第一攝影機31與第二攝影機32求出三組全景應變量測資料。When the first, second, and third cameras 31, 32, and 33 of the image capturing mechanism 3 capture the desired digital image data, the digital image computing module 51 of the processing mechanism 5 can analyze and calculate the displacement and strain. The value, as shown in Fig. 4, the dotted line represents the optical axes of the first, second and third cameras 31, 32, 33, and the image area R is the object to be observed. The digital image computing module 51 is now arranged according to the following rules. The mutual correction parameter unit 511, the panoramic mosaic unit 512, the panoramic array digital image unit 513, and the three-dimensional reconstruction unit 514 perform the digital image as follows: the first, second, and third cameras 31, 32, 33 are arranged according to the present invention. The architecture can improve the two-dimensional image capturing architecture of the two cameras, and the optical axes of the first, second and third cameras 31, 32, 33 are perpendicular to the calibration test piece, corresponding to the second camera 32 and the third camera 33, The first camera 31 and the third camera 33 and the first camera 31 and the second camera 32 can obtain three internal and external parameters of the three-dimensional strain measurement; and the panoramic images are respectively taken by the camera second camera 32 and the third camera The camera 33, the first camera 31 and the third camera 33, and the first camera 31 and the second camera 32 obtain three sets of panoramic strain measurement data.
如第5圖的戶外現地實驗架設示意圖,該第一、二及第三攝影機31、32、33之三點中心光軸共圓交於一點P,採用有固定間隔距離的校正格點41中心座標,作為第一、二及第三攝影機31、32、33校正時所需的校正點,基於戶外現場複雜地形之崎嶇不平及提高重覆量測精度雙重考量之下,分別設置第一校正點301及第二校正點302做為第一、二及第三攝影機31、32、33之量測基準點,第一校正點301用於固定第一、二及第三攝影機31、32、33與第二攝影機32光軸中心(P2)位於同一光軸點,量測初始點(0,0,0)做為量測初始點;第二校正點302做為戶外現場之一參考點(0,0,D),D為待測戶外現場距離第一、二及第三攝影機31、32、33一參考距離(D),且第一校正點301、第二攝影機32光軸中心(P2)及第二校正點302皆位於同一光軸線上,做為初始化量測使用。As shown in the outdoor outdoor experimental erection diagram of FIG. 5, the central optical axes of the three points of the first, second and third cameras 31, 32, 33 are rounded at a point P , and the center coordinates of the corrected grid 41 with a fixed separation distance are used. As the correction points required for the calibration of the first, second and third cameras 31, 32, 33, the first correction point 301 is respectively set under the double consideration of the ruggedness of the outdoor scene complex terrain and the improvement of the repeated measurement accuracy. And the second calibration point 302 serves as a measurement reference point for the first, second and third cameras 31, 32, 33, and the first correction point 301 is used for fixing the first, second and third cameras 31, 32, 33 and The camera 32 optical axis center (P2) is located at the same optical axis point, and the measurement initial point (0, 0, 0) is used as the measurement initial point; the second calibration point 302 is used as a reference point for the outdoor scene (0, 0) , D), D is the reference distance (D) of the first, second and third cameras 31, 32, 33 to be tested, and the first correction point 301, the second camera 32 optical axis center (P2) and the The two calibration points 302 are all located on the same optical axis and are used for initial measurement.
由於第一、二及第三攝影機31、32、33由鏡頭與影像平面所組成,影像平面在鏡頭中心C之前,而焦距則為影像平面至鏡頭中心的距離,物體上某點P(x,y,z)投影至影像平面上為Pc(uc,vc),此點即為C點與P點連線與影像平面之交點,由相似三角形的關係,可得攝影機座標(x,y,z)與影像座標(u 0,v 0)之關係式。Since the first, second and third cameras 31, 32, 33 are composed of a lens and an image plane, the image plane is before the lens center C, and the focal length is the distance from the image plane to the center of the lens, and a point P(x, on the object). y, z) projected onto the image plane as Pc(uc, vc), which is the intersection of the line connecting point C and point P with the image plane. The coordinates of similar triangles can be used to obtain the camera coordinates (x, y, z). ) The relationship between the image coordinates ( u 0 , v 0 ).
式中,(u 0,v 0)為影像平面中心點的座標位置。Where ( u 0 , v 0 ) is the coordinate position of the center point of the image plane.
如第6圖所示,O及O'分別為世界座標系統(world coordinate system)X-Y-Z及影像平面座標系統(image coordinate system)u-v的座標原點,C則為攝影機的鏡頭中心,同時也是第一、二及第三攝影機31、32、33座標系統(camera coordinate system)x-y-z的座標原點,影像座標系統的u軸與v軸分別平行於第一、二及第三攝影機31、32、33座標系統的x軸與y軸,其原點則位於影像平面的左上角。As shown in Fig. 6, O and O' are the coordinate origins of the world coordinate system XYZ and the image coordinate system uv, respectively, and C is the lens center of the camera, and is also the first. The second and third camera 31, 32, 33 coordinate coordinate system xyz coordinate origin, the u coordinate and the v axis of the image coordinate system are parallel to the first, second and third cameras 31, 32, 33 coordinates respectively The x-axis and y-axis of the system have their origin at the upper left corner of the image plane.
對空間中任意點P,經由旋轉矩陣R與平移矩陣t,即可建立世界座標PW(X,Y,Z)與攝影機座標Pca(x,y,z)的轉換關係式
對空間中任一世界座標點P(X,Y,Z),可由式(9)求得其投射於影像平面後的座標位置Pc(uc,vc)。然而在實際取像時,由於鏡頭扭曲的影響,投射點將沿著徑向與切線方向偏移。校正的影像座標(u c ,v c )與扭曲的影像座標(u d ,v d )之間的關係,可以下式表示。For any world coordinate point P(X, Y, Z) in space, the coordinate position Pc(uc, vc) projected on the image plane can be obtained by equation (9). However, in actual image taking, the projection point will be offset from the tangential direction in the radial direction due to the distortion of the lens. The relationship between the corrected image coordinates ( u c , v c ) and the distorted image coordinates ( u d , v d ) can be expressed by the following equation.
利用式(9)與(11),可將在三維世界座標系統下的點P(X,Y,Z),對映到扭曲的影像座標P d (u d ,v d )。然而,使用式(9)與(10),則可將影像平面上扭曲影像座標點(u d ,v d ),轉換到三維空間的投射線上的點,或者是此投射線與已知二維平面的交界點。Using the formula (9) and (11), it may be at the point of the three-dimensional world coordinate system P (X, Y, Z) , the enantiomer distorted image coordinates P d (u d, v d ). However, using equations (9) and (10), the distorted image coordinate points ( u d , v d ) on the image plane can be converted to points on the projection line of the three-dimensional space, or the projection line and the known two-dimensional The junction point of the plane.
使用校正格點的中心為校正點,然而校正點投射到影像平面時,可能不再保持原始圖形,而成為變形圖形。以下將敘述如何求取此變形圖形的中心,以便將三維中間的校正點中心,對映到二維影像平面上。The center of the correction grid is used as the correction point. However, when the correction point is projected onto the image plane, the original image may no longer be retained and become a deformed pattern. The following describes how to find the center of the deformed pattern so that the center of the correction point in the middle of the three-dimensional image is mapped onto the two-dimensional image plane.
而第7圖係顯示一個位於Π'平面上的校正格點R,此校正點之半徑為r,且其校正點中心h0=(X0,Y0,Z0)T係位於平面座標系統x h -y h 的座標原點。在世界座標系統X-Y-Z下,平面座標系統x h -y h 的原點座標為h0=(X0,Y0,Z0)T,而座標軸x h 及y h 的方向餘弦分別為h1=(X1,Y1,Z1)T與h2=(X2,Y2,Z2)T。在x h -y h 座標系統下,此校正點之方程式可表示為
式(12)表示的是一個圓錐曲線的方程式。實際上,校正點投射到影像平面上會是扭曲的校正點,或是在某些特殊範例上是一個圓。使用式(9)與(12),圓點R投影於影像平面的曲線方程式可表示為
在平面座標系統x h -y h 下,校正點中心(u e ,v e )可以求得下式。Under the plane coordinate system x h - y h , the center of the correction point ( u e , v e ) can be obtained as follows.
最後,可使用下式將此控制校正點於影像平面上的投影點座標單位由mm轉換成pixel。Finally, the coordinates of the projection point of the control correction point on the image plane can be converted from mm to pixel using the following formula.
u=M x u d ,v=M y v d (16) u = M x u d , v = M y v d (16)
式中,M x 與M y 分別為在u與v軸方向的比例係數,其單位為pixel/mm,可將實際物理單位元mm轉換為影像單位元pixel。In the formula, M x and M y are proportional coefficients in the u and v axis directions, respectively, and the unit is pixel/mm, and the actual physical unit element mm can be converted into the image unit pixel.
透過以上的轉換,可得到在世界座標中任意控制圓點中心,在影像平面上的座標位置,由上述轉換,可知共有15個攝影機參數,其中M x ,M y ,f,k 1,k 2,p 1,p 2以及(u 0,v 0)為攝影機的內部參數。而(t x ,t y ,t z )與(θ,Φ,φ)則為攝影機的外部參數。Through the above conversion, the coordinate position on the image plane can be obtained at any center of the world coordinates. From the above conversion, there are 15 camera parameters, among which M x , M y , f , k 1 , k 2 , p 1 , p 2 and ( u 0 , v 0 ) are internal parameters of the camera. And ( t x , t y , t z ) and (θ, Φ, φ) are external parameters of the camera.
利用上式(14)~(16),可將一個在三維世界座標下的控制圓點中心,對映到二維影像平面上,可將三維的場景投影到二維影像平面上。然而,此對映的轉換過程係屬於多對一的轉換,如果只使用一部攝影機時,即無法獲取深度的相關資訊。因此,本發明採用一部攝影機光軸垂直於校正試件,求取校正試件上各格點的座標位置,以建立世界座標系統以及一組校正點的座標位置。再使用此組校正點,校正其他兩部攝影機構成之立體視覺系統,各旋轉一個角度,並同時對試件表面的同一區域取像,利用這兩張影像中的相關點,以求取深度資料。Using the above formulas (14)~(16), a control dot center under the three-dimensional world coordinates can be mapped onto the two-dimensional image plane, and the three-dimensional scene can be projected onto the two-dimensional image plane. However, the conversion process of this mapping is a many-to-one conversion. If only one camera is used, the relevant information of the depth cannot be obtained. Therefore, the present invention uses a camera optical axis perpendicular to the calibration test piece to obtain the coordinate position of each grid point on the calibration test piece to establish a coordinate position of the world coordinate system and a set of correction points. Then use this set of calibration points to correct the stereo vision system composed of the other two cameras, each rotating one angle, and simultaneously taking the same area on the surface of the test piece, and using the relevant points in the two images to obtain the depth data. .
若已知空間中M個控制圓點中心的世界座標位置,以及此M個點在攝影機之影像平面上相關的影像座標(U j ,V j ),j=1,…,M,即可以非線性最小平方法,求得攝影機的最佳15個攝影機參數,使得經由攝影機模型方程式求得的影像對座標(u j ,v j )與實驗值(U j ,V j )之間的誤差為最小,亦即使下式之值為最小值。If the world coordinate position of the M control dot centers in the space is known, and the image coordinates ( U j , V j ), j=1,...,M of the M points on the image plane of the camera are known, The linear least squares method is used to obtain the best 15 camera parameters of the camera, so that the error between the coordinate pair coordinates ( u j , v j ) and the experimental values ( U j , V j ) obtained by the camera model equation is minimized. , even if the value of the following formula is the minimum value.
而第8圖係為攝影機校正參數量測示意圖,以第一攝影機31的光軸垂直於校正板4,利用相互校正參數單元511求取校正板4上各校正格點41的座標位置,以建立世界座標系統以及一組校正點的座標位置;之後再使用此組校正點,校正第二攝影機31及第三攝影機33構成之立體視覺系統,其餘編列數位影像關法選用方式亦同;並以第二攝影機32的光軸垂直於校正板4,利用相互校正參數單元511求取校正板4上各校正格點41的座標位置,以建立世界座標系統以及一組校正點的座標位置,再使用此組校正點,校正第一攝影機31及第三攝影機構成之立體視覺系統;再以第三攝影機33的光軸垂直於校正板4,利用相互校正參數單元511求取校正板4上各校正格點41的座標位置,以建立世界座標系統以及一組校正點的座標位置,再使用此組校正點,校正第一攝影機31及第二攝影機32構成之立體視覺系統。8 is a schematic diagram of the camera calibration parameter measurement. The optical axis of the first camera 31 is perpendicular to the calibration plate 4, and the mutual calibration parameter unit 511 is used to obtain the coordinate position of each correction grid 41 on the calibration plate 4 to establish The coordinate system of the world coordinate system and a set of calibration points; then the correction point is used to correct the stereo vision system formed by the second camera 31 and the third camera 33, and the remaining digital image selection methods are also the same; The optical axis of the second camera 32 is perpendicular to the correction plate 4, and the coordinate position of each correction grid 41 on the correction plate 4 is obtained by the mutual correction parameter unit 511 to establish a coordinate position of the world coordinate system and a set of correction points, and then use this. The correction point is corrected to correct the stereo vision system formed by the first camera 31 and the third camera; and the optical axis of the third camera 33 is perpendicular to the correction plate 4, and the correction parameter unit 511 is used to obtain each correction grid on the correction plate 4. The coordinate position of 41 to establish a coordinate position of the world coordinate system and a set of correction points, and then use the set of correction points to correct the first camera 31 and the second camera 32 Into the stereo vision system.
利用校正格點41之中心位置為校正點,且格點投射於影像平面上,建立一世界座標系統以及M個已知世界座標位置之校正點,利用此M個校正點以及攝影機模型方程式,執行立體視覺系統的攝影機校正,以求得各攝影機M x ,M y ,f,k 1,k 2,p 1,p 2與(u 0,v 0)之攝影機內部參數以及(t x ,t y ,t z )與(θ,Φ,φ)之攝影機外部參數。並使用校正格點41中心與影像平面格點中心的相關數學式,以減少第一、二及第三攝影機31、32、33的偏差,此外亦考慮第一、二及第三攝影機31、32、33之徑向與圓周向鏡頭扭曲係數,以增進第一、二及第三攝影機31、32、33的精確度。Using the center position of the correction grid 41 as a correction point, and the grid point is projected on the image plane, a world coordinate system and correction points of M known world coordinate positions are established, and the M correction points and the camera model equation are used to execute Camera correction of the stereo vision system to obtain camera internal parameters of each camera M x , M y , f , k 1 , k 2 , p 1 , p 2 and ( u 0 , v 0 ) and ( t x , t y , t z ) and (θ, Φ, φ) camera external parameters. The mathematical formula of the center of the correction grid 41 and the center of the image plane grid is used to reduce the deviation of the first, second and third cameras 31, 32, 33, and the first, second and third cameras 31, 32 are also considered. The radial and circumferential lens distortion coefficients of 33 are used to improve the accuracy of the first, second and third cameras 31, 32, 33.
以下列方法進行3D視覺系統的校正以及量測誤差評估方式(以校正板4以及第一攝影機31的光軸垂直於校正試件為例,其餘第二攝影機32及第三攝影機33編列數位影像之方式亦同):The correction of the 3D vision system and the measurement error evaluation method are performed in the following manner (the correction plate 4 and the optical axis of the first camera 31 are perpendicular to the calibration test piece, and the remaining second camera 32 and the third camera 33 are arranged for the digital image. The same way):
(1)調整第一攝影機31,使其光軸垂直於校正板4之校正格點4,並使N×N格點陣列在可視範圍內。(1) The first camera 31 is adjusted such that its optical axis is perpendicular to the correction grid 4 of the correction plate 4, and the array of N x N grids is within the visible range.
(2)取得第一攝影機31水平及垂直方向誤差值,設定X=D及X=-D前、後移動,以第一攝影機31求取校正點參數,並以第一攝影機31所攝取影像的水平與垂直方向的Scale Factor(mm/pixel),並產生一世界座標,以及在此座標系統下的N×N個格點座標位置。(2) Obtaining the horizontal and vertical direction error values of the first camera 31, setting X=D and X=-D before and after, moving the first camera 31 to obtain the correction point parameter, and taking the image taken by the first camera 31. Scale Factor (mm/pixel) in horizontal and vertical directions, and produce a world coordinate, and N × N grid coordinates below the coordinate system.
(3)調整第二攝影機32與第三攝影機33,使其觀察範圍與第一攝影機31相同,並配合三維重建單元514校正第二攝影機32與第三攝影機33所形成的三維重建。(3) The second camera 32 and the third camera 33 are adjusted so that the observation range is the same as that of the first camera 31, and the three-dimensional reconstruction formed by the second camera 32 and the third camera 33 is corrected in accordance with the three-dimensional reconstruction unit 514.
(4)以校正後的三維重建,量測剛體位移,以評估其基本量測誤差。(4) Measure the rigid body displacement with the corrected three-dimensional reconstruction to evaluate its basic measurement error.
而當全景拼接製作單元512於進行全景拼接製作時,主要分為兩個步驟(如第9圖所示):When the panoramic splicing and making unit 512 is used for panorama splicing, it is mainly divided into two steps (as shown in FIG. 9):
(1)先藉由第一、二及第三攝影機31、32、33同時拍攝左右兩張影像,先行校正求出鏡頭的參數。(1) First, the left and right images are simultaneously captured by the first, second, and third cameras 31, 32, and 33, and the parameters of the lens are first corrected.
(2)根據求得的參數,將左右兩張影像畫面經過影像匹配的過程投影至同一場景。(2) Projecting the left and right image frames through the image matching process to the same scene according to the obtained parameters.
另亦可輸入一系列的影像,並將輸入之立體影像接合成一全景影像接合,而該系列的每張影像都經過同一點光學中心來拍攝,可任意角度取景,相鄰的影像需要適當的重疊;首先扭曲或投影所輸入的影像到一個特殊的數學模型,然後從第一張扭曲過後的影像和第二張扭曲後影像直到最後一張影像計算梯度,根據這些參數來計算影像接合的參數,找出一組可能的對應解。最後,可以決定最佳的接合點而把相鄰的兩張影像拼接完成具有深度資訊的全景拼接影像。In addition, a series of images can be input, and the input stereoscopic images are joined into a panoramic image, and each image of the series is photographed through the same optical center, and can be viewed at any angle, and adjacent images need appropriate Overlapping; first distorting or projecting the input image to a special mathematical model, then calculating the gradient from the first distorted image and the second distorted image to the last image, and calculating the parameters of the image joint based on these parameters , find a set of possible corresponding solutions. Finally, the best joints can be determined and the adjacent two images can be stitched together to complete the panoramic stitched image with depth information.
再者,當全景編列數位影像單元513於進行編列數位影像之拼接製作時,主要分為兩個步驟:Furthermore, when the panoramic video unit 513 is used for splicing digital images, it is mainly divided into two steps:
(1)先藉由第一、二及第三攝影機31、32、33所擷取之左右影像,先行校正求出第一、二及第三攝影機31、32、33的參數。(1) The parameters of the first, second, and third cameras 31, 32, and 33 are first corrected by the left and right images captured by the first, second, and third cameras 31, 32, and 33.
(2)根據求得的參數,將兩張左右影像畫面經過影像匹配的過程投影至同一場景。(2) Projecting the two left and right image frames through the image matching process to the same scene according to the obtained parameters.
另亦可輸入一系列的影像,並將輸入之立體影像接合成一全景影像接合,該系列的每張影像都經過同一點光學中心來拍攝,可任意角度取景,相鄰的影像需要適當的重疊;首先,扭曲或投影所輸入的影像到一個特殊的數學模型,然後從第一張扭曲過後的影像和第二張扭曲後影像直到最後一張影像計算梯度,根據這些參數來計算影像接合的參數,找出一組可能的對應解。In addition, a series of images can be input, and the input stereoscopic images are joined into a panoramic image. Each image of the series is photographed through the same optical center, and can be viewed at any angle, and adjacent images need appropriate overlap. First, the image input is distorted or projected to a special mathematical model, and then the gradient is calculated from the first warped image and the second warped image until the last image, and the parameters of the image joint are calculated based on these parameters. , find a set of possible corresponding solutions.
以下列方法進行編列數位影像關法之拼接製作步驟:The following steps are used to arrange the splicing steps of the digital image separation method:
(1)調整攝影機N1-N2、N2-N3及N3-N1,進行全景拼接製作得到攝影機N1-N2、N2-N3及N3-N1所形成的局部全景。(1) Adjusting the cameras N1-N2, N2-N3, and N3-N1, and performing panoramic stitching to obtain a partial panorama formed by the cameras N1-N2, N2-N3, and N3-N1.
(2)結合攝影機N1-N2、N2-N3及N3-N1所形成的局部全景,形成的全域式三維重建全景影像畫面。(2) A global panoramic three-dimensional reconstruction panoramic image is formed by combining the partial panorama formed by the cameras N1-N2, N2-N3 and N3-N1.
(3)以三維重建後之全景影像畫面,量測剛體位移,以評估其基本量測誤差。(3) The three-dimensional reconstructed panoramic image is used to measure the displacement of the rigid body to evaluate its basic measurement error.
最後,可以決定最佳的接合點而把相鄰的兩張影像拼接完成具有深度資訊的全景拼接影像;經由以上規則編列後,可以獲得大面積之三維應變量,物件與物件的相對位移當為最主要的監測目標,可使本發明以第一、二及第三攝影機31、32、33進行測量,而具備全域、遠距與便利等優點,可用來進行快速之位移量測,以攝影機量測得到之位移場進行應變分析,計算求得結構在外力作用的應變反應;且至少達到下列之優點:Finally, the optimal joint can be determined and the adjacent two images can be stitched together to complete the panoramic stitched image with depth information; after the above rules are arranged, a large-area three-dimensional strain can be obtained, and the relative displacement of the object and the object is The most important monitoring target enables the present invention to be measured by the first, second and third cameras 31, 32, 33, and has the advantages of global, remote and convenient, and can be used for rapid displacement measurement, with camera volume. The measured displacement field is subjected to strain analysis, and the strain response of the structure acting on the external force is calculated; and at least the following advantages are achieved:
1.實驗儀器及試驗步驟簡單、可行性及準確度高,且實驗裝置具有平衡裝置確保量測結果穩定。1. The experimental instrument and test procedure are simple, feasible and accurate, and the experimental device has a balance device to ensure stable measurement results.
2.量測範圍可任意選取,大面積量測。2. The measurement range can be arbitrarily selected and measured in a large area.
3.藉由多部攝影機內外部參數互相比對,能正確估算位移及應變參數。3. By comparing the internal and external parameters of multiple cameras, the displacement and strain parameters can be correctly estimated.
4.所建立之自動化計算程式,快速將受力前後之數位影像資料進行分析,並計算出位移及應變值。4. The established automatic calculation program quickly analyzes the digital image data before and after the force and calculates the displacement and strain values.
5.攝影機所監測區域受力變形之影像紀錄可永久儲存作為其他方面研究之參考。5. The image record of the force deformation of the area monitored by the camera can be permanently stored as a reference for other research.
綜上所述,本發明編列數位影像關係裝置可有效改善習用之種種缺點;進而使本發明之產生能更進步、更實用、更符合消費者使用之所須,確已符合發明專利申請之要件,爰依法提出專利申請。In summary, the present invention provides a digital image correlation device that can effectively improve various disadvantages of the conventional use; thereby, the invention can be made more progressive, more practical, and more suitable for consumer use, and indeed meets the requirements of the invention patent application. , 提出 file a patent application according to law.
惟以上所述者,僅為本發明之較佳實施例而已,當不能以此限定本發明實施之範圍;故,凡依本發明申請專利範圍及發明說明書內容所作之簡單的等效變化與修飾,皆應仍屬本發明專利涵蓋之範圍內。However, the above is only the preferred embodiment of the present invention, and the scope of the present invention is not limited thereto; therefore, the simple equivalent changes and modifications made in accordance with the scope of the present invention and the contents of the invention are modified. All should remain within the scope of the invention patent.
10...支架10. . . support
11、12...攝影機11,12. . . camera
1...承載機構1. . . Carrying mechanism
11...基座11. . . Pedestal
12...量測桿12. . . Measuring rod
2...調整機構2. . . Adjustment mechanism
21...滑座twenty one. . . Slide
211...滑輪211. . . pulley
212...靠緊輪212. . . Tight wheel
213...調整旋鈕213. . . Adjustment knob
22...調整座twenty two. . . Adjustment seat
23...平衡裝置twenty three. . . Balance device
24...電控裝置twenty four. . . Electric control device
3...影像擷取機3. . . Image capture machine
31...第一攝影機31. . . First camera
32...第二攝影機32. . . Second camera
33...第三攝影機33. . . Third camera
301...第一校正點301. . . First correction point
302...第二校正點302. . . Second correction point
303...標示桿303. . . Marking rod
4...校正板4. . . Calibration board
41...校正格點41. . . Correction grid
5...處理機構5. . . Processing organization
51...編列數位影像運算模組51. . . Digital image computing module
511...相互校正參數單元511. . . Mutual correction parameter unit
512...全景拼接製作單元512. . . Panorama stitching unit
513...全景編列數位影像單元513. . . Panoramic digital video unit
514...三維重建單元514. . . 3D reconstruction unit
52...預警單元52. . . Early warning unit
第1圖,係習用之影像量測結構示意圖。Figure 1 is a schematic diagram of a conventional image measurement structure.
第2圖,係本發明之外觀示意圖。Fig. 2 is a schematic view showing the appearance of the present invention.
第3圖,係本發明影像擷取機與處理機構之方塊示意圖。Figure 3 is a block diagram of the image capturing machine and processing mechanism of the present invention.
第4圖,係本發明編列數位影像之三維應變量測系統示意圖。Fig. 4 is a schematic diagram showing a three-dimensional strain measurement system for digital images according to the present invention.
第5圖,係本發明攝影機校正參數量測之示意圖。Fig. 5 is a schematic view showing the calibration of the calibration parameters of the camera of the present invention.
第6圖,係本發明之攝影機模型與投影原理示意圖。Figure 6 is a schematic diagram of the camera model and projection principle of the present invention.
第7圖,係本發明位於Π'平面上之校正點示意圖。Figure 7 is a schematic view of the calibration point of the present invention on the plane of Π ' .
第8圖,係本發明攝影機校正參數量測示意圖。Figure 8 is a schematic diagram showing the calibration of the camera calibration parameters of the present invention.
第9圖,係本發明全景拼接製作之示意圖。Figure 9 is a schematic view showing the panorama stitching of the present invention.
1...承載機構1. . . Carrying mechanism
11...基座11. . . Pedestal
12...量測桿12. . . Measuring rod
2...調整機構2. . . Adjustment mechanism
21...滑座twenty one. . . Slide
211...滑輪211. . . pulley
212...靠緊輪212. . . Tight wheel
213...調整旋鈕213. . . Adjustment knob
22...調整座twenty two. . . Adjustment seat
23...平衡裝置twenty three. . . Balance device
24...電控裝置twenty four. . . Electric control device
3...影像擷取機3. . . Image capture machine
31...第一攝影機31. . . First camera
32...第二攝影機32. . . Second camera
33...第三攝影機33. . . Third camera
301...第一校正點301. . . First correction point
302...第二校正點302. . . Second correction point
303...標示桿303. . . Marking rod
4...校正板4. . . Calibration board
41...校正格點41. . . Correction grid
5...處理機構5. . . Processing organization
51...編列數位影像運算模組51. . . Digital image computing module
52...預警單元52. . . Early warning unit
Claims (28)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW100125747A TW201310004A (en) | 2011-08-18 | 2011-08-18 | Correlation arrangement device of digital images |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW100125747A TW201310004A (en) | 2011-08-18 | 2011-08-18 | Correlation arrangement device of digital images |
Publications (2)
Publication Number | Publication Date |
---|---|
TW201310004A true TW201310004A (en) | 2013-03-01 |
TWI428562B TWI428562B (en) | 2014-03-01 |
Family
ID=48481928
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW100125747A TW201310004A (en) | 2011-08-18 | 2011-08-18 | Correlation arrangement device of digital images |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW201310004A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104811680A (en) * | 2014-01-28 | 2015-07-29 | 聚晶半导体股份有限公司 | Image obtaining device and image deformation correction method thereof |
CN104811604A (en) * | 2014-01-28 | 2015-07-29 | 聚晶半导体股份有限公司 | Image obtaining device and image deformation correction method thereof |
TWI511081B (en) * | 2014-01-28 | 2015-12-01 | Altek Semiconductor Corp | Image capturing device and method for calibrating image deformation thereof |
TWI548281B (en) * | 2014-11-21 | 2016-09-01 | 財團法人工業技術研究院 | Wide view monitoring system and method thereof |
TWI564647B (en) * | 2015-03-27 | 2017-01-01 | 國立臺北科技大學 | Method of image conversion operation for panorama dynamic ip camera |
TWI579776B (en) * | 2015-10-27 | 2017-04-21 | 財團法人國家實驗研究院 | Method for determinating full scene image by using multiple image acquisiiton devices |
TWI626603B (en) * | 2017-10-24 | 2018-06-11 | 鴻海精密工業股份有限公司 | Method and device for obtaining images |
CN113834433A (en) * | 2020-06-24 | 2021-12-24 | 财团法人工业技术研究院 | Compression state measuring method and compression state measuring system |
TWI777361B (en) * | 2020-01-20 | 2022-09-11 | 新加坡商視覺技術創投私人有限公司 | 3D shooting device, 3D shooting method, 3D display terminal, computer readable storage medium and computer program product |
-
2011
- 2011-08-18 TW TW100125747A patent/TW201310004A/en not_active IP Right Cessation
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104811680A (en) * | 2014-01-28 | 2015-07-29 | 聚晶半导体股份有限公司 | Image obtaining device and image deformation correction method thereof |
CN104811604A (en) * | 2014-01-28 | 2015-07-29 | 聚晶半导体股份有限公司 | Image obtaining device and image deformation correction method thereof |
TWI511081B (en) * | 2014-01-28 | 2015-12-01 | Altek Semiconductor Corp | Image capturing device and method for calibrating image deformation thereof |
US9792685B2 (en) | 2014-01-28 | 2017-10-17 | Altek Semiconductor Corp. | Image capturing device and method for calibrating image deformation thereof |
CN104811604B (en) * | 2014-01-28 | 2018-03-20 | 聚晶半导体股份有限公司 | Image acquiring device and its image deformation bearing calibration |
TWI548281B (en) * | 2014-11-21 | 2016-09-01 | 財團法人工業技術研究院 | Wide view monitoring system and method thereof |
TWI564647B (en) * | 2015-03-27 | 2017-01-01 | 國立臺北科技大學 | Method of image conversion operation for panorama dynamic ip camera |
TWI579776B (en) * | 2015-10-27 | 2017-04-21 | 財團法人國家實驗研究院 | Method for determinating full scene image by using multiple image acquisiiton devices |
TWI626603B (en) * | 2017-10-24 | 2018-06-11 | 鴻海精密工業股份有限公司 | Method and device for obtaining images |
TWI777361B (en) * | 2020-01-20 | 2022-09-11 | 新加坡商視覺技術創投私人有限公司 | 3D shooting device, 3D shooting method, 3D display terminal, computer readable storage medium and computer program product |
CN113834433A (en) * | 2020-06-24 | 2021-12-24 | 财团法人工业技术研究院 | Compression state measuring method and compression state measuring system |
Also Published As
Publication number | Publication date |
---|---|
TWI428562B (en) | 2014-03-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8803943B2 (en) | Formation apparatus using digital image correlation | |
TW201310004A (en) | Correlation arrangement device of digital images | |
Tang et al. | Real-time detection of surface deformation and strain in recycled aggregate concrete-filled steel tubular columns via four-ocular vision | |
US9965870B2 (en) | Camera calibration method using a calibration target | |
Chen et al. | Full-field 3D measurement using multi-camera digital image correlation system | |
Luhmann | Close range photogrammetry for industrial applications | |
CN102003938B (en) | Thermal state on-site detection method for large high-temperature forging | |
CN102032878B (en) | Accurate on-line measurement method based on binocular stereo vision measurement system | |
CA2553477C (en) | Transprojection of geometry data | |
CN109163657B (en) | Round target pose detection method based on binocular vision three-dimensional reconstruction | |
CN112132906B (en) | External parameter calibration method and system between depth camera and visible light camera | |
CN106840011A (en) | Steel tower deformation measuring device and its method | |
CN103983186B (en) | Binocular vision system bearing calibration and calibration equipment | |
Zhou et al. | A novel way of understanding for calibrating stereo vision sensor constructed by a single camera and mirrors | |
Yang et al. | Flexible and accurate implementation of a binocular structured light system | |
CN107328502A (en) | A kind of anchor rod pallet board load visualization digital imaging method | |
TWI526670B (en) | Device and method for measuring three-dimensional images of tunnel deformation | |
CN115272080A (en) | Global deformation measurement method and system based on image stitching | |
CN115540775A (en) | 3D video extensometer of CCD single-phase machine | |
Franco et al. | RGB-D-DIC technique for low-cost 3D displacement fields measurements | |
CN205300519U (en) | Iron tower warp measuring device | |
Gao et al. | Full‐field deformation measurement by videogrammetry using self‐adaptive window matching | |
CN114663486A (en) | Building height measurement method and system based on binocular vision | |
Ge et al. | Calibration of ring multicamera system with transparent target for panoramic measurement | |
Zhou et al. | Optical-path-difference analysis and compensation for asymmetric binocular catadioptric vision measurement |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
MM4A | Annulment or lapse of patent due to non-payment of fees |