201224392 六、發明說明： 【發明所屬之技術領域】 本發明是有關於一種對測量對象物的形狀進行測量的 裝置及方法，尤其是有關於一種即使測量對象物的背面侧 的形狀亦高精度地測量的裝置及方法。 【先前技術】 業界期待於利用格子投影法的非接觸形狀測量裝置、 利用干涉條!文分析的非接觸應力·應變測量裳置^ 數位全像術的位移測量裝置等中，將改 件 =多個條纹圖像測量資料加以合成，藉此獲得 但是，於先前的條紋圖像分析方法中，例如 已每個像素進行 有如=該前的技術，於專利文獻1中揭示 處理,計算彼此:二區域來進行 分析方法卜存在無法針對每個=進=成=前的 -般認為藉由不僅將資料加：仃：成的問碭。 素決定評價值並採用評價值高的資料 母個像 $的:為評價圖像的先前技術利文獻以好^结 下的方法:使圖像剖_農淡分布通過;==: 201224392 頻率的=敏度的特性的帶通m，根_帶通遽波器的 輸出的農淡分布而算出表示圖像品質的評價值。該先前的 圖像評價方法具有可_評價絲評價並表錄位圖像的 圖像=質的特徵。但是，該先前的圖像評價方法存在如下 的問題·n用於由相機所拍攝的數位圖像的圖像品質的 評價’無法評價於條紋圖像測量中進行了分析的相位資料。 因此’於專利文獻3中提出有如下的技術：對經相移 的條賴像資料的各個進行相位分析*獲得多個相位 如布資料後根據與多個相位分布資料的各個相關的條紋 圖像負料來對每個像素蚊值，與每個像素相對 應的#價錄料，n此可將乡_齡布㈣高精度地 合成。 [先前技術文獻] [專利文獻] [專利文獻1]日本專利特開細丨捕㈣號說明書 [專利文獻2]日本專利特開平π·325922號說明書 [專利文獻3]日本專利第3837565號公報 然而近年來’存在對測量對象物的整麵面的形狀 進期望並非僅對測量對象物的， 進打測量，._輯象物的整周的雜進行 但是1於專利文獻3所記載的方法中，由於自一個方 向拍攝測篁對象物，因此在無法掛不造相 量對象物的背面側測定形狀方面留τ課題。、視场的測 【發明内容】 201224392 因此’本發明的目的在於提供一種即使不進入一台相 機的視場的測量對象物的背面_形狀亦高精度地測量的 裝置及形狀測量方法。 發明者等人對解決上述課題的方法進行努力研究的結 ，發現有效的是於測量對象物的周圍配置鏡子，並將相 機配置成可同時拍制量對象物的圖像與映在鏡子中的鏡 像=圖像，賴拍攝測量對象物及映在鏡子中的鏡像，從 而完成本發明。 …=本發明的形狀測量裝置是對測量對象物的形狀進 2置的裝置，其特徵在於：其包括格子圖案投影部將 格子圖案投影於上刺量縣物上；至少—個鏡子，配置 測量對象物的周圍；至少—個攝影部拍攝上述測 =象物及映在上述鏡子中的上述測量對象物的鏡像；以 :析處理σ卩’對所拍攝的上述測量對象物及上述鏡像的 ^象的各個實施相位分析處理，算出上述測量對象物的形 ^並且將所拍攝的上述測量對象物的圖像與上述鏡像的 ^、、加以合成，且將上述攝影部的各個配置成可同時拍攝 述測量對象物的至少—部分的區域與該至少—部分的區 场的鏡像。 另外’本發明的形狀測量裝置是對測量對象物的形狀 量的裝置其特徵在於：其包括格子®案投影部， 里格子圖案投影於上述測量對象物上；至 =上述測量對象物的周圍；以及至少一個攝影部，拍： 地測置對象物及映在上述鏡子巾的上述測量對象物的鏡 201224392 像；且將上述攝影部的各她置成可同時㈣上述測量對 象物的至部分的區域與該至少—部分的區域的鏡像。 另外’如本發明的形狀測量裝置，其中對上述鏡子的 各個設置雜散光防止壁。 另外’如本發明的形狀測量裝置，其具備兩個鏡子與 兩個攝影部。 另外’如本發__測量裝置，其具備兩個鏡子盥 一個攝影部。 另外，本發明的形狀測量方法是對周圍配置有至少一 :鏡:的測量對象物的形狀進行測量的方法，其特徵在 包括格子圖案投影步驟，將格子圖案投影於上述測 ^物上，圖像拍攝步驟，同時拍攝上制量對象物及 、在上述鏡子巾的上述測量對象⑽麟；祕步驟，僅 格子圖案移動規定的大小；以及形狀算出步驟，對 ，上述實像及上述鏡像的圖像的各個實施相位分析 3 ’算出上述測量對象物的形狀；且該形狀算出步驟是 述圖像拍攝步驟與上述轉步驟僅重複狀的次數 设進行。 處理本發明的形狀測量方法，其中上述相位分析 疋，由如下方式而進行：根據全空間表格化方法事 十對每個像素製作使相位值與”絲產生關聯的表 值求=3^格，根據由相移法所求出的各像素的相位 另外，如本發明的形狀測量方法，其更包括根據與所 201224392 拍攝的®I像的鮮成分的比相對應的評價值，將上述測量 對象物的圖像與上述鏡像的圖像合成為—_像的合成處 理步驟。 [發明的效果] 根據本發明，藉由將鏡子配置於測量對象物的周圍， 可同時自錢方向制量縣物的形狀進行測量，因此可 測定於自-财向_射無法獲得的測量對象物面 侧的形狀。 另外，藉由將利用多個相機所拍攝的圖像加以合成， 即使於測量對象物巾含有金制情況下、或者測量^象物 的形狀具有曲面的情況下，亦可降低光暈的影響而高精度 地算出測量對象物的形狀。 為讓本發明之上述和其他目的、特徵和優點能更明顯 易懂，下文特舉較佳實施例，並配合所附圖式，作詳細說 明如下。 【實施方式】 [形狀測量裝置] 以下，參照圖式來對本發明的實例進行說明。 本發明的形狀測量裝置包括：格子圖案投影部，將格 子圖案投影於測量對象物上；至少一個鏡子，配置於測量 對象物的周圍；至少一個攝影部，拍攝測量對象物及映在 鏡子中的測里對象物的鏡像；以及分析處理部，對測量對 象物的圖像及鏡像的圖像的各個實施相位分析處理，算出 測量對象物的形狀，並且將所拍攝的上述測量對象物的圖 201224392 像與上述鏡像的圖像加以合成。此處，將各個攝影部配置 成可同時拍攝測量對象物的至少一部分的區域與該至少一 部分的區域的鏡像較重要。 圖1是表示本發明的一實例的形狀測量裝置的圖。該 形狀測量裝置1〇〇包括：z軸用光源1、z軸用格子玻璃2、 第1透鏡3、壓電平台(piezo stage)4、帕耳帖(Peltier)冷卻 裝置5第1相機6及第2相機7、基準點投光用雷射器.8、 第1鏡子9及第2鏡子10、基準平板Η、xy軸用光源、 =轴用格子玻璃13、z轴移動平台14、以及分析處理部 準平^形狀測量裝置_中，將測量對象物0載置於基 對象，且於德準平板11的周圍配置有映照測量 十象物0的兩面鏡子（第1鏡子9及第2鏡子1〇)。 格子準平板11的上方，設置有ζ軸用光源卜Ζ轴用 格子坡蹲2、fl透鏡3 釉用 圖案投影於__ 〇上的將格子 車由用光源賴射方向朝向_2方的向格子圖案投影部。此處，z 作為^ ’設置有兩台相機U 1相機6及第2相機7) 下為攝影部。各個相機是 相機7) ，呈45。的角度的方式朝下(:方向=丨拍攝方向與 1相機6及第2相機7是以自z方;酉己置，且第 正交的方切* 自Z方向觀察’拍攝方向相互 』万式配置。此處，將第1相 土 配置成可同時拍_量對象4相機7分別 續、及該所期制區域的鏡像。 」關望的 201224392 另方面，於基準平板11的下方設置有xy轴用格子 13與xy軸用光源12，於其更下方設置有連接於基準 、’反11且使基準平板11 (即，基準面11a)在z軸方向 上移動1 z轴移動平台14。該些是為了根據全空間表格化 方法，製作作為相位值與空間座標的相關聯的空間座標表 格而使帛卩下，冑形狀測量裝置刚的各構成要 說明。 Z軸用光源1經由z軸用格子玻璃2,照射用於將格子 圖案，影於測量對象物0及基準面丨i a上的光。該光只要 可穩定地投影格子圖案，則並無特別限定。例如，可使用 發光一極體（Light Emission Diode，LED)。於本實例中， z軸用光源1是以其照射方向與基準平板u的法線方向平 行的方式配置，但只要可將格子圖案投影於測量對象物〇 及基準面lla上，則配置並無特別限定。 z軸用格子玻璃2具有如圖2所示的一維的格子圖 案，其用於使由z軸用光源1所照射的光通過，而將一維 的格子圖案投影於測量對象物〇上。 第1透鏡3是以使自z軸用光源丨照射且通過了 z軸 用格子玻璃2的光聚光，將格子圖案投影於測量對象物〇 及基準面lla上的方式而配置。 壓電平台4與z軸用格子玻璃2結合，並使z軸格子 玻璃2移動。藉此，可使投影於測量對象物〇及基準面Ua 上的一維的格子圖案移動。該壓電平台4是以使如圖2所 示的z軸用格子玻璃2中所描繪的一維的格子圖案移動的 201224392 多;条直線朝不子圖案的以規定的間隔配置的 玻璃2結合。、 仃的方向移動的方式與Z轴用格子201224392 VI. Description of the Invention: [Technical Field] The present invention relates to an apparatus and method for measuring a shape of a measuring object, and more particularly to a shape of a back side of a measuring object, which is highly accurate Apparatus and method of measurement. [Prior Art] The industry is looking forward to the use of the non-contact shape measuring device using the lattice projection method, the non-contact stress and strain measurement using the interference beam analysis, the displacement measuring device for the digital holography, and the like. The stripe image measurement data is synthesized and obtained by this. However, in the previous stripe image analysis method, for example, the technique has been performed for each pixel as before, and the processing is disclosed in Patent Document 1, and the calculations are performed: The area to analyze the method of existence can not be considered for each = in = = = before - generally by adding not only the information: 仃: into the question. The prime value is determined by the evaluation value and the data value of the evaluation value is high: the method for evaluating the image of the prior art is well-formed: the image is dissected and the distribution is passed; ==: 201224392 frequency The bandpass m of the characteristic of the sensitivity and the agro-light distribution of the output of the root_bandpass chopper are used to calculate an evaluation value indicating the image quality. This prior image evaluation method has a feature that the image can be evaluated and the image of the recorded image is qualitative. However, this prior image evaluation method has the following problems: n Evaluation of image quality for digital images captured by a camera 'The phase data analyzed in the fringe image measurement cannot be evaluated. Therefore, in Patent Document 3, there is proposed a technique of performing phase analysis on each phase of the phase-shifted image data*, obtaining a plurality of phases such as cloth data, and based on the stripe image associated with each of the plurality of phase distribution data. The negative material is used for each pixel mosquito value, and the # price recording corresponding to each pixel, n can be synthesized with high precision of the township-age cloth (four). [PRIOR ART DOCUMENT] [Patent Document 1] Japanese Patent Laid-Open Publication No. Hei. No. 325922 (Patent Document 3) Japanese Patent No. 3837565 In recent years, it is expected that the shape of the entire surface of the object to be measured is not only the object to be measured, but the measurement of the entire surface of the object is performed. However, in the method described in Patent Document 3 Since the object to be measured is photographed from one direction, the τ problem is left in the measurement of the shape on the back side of the object that cannot be attached. Measure of the field of view [Description of the Invention] The present invention has an object to provide a device and a shape measuring method for measuring the back surface shape of a measurement object without entering the field of view of one camera. The inventors of the present invention have made an effort to study the method for solving the above problems, and found that it is effective to arrange a mirror around the object to be measured, and to arrange the camera to simultaneously capture an image of the object and the image reflected in the mirror. The image = image, which is a photograph of the measurement object and a mirror image reflected in the mirror, thereby completing the present invention. The shape measuring device according to the present invention is a device for inserting a shape of a measuring object, and includes a grid pattern projecting portion for projecting a lattice pattern on the upper thorn count; at least one mirror, configuration measurement Around the object; at least one of the photographing units captures the image of the object to be measured and the object to be measured reflected in the mirror; and the image of the object to be measured and the image of the image taken by the processing σ卩' Each phase of the image performs phase analysis processing, calculates the shape of the object to be measured, and combines the captured image of the object to be measured with the image of the image, and arranges each of the image capturing units to be simultaneously photographed. A mirror image of at least a portion of the area of the object to be measured and the at least part of the field. Further, the shape measuring device according to the present invention is characterized in that the shape measuring device is a device for measuring the shape of the object, and includes a lattice projection portion, wherein the lattice pattern is projected on the object to be measured; and the periphery of the object to be measured; And at least one photographing unit that photographs the object and the mirror 201224392 image of the object to be measured reflected on the mirror towel; and sets each of the photographing portions to be simultaneously (four) to the portion of the object to be measured A mirror image of the area and the at least part of the area. Further, the shape measuring device according to the present invention, wherein each of the above mirrors is provided with a stray light preventing wall. Further, the shape measuring apparatus according to the present invention is provided with two mirrors and two photographing sections. Further, as in the present invention, the measuring device has two mirrors and one photographing unit. Further, the shape measuring method of the present invention is a method of measuring a shape of a measuring object in which at least one mirror is disposed, and includes a grid pattern projection step of projecting a lattice pattern on the object. In the image capturing step, the object to be measured and the object to be measured (10) in the mirror towel are simultaneously photographed; in the secret step, only the grid pattern is moved by a predetermined size; and the shape calculating step, the image of the real image and the mirror image are Each of the implementation phase analysis 3' calculates the shape of the object to be measured; and the shape calculation step is performed by the image capturing step and the number of repetitions of the above-described turning step. The shape measuring method of the present invention is processed, wherein the phase analysis 上述 is performed by: according to the full-space tabulation method, for each pixel, a table value of the phase value associated with the silk is made = 3^, Further, according to the phase of each pixel obtained by the phase shift method, the shape measuring method according to the present invention further includes the above-mentioned measuring object based on the evaluation value corresponding to the ratio of the fresh component of the ®I image taken in 201224392 The image of the object and the image of the image are combined into a composite processing step of the image. [Effect of the Invention] According to the present invention, by arranging the mirror around the object to be measured, the county can be simultaneously produced from the money direction. Since the shape is measured, it is possible to measure the shape of the object surface side of the measurement object that cannot be obtained from the self-financing direction. In addition, by combining images captured by a plurality of cameras, even the object to be measured contains In the case of the gold system or when the shape of the measuring object has a curved surface, the shape of the object to be measured can be accurately calculated by reducing the influence of the halation. The other objects, features, and advantages will be more apparent and understood. The following detailed description of the preferred embodiments and the accompanying drawings will be described in detail below. [Embodiment] [Shape Measuring Device] Hereinafter, the present invention will be described with reference to the drawings. The shape measuring apparatus according to the present invention includes: a lattice pattern projection unit that projects a lattice pattern on a measurement object; at least one mirror is disposed around the measurement object; and at least one imaging unit captures the measurement object and a mirror image of the object to be measured reflected in the mirror; and an analysis processing unit that performs phase analysis processing on each of the image of the object to be measured and the image of the mirror image, calculates the shape of the object to be measured, and takes the above-described measurement The image of the object 201224392 is combined with the image of the above-mentioned mirror image. Here, it is important to arrange each of the image capturing sections so that at least a part of the object to be measured can simultaneously capture a mirror image of the at least a part of the image. A diagram of a shape measuring device according to an example of the present invention. The shape measuring device 1 includes: z-axis light 1. Z-axis lattice glass 2, first lens 3, piezoelectric stage (piezo stage) 4, Peltier cooling device 5 first camera 6 and second camera 7, and reference point projection laser .8, the first mirror 9 and the second mirror 10, the reference plate Η, the xy-axis light source, the =-axis lattice glass 13, the z-axis moving stage 14, and the analysis processing unit quasi-level shape measuring device _ The object 0 is placed on the base object, and a double-sided mirror (the first mirror 9 and the second mirror 1〇) that reflects the ten-object object 0 is disposed around the Dickson plate 11. The top of the lattice plate 11 is provided with a crucible The axis light source is used for the dice axis, the lattice plate 2, the fl lens 3, and the glaze pattern is projected on the __ 〇, and the lattice car is projected toward the plaid pattern by the light source ray direction toward the _2 side. Here, z is ^ 'Set two cameras U 1 camera 6 and 2nd camera 7) Under the photography department. Each camera is a camera 7) and is at 45. The angle of the way is downward (: direction = 丨 shooting direction with 1 camera 6 and the second camera 7 is from the z-square; 酉 set, and the orthogonal square cut * observe from the Z direction 'shooting direction mutual 』 million In this case, the first phase soil is arranged so that the target object 4 camera 7 can be simultaneously photographed and the mirror image of the planned area is continued. "201224392 of the look-ahead. In addition, the lower side of the reference flat plate 11 is provided. The xy-axis grid 13 and the xy-axis light source 12 are provided below the d-axis 11 and the reference flat plate 11 (i.e., the reference surface 11a) is moved by the z-axis moving platform 14 in the z-axis direction. In order to create a space coordinate table as a phase value and a space coordinate according to the full-space table method, each of the configurations of the 胄-shaped shape measuring device will be described. The Z-axis light source 1 is via the z-axis. Light is applied to the measurement object 0 and the reference surface 丨ia by the lattice glass 2. The light is not particularly limited as long as the lattice pattern can be stably projected. For example, a light-emitting pole can be used. Light Emission Diode (LED). In this example The z-axis light source 1 is disposed such that the irradiation direction thereof is parallel to the normal direction of the reference plate u. However, the arrangement is not particularly limited as long as the lattice pattern can be projected on the measurement target object and the reference surface 11a. The lattice glass 2 for a shaft has a one-dimensional lattice pattern as shown in FIG. 2 for passing light irradiated by the z-axis light source 1 and projecting a one-dimensional lattice pattern onto the object to be measured. In the first lens 3, the light is condensed by the light source 丨 from the z-axis and passed through the grating glass 2 for the z-axis, and the lattice pattern is projected on the object to be measured and the reference surface 11a. The z-axis lattice glass 2 is combined with the z-axis, and the z-axis lattice glass 2 is moved. Thereby, the one-dimensional lattice pattern projected on the measurement target object and the reference surface Ua can be moved. As shown in Fig. 2, the z-axis is more than 201224392 moved by the one-dimensional lattice pattern depicted in the lattice glass 2; the straight lines are joined to the glass 2 arranged at a predetermined interval in the sub-pattern, and the direction of the movement of the crucible Lattice with Z axis
3 ^ f 1 iiH 成，他 子圖案投影部並不限定於上述構 吏格子圖案移動的方式構 投影儀來代替上述構成。 例如亦叮使用液曰曰 高的位ί朝4二Ϊ置5疋於基準平板11的上方的裝置的最 卻。 Μ ’且以不使整個裝置過熱的方式進行冷 物6隹及第2相機7如圖3所示，拍攝測量對象 的圖傻平板11的周圍所配置的鏡子中的鏡像〇, :::因此’將各個相機配置於可將測量對象物。與映 卜兄：：：鏡像〇,兩者同時納入視場中進行拍攝的位置 ^於^實例中’如上所述’第！相機6及第2相機7分 別疋以自y方向觀察，其拍攝方向與z軸呈45。的 方式朝下（即，-z軸方向）配置，另外，帛4機^第 2相機7是以自2方向觀察，各自的拍攝方向相互正交且 與相對的鏡面的法線方向平行的方式而配置。但是，口、要 上述要件，即，各個相機被配置於可將測量對象物〇與映 在鏡子中的鏡像0,兩者同時納入視場中進行拍攝的^立置 上’則相機的設置位置或拍攝方向等並無限定。 作為該些第1相機6及第2相機7，例如可使用互補 201224392 ·>/ vi^r^v/L/Λχ. 金氧半導體（Complementary Metal Oxide Semiconductor， CMOS)相機或電何輕合元件（charge Coupled Device， CCD)相機。 再者，自相機觀察，如圖5所示，測量對象物〇與其 鏡像〇’無需完全分離，鏡像〇'的一部分亦可被測量對象物 〇遮蔽。於此情況下，只有未被遮蔽的區域成為相位分析 處理的對象。 . 另外，欲測量形狀的區域（所期望的測量區域）未必 是測量對象物0的整體。於此情況下，只要將各個相機配 置於可將測量對象物0的所期望的測量區域（即，測量對 象物0的至少一部分的區域）與映在鏡子中的所期望的測 量區域的鏡像同時納入視場中進行拍攝的位置上即可。另 外，自相機觀察，上述所期望的測量區域與其鏡像無需與 上述所期望的測量區域完全分離，鏡像的—部分亦可被所 期望的測量區域遮蔽。於此情況τ，亦只有未被遮蔽的區 娀成為相位分析處理的對象。 基準點投光肖雷射11 8是設置於基準平板11的上方， 對測量對象物0或基準面Ua上的適當的位置照射 光’且於將所拍攝的測量對象物〇的圖像與映在鏡子^ 置這樣的基準點（記號）。於本實例中，藉由位 基但並不限定於此，例如亦可事先於 上或基準面11a上的適#的位置標示基準點。 第1鏡子9及第2鏡子10是設置於基準平板U的周 12 201224392 圍’且為了映照測量對象物Ο的鏡像0,而設置。於本實例 中如圖1所示’兩塊鏡子是以與基準面iia正交，進而 彼此的鏡面正交並接觸的方式而配置。該些鏡子的配置如 上所述，只要以使各個相機可同時拍攝測量對象物Ο與映 在鏡子中的測量對象物〇的鏡像的方式而配置，則並I特 別限定。 曰圖5表示將格子圖案投影於基準平板u上所載置的測 置對象物0上，且鏡像〇,映在第2鏡子1〇中的樣子。使 用該些鏡子，並利用一台相機同時拍攝測量對象物〇與映 在鏡子中的鏡像等同於使用兩台相機，自兩個不同的位置 拍攝測量對象物〇。即，僅使用_台相機，便可獲得利用 配置於不同位置上的兩台相機的測量結果。 n ， 仰叫啊又且，处口J狗攝無鏡子時I法拍蟲 的部分。因此，藉由以可拍攝測量對象物〇的整體的^ 而配置鏡子，亦可測定測4對象物0整 此處，如圖1麻，當使兩錢子於極其 位置 接近（或接觸）來進打配置時，若自z軸用光源】昭射 則存在由測量對象物〇或第！鏡子9及第2鏡子忉 射或散射的光彼此干涉’無法高.精度地撕形 況。於此情況下，如圖6所示，藉由在鏡子中設置= 防止壁，可使阻礙形狀測量的光的干涉受卜= =止壁16只要可抑制上述干涉，則大小或形丄3 的形狀進行測量時载 基準平板11於對測量對象物〇 201224392 置該測量對象物ο，並且其表面具有根據全空間表格化方 法製作作為相位值與空間座標的相關聯的空間座標表格時 所使用的基準面11a。於本實例中，用於測量對象物〇的 形狀測量與空間座標表格的製作的光源不同，形狀測量用 的光源（z軸用光源1)是配置於基準平板u的上方，另 外二間座私表格製作用的光源（xy軸用光源12)是配置 於基準平板11的下方。因此，作為基準平板u，使用 mmx50mm的蛋白石層玻璃，以可自上下方向（即，士z 軸方向）將格子圖案投影於基準面lla上的方式構成 外’為了確保後述的校正的精度，雜準平板u的表 為散射面之類的加工處理，並將所獲得的 政射面作為基準面lla而朝上（即+z方向）配置。 表示藉由z軸用光源」，自基準面m的上方經 由z軸用格子玻璃2而投影了—維格子圖案的、， 以及映在第2鏡子1()中的鏡像。另外，圖8表 a’ 軸用光源12 ’自基準面lla #下方經由xy細格 13而投影了二維格子圖案的基 :玻璃 子10中的鏡像。 早曲lla以及映在第2鏡 只要並不限定於蛋白石層玻璃， 圖案，自下方投影用於校正的二維格^牵篁的一維格子 限定。 料子_ ’職無特別3 ^ f 1 iiH is formed, and the sub-pattern projection portion is not limited to the configuration in which the above-described configuration lattice pattern is moved, instead of the above configuration. For example, the height of the liquid 曰曰 ί 朝 朝 朝 朝 朝 朝 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。. Μ 'and the cold object 6 隹 and the second camera 7 are not overheated by the entire apparatus. As shown in FIG. 3, the mirror image in the mirror disposed around the figure 11 of the measurement object is photographed, ::: 'Configure each camera to measure the object. And Ying Buxiong:::Mirroring, both of which are included in the field of view for shooting. ^In the ^ example, as mentioned above, the first! The camera 6 and the second camera 7 are viewed from the y direction, respectively, and the shooting direction is 45 with respect to the z-axis. The mode is arranged downward (that is, in the -z axis direction), and the second camera 7 is viewed from the two directions, and the respective imaging directions are orthogonal to each other and parallel to the normal direction of the opposing mirror surface. And configuration. However, the above-mentioned requirements are required for the camera, that is, each camera is disposed in a mirror image 0 that can be measured in the mirror and both are incorporated into the field of view for shooting. There is no limit to the shooting direction or the like. As the first camera 6 and the second camera 7, for example, a complementary 201224392 ·>/ vi^r^v/L/Λχ. Complementary Metal Oxide Semiconductor (CMOS) camera or an electric light-emitting component can be used. (charge Coupled Device, CCD) camera. Further, as seen from the camera, as shown in Fig. 5, the object to be measured is not completely separated from the mirror image 〇, and a part of the mirror image 亦可 can be shielded by the object to be measured. In this case, only the unmasked area becomes the object of phase analysis processing. Further, the area to be measured (the desired measurement area) is not necessarily the entirety of the measurement object 0. In this case, each camera is disposed at the same time as the mirror of the desired measurement area of the measurement object 0 (that is, the area of at least a part of the measurement object 0) and the mirror image of the desired measurement area reflected in the mirror. It can be placed in the field of view for shooting. In addition, from the perspective of the camera, the desired measurement area and its mirror image need not be completely separated from the desired measurement area, and the mirrored portion can also be obscured by the desired measurement area. In this case τ, only the unmasked area becomes the object of the phase analysis processing. The reference point projection light ray 11 8 is disposed above the reference plate 11 and illuminates the appropriate position on the measurement object 0 or the reference surface Ua and images the image of the object to be imaged. Set such a reference point (symbol) in the mirror. In the present example, the position is not limited thereto by the position, and for example, the reference point may be indicated in advance on the upper or the position of the reference surface on the reference surface 11a. The first mirror 9 and the second mirror 10 are provided on the circumference 12 201224392 of the reference flat plate U and are provided to reflect the mirror image 0 of the measurement object. In the present example, as shown in Fig. 1, the two mirrors are arranged in such a manner as to be orthogonal to the reference plane iia and further orthogonal to and in contact with each other. The arrangement of the mirrors is specifically limited as long as the respective cameras can simultaneously image the measurement object and the image of the measurement object reflected in the mirror. Fig. 5 shows a state in which the lattice pattern is projected onto the object to be placed 0 placed on the reference flat plate u, and mirrored to reflect the image in the second mirror. Using these mirrors and simultaneously taking a measurement object with a single camera and an image reflected in the mirror is equivalent to using two cameras to take measurements of objects from two different locations. That is, using only the _ camera, measurement results using two cameras arranged at different positions can be obtained. n, yelling and ah, and the mouth of the J dog when shooting without a mirror. Therefore, by arranging the mirror so that the whole object of the object to be measured can be photographed, it is also possible to measure the object 4 as shown here, as shown in Fig. 1, when the money is approached (or contacted) at an extreme position. When entering the configuration, if the light source from the z-axis is used, the object to be measured or the first is present! The mirror 9 and the second mirror illuminate or scatter the light to interfere with each other. In this case, as shown in Fig. 6, by providing the = prevention wall in the mirror, the interference of the light that hinders the shape measurement can be made = = the stop wall 16 can be suppressed in size or shape as long as the above interference can be suppressed When the shape is measured, the load reference plate 11 is placed on the object to be measured 〇201224392, and the surface thereof is used to create an associated space coordinate table as a phase value and a space coordinate according to the full-space tabulation method. Reference plane 11a. In the present example, the shape measurement for measuring the object object is different from the light source for making the space coordinate table, and the light source for shape measurement (the light source 1 for the z-axis) is disposed above the reference plate u, and the other two are private. The light source for the table creation (the xy-axis light source 12) is disposed below the reference flat plate 11. Therefore, as the reference plate u, an opal layer glass of mmx50 mm is used, and the lattice pattern can be projected from the vertical direction (that is, the z-axis direction) onto the reference surface 11a. To ensure the accuracy of the correction described later, The table of the quasi-plate u is a processing such as a scattering surface, and the obtained political surface is arranged upward (ie, in the +z direction) as the reference plane 11a. The light source for the z-axis is projected from the upper side of the reference plane m by the lattice glass 2 for the z-axis and the mirror image of the second mirror 1 (). Further, in Fig. 8, the a'-axis light source 12' projects a base of the two-dimensional lattice pattern from the lower surface of the reference plane 11a# via the xy fine grid 13: a mirror image in the glass sub-frame 10. The early curve lla and the second mirror are not limited to the opal layer glass, and the pattern is defined by the one-dimensional grid of the two-dimensional grid for correction from below. Material _ ’ job no special
Xy軸用光源!2如圖2所示，具 針孔板^、以及第—可照的 201224392 光的方式構成。即，自τ p 孔板12b的π & & 先源12a所照射的光通過針 it t 光源’藉由透鏡仏而成為大致平 灯先後，通過xy軸用格子玻璃13。 ㈣H用，子麵13是於根據全空間表格化方法製作 二早時使用’且插緣有用以投影於基準面m上 平板、11的不圖案。該Xy軸用格子玻璃13是配置於基準 的正下方。xy軸用格子玻璃13如圖2所示，藉由 好刀二有一維的格子圖案的兩; =3b)以各自的格子圖案彼此正交的m :=::用描缘有二維的格子圖案的-塊格子玻璃 z軸移動平台14與基準平板u社入， 12的下方。該z軸移動平台Μ如二所示\ Ϊ:Γ使基準面110於ζ軸方向上移動。於： 單轴平台作1 Γ ο.1 Mm的精度定位的帶有反饋功能的 對象物^ ‘、LZ移動平台14t’且用於形狀測量時的測量 ，象物〇的向度（即’Z座標）調整 刀析處理部15對由相機所賴 鏡子μ鏡像_像實施相位分析^物 ㈣=物0的形狀。此處，相位分析二= 疋’於本實例中’㈣魏_移 ’事先針對各像素製作物= 工間私的相關聯的空間座標表格並加以保存。藉此，若 201224392 =ri出圖像的各像素的相位值，則可立即求出空 形狀。丁 y座標及z座標）’即’測量對象物0的 另外，於本實例中，藉由使用兩台相 方；!所拍攝的多幅圖像。因此，分析處理部 人忐。二Γ…則量對象物0的圖像與鏡像〇,的圖像加以 二拍摄般’該合成是於對由進行拍攝的不同的相機 =的圖像或鏡像〇,的圖像適當地實施旋轉或反轉處 ‘二nr定，例如可藉由將圓像資 的後述的規定的=二’亦可根據本發”所使用 的測ί ^ ϋ用乡個鏡子及相機將不進人—自相機的視場 的圖像力以纟^的背面侧映在鏡子上，因此藉由將所拍攝 定組合，可對測量對象物0的整周的形狀進行測 的ϋ對象物例如為電子零件’於含有金屬部分 於二、則詈ίί測量對象物具有曲面的情況下’若照射用 則有時會產生光晕(haiati〇n)，但若相機 罟姑;t地5，則產生光暈的位置亦不同，因此將由不同位 塑^ 像加以合成’藉此亦可降低光晕的影 曰’而而精度地求出測量對象物的形狀。 的兩3機形狀測量裝置100具備作為攝影部 (第1相機6及第2相機7)、及兩面鏡子（第 兄9及第2鏡子10)，但相機及鏡子的數量並無限定。 201224392 :如’可如圖9所示的形狀 機（相機26)來代替作為第丨^屣置200般，具備一台相 的兩台相機（第1相機^及第2】的形狀測量裝置100中 (鏡子30)來代替兩面鏡子（ 機7) ’並具備一面鏡子 且將其他構成設定為與形狀測鏡子9及第2鏡子1〇)， 對構成與形狀測量裝置1〇〇不裝置丨〇〇相同。此處，僅 另外.，亦可如圖10所示的=的部分附加符號。 一台相機（相機36)來代替形狀測量裝置300般，具備 機（第1相機6及第2相機/)，剛量裝置咖中的兩台相 狀測量裝置100相同，且以可並將其他構成設定為與形 個鏡像的方式構成相機36。°此^時拍攝測量對象物〇及三 置100不同的部分附加符號’僅對構成與形狀測量裝 與由朝向不同的方向的四二彳:此，可利用一台相機獲得 藉由使用此種形狀測拍攝相同的效果。 對測量對象物的形狀進行測I G’可同時自多個方向 的拍攝中無法獲得的測量對象===一個方向 :便於測量對象物中含有金屬=== 的形狀具有曲面的情況下，亦 2者射，象物 地算出測量對象物的形狀。 -箪的影響而向精度 (相移法） 此處’對本發明中用作相位分析處理 及全Π表3方法進行說明。首先，對相移法進行說明法 圖11疋表不格子圖像的亮度分布與相位分布的關係 201224392 的圖。圖11 (a)表示格子的亮度分布，圖n (b)表示格 子的相位分布。另外，圖12是表示相移量與已相移時的亮 度變化的關係的圖。 格子或干涉條紋的亮度值I (x，y) 一般如圖12 (a) 所示，於空間（x,y)上分布成餘弦波狀。若利用式子來對 其加以表示’則成為如式（1 )般。 ^{x,y) = ^x,y)cos(0(x>y)) + b(xyy) ( 1 ) 此處’點（x，y)是所拍攝的圖像内的一點，a (X，y) 及b (x，y)分別表示亮度振幅與背景亮度，Θ (x，y)表示 格子的相位值。於拍攝有格子的圖像（以下，稱為「格子 圖像」）的情況下’相位可藉由實數整體來表示，但亦可看 作自0至2π為止的2π週期的重複。圖n(b)是將0(x，y) 的分布表現為自〇至2π為止的重複的圖。 相移法是一方面使格子的相位僅變化1個週期一方面 拍攝多幅格子圖像’並根據所獲得的多幅圖像求出相位分 布的方法。於所有像素中，亮度變化1個週期’因此可根 據，亮度變化’針對各點獨立地，即，不使關圍的像素 的冗度變化的資訊地求出相位值。因此，該方法是對於具 有階差或不連續的物體的形狀測量有效的方法 。此處，以 根據最普遍使用的每次相移π/2而成的4個亮度值求出相 位值的情況（即’相移次數為4次的情況）為例，對相移 法的原理進行說明。 201224392 詈若示的格子的亮度分布的式中追加相移 sa，則成為式（2)。 I(x,y,a) = a(x, cos(0(x, ;；) + «) + b(x, y) (2) 圖12表示具有初始相位θ的點（像素）的相 與亮度變化的關係。所謂初始相位，是指相移量為 格子的相位。若將相移量自〇起每次變化π/2時的亮戶八 別設定為I〇、Ii、I2及Ι3，則該些可分別如式（3)〜^卞刀 般表示。再者，於以下的式中省略（x，y)的記載。"6) I0(xyy) = acos6-\-b (3) ’1 (又’ >0 = a c〇s〔^+昏)+ 6 (4) 八(x，少)=β c〇s0 + π)+δ (5) /3 〇χ，3〇 = a cos(0 + 专)+ 6 (6) 可自該些式獲得以下的式（7)及式（8)。 —/〇 = —2acos0 (7) (8) (8)201224392 h =2asin^ 進而’可自式（7)及式（8)導出以下的式（9)，並 根據該關係式而求出相位值θ。即，若可獲得相移量為〇、 f、π及3π/2時的亮度l〇、L、^及乌，則求出針對該像 素的相位值Θ。 J2~h (9) 此處，藉由増多相移的次數（即，自〇至271為止的刻 度數）’可降低相機的隨機雜訊的影響。若將相移次數設定 為N，將相移量為27lk/N時的亮度設定為lk，則可導出式 (10)，並根據該關係式而求出相位值θ。 =-tan0 do) ^。如此，可藉由相移法而求出圖像上的各像素的相位值 (校正方法） 算出t二根據藉由相移法對各像素所求出的相位值， 座標（x、y、z)’即’測量對象物的形狀的方法 20 201224392 *於t發明中，不使用如先前技術般，利用使相位值與 空間座標產生對應的變換式求出該變換式中所使用的表i 二’於本發3种，採用事先針對相機的每個像素 i梭f投影的格子的相位值與空間座標的對應_並加以 二!i的方，’將此稱為校正（caiibrati〇n)。若藉由該校 ^ =得投影於測量對象物上的格子的相位值，則不 變換式的座標的計算，參照所獲 門;r ΐ 0 ί、表格，便可立即求*測#對象物的表面的空 間座軚，即，測量對象物的形狀。 (全空間表格化方法） 專利====法(例如，參照曰本 如圖^示，使垂直t 量:原卿 置的基準面於ζ軸方向上^度方向）而設 儀是事先固定於基準面的上多動。相機與投影 於基準面上。此時，並不特 技影儀將格子圖案投影 =一格 示的直線L上的點。對==攝圖13(b)所 R2、...Rn，分別拍攝點Pq、p ‘、、' p基丰面的位置Ro、心、 值θ〇、θ!、Θ2、.，^可藉由相^二fN。各個點的相位 圖13 (b)表示―個傻財出° 間座標（X、y，Z)的關係。針f攝Hi與基準面上的空 1個像素所拍攝的各基準 21 201224392 面上的點 P〇、p,、Pi、.. D , ' a ·ΡΝ，自投影於基準面上的二維的 4子圖案翁X鍊及y座標， L 針對㈣以絲，可藉二^^ 變換格子法料财出x方向及y方向的 連接此分別獲得各_x座標 及ϋ (例如’參照日本專利第规918號公報）。 相對二ΐίΐ準面的每個位置’可針對每個細得 相對於投#好_域θ的χ座標 ==格:圖案的相位值”能於基準面二 付’仁猎由縮小基準面的間隔，並對其中間進行内插，可 咼精5地求出相對於所有相位值的空間座標。 藉由上述权正方法’可針對每個像素分別獲得相位值 θ〇 1 °2 知與 Χ 座標 Xq、χι、χ2、"·Χν，y 座標 y〇、 (a)〜圖14 (c)中以黑點來表示該些的關係。 yi、y2、’^及z座標z〇、Ζι、z，、…Zn的關係。於圖14 相對於相位值Θ的x座標、y座標及z座標分別平滑 地變化。將相位值θ _隔地分割，且分職行内插而求 出所對應的X座標、y座標及z座標，藉此製作空間座標 表格。於格子投影法的情況下，在一般的配置中，相位= 與座標的對應關係成為平滑的曲線狀，因此可藉由線性内 插而期待足夠的精度，當然亦可進行高階内插。於圖14 中，進行線性内插，以連結黑點間的直線上的點的形式製 作空間座標表格。針對圖丨4中由κ所表示的區間，可藉 由進行外插而求出相位值與空間座標的關聯。 22 201224392 如此，可根據全空間表格化方法， 範圍 與空間座標U座標、y座標及z座標）的2相位值θ 標表格。終㈣於進彳爾袼子_=^座 亦可根據㈣㈣解，將她連接後的她的障座兄下’ 關係表格化。藉由進行相位連接，可擴h料的 (測量對象物的形狀測量） 面上先r m乂所不’將測量對象物設置於基準 圖幸二二％、、权正處理時的袼子圖案相同的格子 ::像素拍攝測量對象物上的點影丄= 圖案^錄ep可料轉素_位=^點上的格子 藉由參照事先製作的空間座標表格，而求出 得與㈣的表格便可立即獲 y座標與z奸，目對應的χ座標χΡ。同樣地，關於 而立即自二。:r堇參照以…^^ 得的事先製作的空間座標表格，可自所獲 象物的，出空間座標’從而可高速地測量該測量對 相機 (再抽樣處理） 藉由上述形狀測 的像素所拍攝的 量方法所獲得的形狀測量結果可作為 點的座標分布而求出，因此如圖15 23 201224392 -- -X-- (a )所示 而獲得。㈣m於xy平面鱗㈣關座標分布 崎付巧4由配置於不同位置料個 η或實像的圖像與鏡像的圖像加以 間座標。因此，進行如下地抽樣的點的空 點S’及找P出包^知道X座標及^標的點Ps的3個 〆 c 、人，將通過该3點的平面與自抽;& 抽樣點的z座標。藉===線的交點作為再 一 楮田以上述方式進行再抽樣處理，如圖 所不’可獲得於xyif面巾相等間隔地抽樣的點的 另外，於本發明的一實例的形狀測量裝置100中，以 自Z方向觀察兩台相機的拍攝方向相互正交的方式配置該 兩台相機’因此為了將由第丨相機所拍攝的圖像與由第2 相機所拍攝的圖像加以合成，必需對_個圖像實施僅旋轉 90°的旋轉處理，另外，於由各個相機所拍攝的各圖像中， 對鏡像的®像實施圖像反轉處理。當第i相機與第2相機 之間的配置關係不同於形狀測量裝置謂時，根據該配置 關係，適當地進行圖像的旋轉及旋轉處理。 (根據相位評價值的圖像合成方法） 對於來自多個方向的形狀測量結果，根據實施了上述 =抽樣處理或圖像反轉處理及/或圖像旋轉處理的圖像的 f間座標’進行_的合成處理。此處，®像的合成處理 可藉由針對每個像素而合成高度資訊（z座標）來進行。 201224392 方定’例如可將圖像資料簡單地加以 可11由進行加權並純平均來進行。 具有可靠$出1為判定所拍攝的多幅圖像的圖像 Γ目位；俨i進行圖德的指標的相位評價值，根據所獲得的 亦無限;。於本發明Γ:各;蝴相位評價值 傅里葉變換，並作為-次頻率ω1的功率 二義)：情況下，當亮度的振 的，和度或亮度的振幅小時(即，亮度的變化小時):= 二料:包含隨機雜訊時’相位評價值變小。因此，相位評 =值是可確認相位值的可靠性的值’於圖像合成時，可作 為用於合成的圖像資料的取捨選擇時的指標。 各方:圖像的合成處理是針對二象素求出來自 各方向的冋度（Ζ方向的位置）資訊 :r=r_ 除…： J本二月中，將以下的式⑻用於相位評價值E⑹ 、十异。此處，F (ωη)表示頻率ωη的功率頻 ， 另外，Ν表示與最高的頻率相對應的次數。 、 (11) 25 201224392 如此 ;平僧僧招、^像素中’僅合成由式（11)所求出的相位 靠性低圖:此可獲得排除了可 产（Z 將基準面作為測量對象物，—方面變更高 。數Si::面進行基準面的形狀測定。此處，將相 時的f 16(e)表不基準面的Z轴的位置為500 μιη Θ (即布圖。此處，圖16⑷表示將自四個方 = i實像與鏡像、以及第2相機的實像與 i 的圖像的圖像資料的全部加以平均而成的 表示藉由本發明的方法，根據相位評價值 =1幅圖像加以合成所獲得的結果。如根據該些 =㈣I可知藉由實施本發明的合成處理，高度 均減少’測量精度得到提昇。w 16⑷表示 位置的實施了本發明的合成處理的z座“布，、可z 知於任何位置’測量精度均得到提昇。 其次，對圖17所示的作為包含超硬合金的測 的試樣的形狀測量結果進行說明。於該試樣的表面 先使用萬能投影機（V-12BS型Ν〇·13〇〇〇21)及 高/測厚計_-5觀〇遍523)所測定的圖17的表^ 的階差，伴隨自U)前進至⑺，每次增高約10卿。此 26 201224392 處’對該階差部的形狀測量結果加以關注。再者，校正及 測量疋將相移次數設定為16次，如上述般連續地拍攝五幅 圖像並將圖像資料加以平均，降低隨機雜訊的影響。 圖18(a)〜圖18(e)是表示所測量的試樣整體的高度 座標）分布圖的圖。另外，圖19(a)〜圖19(e)是根據圖18(a) 〜圖18(e)所示的結果，表示圖18(a)〜圖18(e)所示的位置 j上的X方向的位置i與高度（z座標）分布的圖，。於各個 圖中’圖19 (a)表示第1相機的實像的圖像，圖19 (匕） 表示第1相機的鏡像的圖像，圖19 (c)表示第2相機的 實像的圖像，圖19 (d)表示第2相機的鏡像的圖像，圖 19 (e)表示所合成的圖像，此處，合成時的相位評價值的 閥值為15。 如該圖19 (a)〜圖19 (d)所示，可知自四個方向所 拍攝的各圖像中的高度的不均較大，相對於此，如圖19(e) 所示，可知藉由本發明的圖像合成處理，高度的不均 測量精度得到提昇。 圖20是拍攝作為包含銘的測量對象物〇的試 得的圖像。如根據該圖而明確般’產生了光暈，於該狀^ 下二無法趣據該圖料確地制量對象物0的形狀進行 測量。 1』 的本〜圖耶)表示藉由具備兩台相機及兩塊鏡子 的本發明的形狀測量裝£丨⑻，自四 成的圖像。圖21 (a)〜圖21⑷矣一二攝以樣而 相;C )表不由弟1相機及第2 相機所拍攝的實像及鏡像的圖像與各自 27 201224392 21 (e)表示由本發明的方法所合成的圖像。如根據圖2i (a)〜圖21 (d)的圖像而明確般，各圖像的高度分布因 光暈而受到影響並包含雜訊，但於圖21 (e)所示的經合 成的圖像中’可知光暈的影響大幅度降低。如此，藉由本 發明的形狀測量裝置1〇〇,即便於測量對象物〇包含金 屬、產生光暈的情況下，亦可高精度地對測量對象物〇的 形狀進行測量。 [形狀測量方法] 其次，對本發明的形狀測量方法進行說明。本發明的 形狀測量方法是於相位分析處理中，使用根據全空間表格 化方法的相移法作為其一例，但並不限定於此。當使用根 據全空間祕化方法的相移法時，必需進行事先針對每個 像素製作作為相位值與空間座標的相關聯的空間 的校正，並保存所獲得的空間座標表格。因此， ::多個基準面位置上的z軸與xy軸的 : =光圖像。首先’對校正方法,的拍攝格子= /、基準點投光圖像的校正方法進行說明。 (校正用圖像的拍攝處理） 準：格子投影圖像及基 此使Z軸用棬早诚描1電干〇 3移動至規定的位置，藉 後，點亮作為Z轴用光源 藉由第1相機6及第 28 201224392 2相機7減投影於基準面u上的格子随及映在鏡子中 的格子圖案的鏡像。此處，圖像的拍攝並非僅進行一次， 藉由進減定的次數，例如五次，絲圖像資料的平均來 重新作為时:#料，亦可降仙制隨__影響。拍 束後，媳減z軸用光源1，將所拍攝的圖像保i於分 =處理部15 t。反覆進行自上述壓電平台14的移動至圖 ΪΐΓί為止的—連串的處理，直至達到事先所設定的相 移次數為if·。 圖像的拍攝，不改變基準 維格子圖像的拍攝與基準 緊接著上述z軸用格子投影 面的Z座標，連續進行xy軸用二 點圖像的拍攝。 即，首先 點冗叮軸用光源12,藉由第i相機6及 2相機7拍攝投影於基準面山上的二維的格子圖案及 光=子：的格子圖案的鏡像。拍攝結束後，熄滅xy軸用 先源Π，將所拍攝的圖像保存於分析處理部15中。 及第m’7點絲準點投光8，藉由第1相機6 及第2相機7拍攝投光於基準自山上的基準點及 鏡像。此處，與校正處理的情況同樣地， 3 進行—次，藉由進行規定的次數，例如 低:機的隨機均資料，亦:降 用雷:器8，將所拍攝的圖像保存於“ΐΐίίΤ!先 的圖像面lla的―個位置，可獲得校正所需 繼而’错由z轴移動平台14而使基準面lla僅移 29 201224392 動規定的大小後，重複上述所有處理。 _=將"1 上處理僅重複規定的次數（所規定的基準面 的面數），而兀成校正時所需的圖像的拍攝處理。 (校正處理） 圖23表不使用藉由圖22所示的處理所拍攝的圖 行校正，製作空間座標表格的處理的流程圖。 首先，藉由上述相移法，對圖22所示的處理中所拍攝 的對應於相移次數賴卵像紐她分 像素的相仂徭。 山口丁合 其次L於xy軸座標的相位分析前，於圖23的處理中， 根據在點亮基準點投光用雷射n 8陳態下所拍攝的圖 像，搜索圖像中的基準點的位置。 繼而，藉由例如傅里葉變換格子法’根據所拍攝的二 維格子圖像求出X軸及y軸的相位值。 其後，對所獲得的X軸及y軸的各相位值進行相位連 接，且相對於各方向轉換成相同的相位。此時，將所搜索 的基準點用於相位連接的中心點。 將上述所有處理僅重複對應於規定的基準面的面數的 次數’藉此可針對所有像素’求出基準面的各位置上的χ 方向、y方向及ζ方向的相位值。 求出各基準面位置上的相位值後，算出空間座標（x ，標、y座標及z座標）。關於空間座標，首先，z座標可 藉由如下方式求出：求出各像素中所需的表格位置的ζ軸 的相位值存在於藉由上述處理所求出的多個相位中的哪個 30 201224392 =二處理所求出的相位值與 =r:•軸二 正ί理。* 11作所指疋雜數的空間座標表格後完成校 格化方法一 流程^24表示對測量對象物0的形狀進行測量的處理的 首先，將測量對象物〇載置於基準面lla上，豆4， 對應於測量對象物0的高度，藉由Z轴移動平台14、= 基準面11a移動至任意的位置。 攝圖19(心〜圖19(6)所示的2轴用投影格子拍 攝f理同樣地，使壓電平台3移動至規定的位置，科此 H格2僅移動對應於相移次數的移動量二點 冗作為z轴用光源！的㈣，藉由第i相機6及 7拍攝投影於測量對象物〇上的格子圖_ 像保存於分析處理部15中。反覆進行自上述壓 Μ ===的"連串的處理’直至達到。事先 其後，根據所拍攝的對應於相移次數的圖像，藉由例 31 201224392 如相移法it行她分析，求料對各 nt圖22及圖23所示的處理所製作的二4 據所求出的相位值獲得空間座標，從而^ 最後’對多個形狀測量結果實施圖25所示的合成處 理’藉此可獲得一個形狀測量結果。 (合成處理） 為了好㈣像加以合成，必需事先搜索各圖像中成 為同-=置的-點。因此，與校正處理的情關樣地，首 先’點⑨基準點投光用f射!！ 8，其次拍攝投光於基準面 11a上的基準點的圖像，然後熄滅基準點投光用雷射器8 並將所拍攝的圖像保存於分析處理部15中。 其次，搜索所拍攝的圖像上的基準點。 繼而’算出所拍攝的所有圖像的相位評價值（PEV)。 其後，進行所拍攝的圖像是否為拍攝鏡像而成的圖像 的判定’於其為鏡像的航下進行使圖像反轉的圖像反 轉處理。 繼而’進行所拍攝的圖像是否為由第1相機所拍攝的 圖像的判定’於其並非由第丨相麟祕的圖像的情況 下，進行使圖像旋轉90。的處理。此處，旋轉9〇。是為了於 形狀測量裝置1GG巾’如圖！及圖4所示般配置第1相機 及第2相機’當第1相機與第2相機之間的配置關係不同 於形狀測量裝f _時’根據該配置關係，適當地進行圖 32 201224392 像的旋轉及旋轉處理。 繼而，對所有圖像實施上诚森4 xy座標再次配置於同-個座根系中，&理H則量的 ::==:低於規定_時捨棄;= 如此’藉由本發明的形狀測量方法，可同時自多 向對測量對象㈣形狀進行測量，因此可測定於自一 向的拍攝中無法獲得的測量對象物的背面側的形狀。 另外，藉由將利用多個相機所拍攝的圖像加以合 即便於測量對象财含有金屬的情況下、或者測量對象物 的=狀具有曲面的情況下’亦可降域暈的影響而高精度 地算出測量對象物的形狀。 以上’列舉具體例而對本發明進行了詳細說明，但本 領域從業人貞刊確得知只要不麟本發_專利申請的 範圍，則可進行任何變形或變更。例如，亦可進行構成。 因此’本發明並不限定於上述實施形態。 [產業上之可利用性] 根據本發明，可同時自多個方向對測量對象物的形狀 進行測量，另外，可降低光暈的影響，高精度地對測量對 象物的形狀進行測量，因此於電子零件的檢查、人體測量、 醫療用測量、以及小型生物的立體觀察或立體測量等時有 用0 33 201224392 —雖然本發明已以較佳實_揭露如上，财並非 發明♦任何熟習此技藝者’在不脫離本發明之精神 範圍當視後附之申請專利範圍所界定者為準。發月之保濩 【圖式簡單說明】 圖1，本發明的—實例的形狀測量裝置的立體圖。 圖2是本發明的—實例的形狀測量裝置的前視圖。 詳細^本發明的—實例的形狀測量裝置的構成的 圖4是本發明的—實例的形狀測量裝置的俯視圖。 禮2 5(心〜圖柳是表示投影有格子圖案的試樣及其鏡 课的圖。 圖6是表示設置有雜散光防止壁的鏡子的圖。 圖7是表示投影有用於形狀測量的一維格子圖案的基 準面及其鏡像的圖。 圖8是表示投影有全空間表格化方法中所使用的二維 格子圖案的基準面及其鏡像的圖。 圖9是針對使用-台相機與一面鏡子的情況的形狀測 量裝置的簡易立體圖。 圖10是針對使用-台相機與兩面鏡子的情況的形狀 測量裝置的簡易立體圖。 圖11 (a)是表示亮度分布的圖，圖u (b)是表示相 位分布的圖。 圖丨2是表示相移量與亮度的關係的圖。 34 201224392 圖13(a)〜圖13(b)是表示利用全空間表格化方法的形 狀測量的原理的圖。 圖14(a)〜圖14(c)是表示針對某一像素的相位值與χ 座標、y座標及z座標的對應關係的圖。 圖15 (a)是表示再抽樣前的相位分布資料的測量點 的圖，圖15 (b)是表示再抽樣後的相位分布資料的測量 點的圖。 · 圖16 (a)是表示使用測量結果的平均值時的圖，圖 16 (b)是表示利用本發明的合成處理的z座標分布的圖， 圖16 (c)是表示本發明的各z座標位置上的z座標分布 的圖。 圖17是表示作為測量對象物的試樣及其表面形狀的 圖。 圖18 (a)〜圖18 (d)是表示由兩台相機所拍攝的試 樣的實像及鏡像的圖像的圖，圖18 (e)是表示所合成的 圖像的圖。 圖19 (a)〜圖19(d)是表示針對由兩台相機所拍攝 的試樣的實像及鏡像的圖像的高度分布及相位評價值的 圖，圖19(e)是表示針對所合成的圖像的高度分布的圖。 圖20是產生了光暈時的測量對象物的圖像。 圖21(a)〜圖21(e)疋表示藉由本發明的一實例的形狀 測量裝置所獲得的形狀測量結果與所合成的圖像的圖。 圖22是表不杈>正處理中的拍攝圖像的流程圖的圖。 圖23是表不校正處理中的製作空間座標表袼的流程 35 201224392 圖的圖。 圖24是表示本發明的形狀測量方法的流程圖的圖。 圖25是表示本發明的形狀測量方法中的圖像的合成 處理的流程圖的圖。 【主要元件符號說明】 I : z軸用光源 2:2轴用格子玻璃/2軸格子玻璃· 3:第1透鏡 4:壓電平台 5:帕耳帖冷卻裝置 6 :第1相機 7 :第2相機 8:基準點投光用雷射器 9:第1鏡子 10 :第2鏡子 II :基準平板 11a :基準面 12 : xy軸用光源/xy軸用投影儀 12a : LED點光源 12b :針孔板 12c :第2透鏡 13 : xy軸用格子玻璃 13a : X軸用格子玻璃/x軸玻璃 13b : y軸用格子玻璃/y軸玻璃 36 201224392 14 : z軸移動平台 15 :分析處理部 16 :雜散光防止壁 26、36 :相機 30 :鏡子 100、200、300 :形狀測量裝置 I、Ι〇、Ιι、I2、I3、Ih、II :党度值 K :區間 L :直線/攝影線 Ο:測量對象物 0':測量對象物的鏡像 P、P〇、Pi、P2、…PN :點The Xy-axis light source! 2 is constructed as shown in Fig. 2, with a pinhole plate and a first-lightable 201224392 light. That is, the light irradiated from the π && source 12a from the τ p orifice plate 12b passes through the needle 光源 light source ’ to be substantially flat by the lens ,, and passes through the lattice glass 13 through the xy axis. (4) For H, the sub-surface 13 is used in the case of making a two-earth morning according to the full-space tabulation method, and the non-pattern is used to project the flat plate and the 11 on the reference plane m. The lattice glass 13 for the Xy axis is disposed directly below the reference. The xy-axis lattice glass 13 is as shown in FIG. 2, by two of the two-dimensional lattice patterns of the good knife; =3b) m:=: with the respective lattice patterns orthogonal to each other; The pattern-block plaid glass z-axis moving platform 14 is connected to the reference plate u, below the 12th. The z-axis moving platform, as shown in FIG. 2, moves the reference plane 110 in the direction of the x-axis. On: Single-axis platform for 1 Γ ο.1 Mm precision positioning of the object with feedback function ^ ', LZ moving platform 14t' and used for shape measurement, the orientation of the object (ie 'Z The coordinate adjustment processing unit 15 performs a phase analysis method (four)=object 0 shape on the mirror mirror image of the camera. Here, the phase analysis 2 = 疋 ' in the present example '(4) Wei_Shift' is previously stored for each pixel object = the associated space coordinate table of the work space. Thereby, if 201224392 = ri out the phase value of each pixel of the image, the empty shape can be obtained immediately. In addition, in the present example, a plurality of images taken by two phases are used. Therefore, the analysis and processing department is awkward. In the case of the image of the object 0 and the image of the mirror image, the image is imaged twice. The composition is appropriately rotated on the image of the camera or image mirrored by the camera. Or the reversal of the 'two nr', for example, by using the stipulations of the round stipulated below = two 'can also be used according to the hair ” ^ 乡 乡 乡 乡 乡 乡 乡 乡 乡 及 及 及 乡 乡 乡 乡 乡 乡 乡 乡Since the image force of the field of view of the camera is reflected on the mirror on the back side of the image, the object to be measured for the entire circumference of the object to be measured 0 is, for example, an electronic part. In the case where the metal part is included in the second part, then the measurement object has a curved surface. 'If the illumination is used, a halo may occur. However, if the camera is awkward; if the ground is 5, a halo is generated. Since the position is also different, the shape of the object to be measured can be accurately obtained by synthesizing the image of the different positions, thereby reducing the influence of the halo. The two-machine shape measuring device 100 is provided as a photographing unit ( The first camera 6 and the second camera 7), and the two mirrors (the first brother and the second mirror 10), The number of cameras and mirrors is not limited. 201224392 : For example, the shape machine (camera 26) shown in Fig. 9 can be used instead of the two cameras (one camera) with one phase as the second device 200. In the second shape measuring device 100 (mirror 30), instead of the two mirrors (machine 7)', one mirror is provided and the other configuration is set to the shape measuring mirror 9 and the second mirror 1), and the configuration and shape are The measuring device 1 is not the same as the device. Here, only the part of the = as shown in Fig. 10 may be attached. A camera (camera 36) is provided instead of the shape measuring device 300. (The first camera 6 and the second camera/), the two phase measuring devices 100 in the rigid device are the same, and the camera 36 can be configured such that the other configurations are mirror images. Shooting the object to be measured and the different parts of the three sets of 100 additional symbols 'only for the shape and shape measurement with the direction of the direction of the four or two: This can be obtained by using a camera to capture the same by using this shape measurement The effect of measuring the shape of the object to be measured IG 'Measuring object that can be obtained from multiple directions at the same time === One direction: When the shape of the object containing metal === has a curved surface, it is also possible to calculate the object to be measured. The shape of the object - the influence of the influence of 箪 (phase shift method) Here, the method for the phase analysis processing and the full table 3 in the present invention will be described. First, the phase shift method will be described. Fig. 11 (a) shows the luminance distribution of the lattice, and Fig. 11 (b) shows the phase distribution of the lattice. In addition, Fig. 12 shows the phase shift amount and the phase shift. A graph of the relationship of brightness changes. The luminance value I (x, y) of the lattice or interference fringes is generally distributed as a cosine wave in space (x, y) as shown in Fig. 12 (a). If it is expressed by the formula, it becomes as in the formula (1). ^{x,y) = ^x,y)cos(0(x>y)) + b(xyy) ( 1 ) where 'point (x,y) is a point within the captured image, a ( X, y) and b (x, y) represent the luminance amplitude and the background luminance, respectively, and Θ (x, y) represents the phase value of the lattice. In the case of photographing a grid image (hereinafter referred to as "lattice image"), the phase can be represented by a real number as a whole, but it can also be regarded as a repetition of 2π periods from 0 to 2π. Figure n(b) is a diagram showing the distribution of 0(x, y) as a repetition from 〇 to 2π. The phase shift method is a method of taking a plurality of grid images on the one hand and changing the phase of the grid by one cycle, and obtaining a phase distribution based on the obtained plurality of images. In all of the pixels, the luminance changes by one cycle. Therefore, the phase value can be obtained independently of each point, i.e., without changing the redundancy of the pixel to be closed, based on the change in luminance. Therefore, the method is an effective method for measuring the shape of an object having a step or discontinuity. Here, the case where the phase value is obtained from the four luminance values obtained by the most commonly used phase shift of π/2 (that is, the case where the number of phase shifts is four times) is taken as an example, and the principle of the phase shift method is applied. Be explained. 201224392 When the phase shift sa is added to the equation of the luminance distribution of the grid, the equation (2) is obtained. I(x,y,a) = a(x, cos(0(x, ;;) + «) + b(x, y) (2) Figure 12 shows the phase of a point (pixel) with an initial phase θ The relationship between the brightness changes. The initial phase means that the phase shift amount is the phase of the grid. If the phase shift amount is changed from 〇/2 to 亮/2, the bright households are set to I〇, Ii, I2, and Ι3. Then, these can be expressed as in the equations (3) to 卞, and the description of (x, y) is omitted in the following equation. "6) I0(xyy) = acos6-\-b (3 ) '1 (also ' >0 = ac〇s [^ + faint) + 6 (4) Eight (x, less) = β c〇s0 + π) + δ (5) /3 〇χ, 3〇 = a cos(0 + specific)+ 6 (6) The following formulas (7) and (8) can be obtained from these formulas. —/〇= —2acos0 (7) (8) (8)201224392 h =2asin^ Further, the following equation (9) can be derived from equations (7) and (8), and the phase is obtained from the relational expression. The value θ. That is, if the luminances l〇, L, ^, and 时 when the phase shift amounts are 〇, f, π, and 3π/2 are obtained, the phase value Θ for the pixel is obtained. J2~h (9) Here, the influence of random noise of the camera can be reduced by the number of times of multi-phase shift (i.e., the number of degrees from 〇 to 271). When the number of phase shifts is set to N and the luminance when the phase shift amount is 27 lk/N is set to lk, the equation (10) can be derived, and the phase value θ can be obtained from the relational expression. =-tan0 do) ^. In this way, the phase value of each pixel on the image can be obtained by the phase shift method (correction method). Calculate the phase value (t, y, z) obtained for each pixel by the phase shift method. a method of measuring the shape of an object 20 201224392 * In the invention of t, the table i used in the conversion is obtained by using a transformation formula that produces a phase value corresponding to a space coordinate as in the prior art. In the three types of the present invention, the phase value of the lattice projected by the pixel f for each pixel of the camera and the space coordinate are used _ and the square of the second is given, 'this is called correction (caiibrati〇n). If the phase value of the grid projected on the object to be measured is obtained by the school ^ =, the calculation of the coordinates of the unconverted type can be obtained by referring to the obtained gate; r ΐ 0 ί, the table, and the object can be immediately evaluated. The space of the surface is the seat, that is, the shape of the object to be measured. (Full space tabular method) Patent ==== method (for example, refer to the 曰本图, so that the vertical t amount: the original reference plane is in the direction of the x-axis direction) and the instrument is fixed in advance. Multi-motion on the datum. The camera and projection are on the reference plane. At this time, the grid pattern is not projected by the special image projector = the point on the straight line L of one frame. For the == picture R2, ... Rn of the picture 13(b), the positions Pq, p ', and 'p base plane position Ro, heart, value θ 〇, θ!, Θ 2, . By phase two fN. The phase of each point Figure 13 (b) shows the relationship between the coordinates (X, y, Z). The point f is taken by Hi and the pixels 21 on the reference plane. The points on the surface of the 2012 21392 face are P〇, p, Pi, .. D, ' a · ΡΝ, self-projected from the reference surface. The 4 sub-patterns of the Weng X chain and the y coordinate, L for the (4) silk, can be used to obtain the x-direction and the y-direction connection by the two ^^ transformation lattice method, respectively, to obtain the respective _x coordinates and ϋ (for example, 'refer to Japanese patents Bulletin No. 918). Each position relative to the two ΐ ΐ 可 can be 细 相对 ====================================================================================== Interval, and interpolating between them, the spatial coordinates relative to all phase values can be obtained by 咼5. By the above positive method, the phase value θ〇1 °2 can be obtained for each pixel separately. The coordinates Xq, χι, χ2, "·Χν, y coordinates y〇, (a) ~ Figure 14 (c) show the relationship by black dots. yi, y2, '^ and z coordinates z〇, Ζι The relationship between z, z, and Zn is smoothly changed with respect to the x coordinate, the y coordinate, and the z coordinate of the phase value Θ in Fig. 14. The phase value θ _ is divided into regions, and the divisional line is interpolated to obtain the corresponding The X coordinate, the y coordinate, and the z coordinate are used to create a space coordinate table. In the case of the lattice projection method, in a general configuration, the phase = coordinate relationship with the coordinate becomes a smooth curve, so linear interpolation can be performed. And expect sufficient accuracy, of course, can also perform high-order interpolation. In Figure 14, linear interpolation A space coordinate table is created in the form of a point on a straight line connecting black points. For the section indicated by κ in Fig. 4, the correlation between the phase value and the space coordinate can be obtained by extrapolation. 22 201224392 Thus, According to the full space tabular method, the range and space coordinates U coordinate, y coordinate and z coordinate) 2 phase value θ table. Finally (four) in the entrance 袼 袼 _ = ^ seat can also be based on (four) (four) solution, connect her After her disobedience brothers' relationship table. By phase connection, the material can be expanded (measurement of the shape of the object to be measured), the surface is first rm乂, and the object to be measured is set to the reference map. 2%, the grid with the same pattern of the dice in the processing of the right:: The pixel on the object to be measured is the shadow of the object 图案 = pattern ^ ep can be converted to _ bit = ^ grid on the point by reference to the pre-made The space coordinate table, and the table obtained by (4) can be immediately obtained by the y coordinate and the scorpion mark corresponding to the target. Similarly, immediately after the second: :r堇 reference to the pre-production of ...^^ Space coordinate form, which can be used to obtain space from the object The measurement results of the measurement can be measured at high speed for the camera (re-sampling process). The shape measurement obtained by the method of the above-described shape-measured pixel can be obtained as the coordinate distribution of the point, so as shown in Fig. 15 23 201224392 ---X-- (a) is obtained. (4) m in the xy plane scale (four) off-coordinate distribution Saki Fu Qiao 4 is inter-coordinated by the image of the η or real image and the mirror image placed at different positions. , the empty point S' of the point sampled as follows, and the 3 pieces of the point Ps that know the X coordinate and the ^ point, and the person will pass the plane of the 3 points and self-extraction; & the sampling point The z coordinate. The intersection of the === line is re-sampling in the above manner as another field, as shown in the figure, in addition to the point where the xyif face is sampled at equal intervals, in the shape of an example of the present invention. In the measuring device 100, the two cameras are arranged such that the imaging directions of the two cameras are orthogonal to each other in the Z direction. Therefore, in order to combine the image captured by the second camera with the image captured by the second camera, Must implement only rotation on _ images The 90° rotation process is performed, and in each image captured by each camera, image inversion processing is performed on the mirrored image. When the arrangement relationship between the i-th camera and the second camera is different from that of the shape measuring device, the image rotation and rotation processing are appropriately performed in accordance with the arrangement relationship. (Image Synthesis Method According to Phase Evaluation Value) The shape measurement results from a plurality of directions are performed based on the f-coordinate ' of the image subjected to the above-described = sampling processing, image inversion processing, and/or image rotation processing Synthesis of _. Here, the synthesis processing of the ® image can be performed by synthesizing height information (z coordinate) for each pixel. For example, the image data can be simply weighted and purely averaged. The image having the reliability of $1 is the image of the plurality of images captured by the determination; the phase evaluation value of the index of the map is 俨i, and the obtained value is also infinite; In the present invention: each; the butterfly phase evaluation value Fourier transform, and as the power of the -second frequency ω1 is ambiguous): In the case, when the luminance of the luminance, the amplitude or the amplitude of the luminance is small (ie, the change in luminance) Hour):= Two materials: When the random noise is included, the phase evaluation value becomes smaller. Therefore, the phase evaluation value is a value that can confirm the reliability of the phase value. When the image is synthesized, it can be used as an index for the selection of the image data for synthesis. All parties: The image synthesis process is to obtain the information of the enthalpy (position in the Ζ direction) from each direction for two pixels: r=r_ except...: In the second month of February, the following equation (8) is used for phase evaluation. The value E(6) is ten. Here, F (ωη) represents the power frequency of the frequency ωη, and Ν represents the number of times corresponding to the highest frequency. (11) 25 201224392 In this way, in the flat pixel, the pixel is only synthesized by the equation (11): the phase can be obtained by subtracting the yield (Z is the reference surface as the measurement object) The number of changes is high. The number of Si:: faces is used to measure the shape of the reference plane. Here, the position of the Z axis of the reference surface f 16(e) is 500 μm Θ (that is, layout). 16(4) shows an average of all the image data of the image of the real image of the second camera and the image of the second camera and the image of i, and the method of the present invention is based on the phase evaluation value=1. The result obtained by synthesizing the image is as follows. According to the above = (IV) I, it is known that the measurement process is improved by the synthesis process of the present invention, and the measurement accuracy is improved. w 16 (4) indicates the position of the z-station in which the synthesis process of the present invention is carried out. "Wrap, can be known at any position." The measurement accuracy is improved. Next, the shape measurement result of the sample including the superhard alloy shown in Fig. 17 is explained. The surface of the sample is used first. Universal projector (V-12BS type Ν〇·13〇〇 21) and the height/thickness gauge _-5 〇 〇 523) The step of the table of Figure 17 measured, with the advancement from U) to (7), each increase by about 10 qing. This 26 201224392 at 'this The shape measurement result of the step is paid attention to. In addition, the correction and measurement 设定 sets the number of phase shifts to 16 times, and continuously shoots five images and averages the image data as described above to reduce the influence of random noise. 18(a) to 18(e) are diagrams showing a distribution map of the measured height coordinates of the entire sample. Further, Fig. 19(a) to Fig. 19(e) are based on Fig. 18(a)~ The result shown in FIG. 18(e) is a view showing the position i and the height (z coordinate) distribution in the X direction at the position j shown in FIGS. 18(a) to 18(e). Fig. 19 (a) shows an image of a real image of the first camera, Fig. 19 (匕) shows an image of a mirror image of the first camera, and Fig. 19 (c) shows an image of a real image of the second camera, Fig. 19 (d) An image showing the mirror image of the second camera, and Fig. 19(e) shows the synthesized image. Here, the threshold value of the phase evaluation value at the time of composition is 15. As shown in Fig. 19 (a) to Fig. 19 (d) As shown, it can be seen that shooting from four directions As shown in Fig. 19(e), as shown in Fig. 19(e), the height unevenness measurement accuracy is improved by the image composition processing of the present invention. The image of the object to be measured of the object of the measurement is as follows. As shown in the figure, the halo is generated, and it is impossible to measure the shape of the object 0 according to the figure. 1 』本图图) shows the image from the 40% of the shape measurement device of the present invention having two cameras and two mirrors. Figure 21 (a) ~ Figure 21 (4) 矣 二 摄 ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; 1 1 1 1 1 27 1 1 27 27 The resulting image. As is clear from the images of FIGS. 2i(a) to 21(d), the height distribution of each image is affected by halation and contains noise, but is synthesized as shown in FIG. 21(e). In the image, it is known that the effect of halo is greatly reduced. According to the shape measuring apparatus 1 of the present invention, even when the object to be measured contains metal and a halo is generated, the shape of the object to be measured can be measured with high precision. [Shape Measuring Method] Next, the shape measuring method of the present invention will be described. The shape measuring method of the present invention is an example in which the phase shift method according to the full space tabulation method is used in the phase analysis processing, but the present invention is not limited thereto. When the phase shift method according to the full space secretization method is used, it is necessary to make a correction for the associated space of the phase value and the space coordinate for each pixel in advance, and save the obtained space coordinate table. Therefore, ::Z-axis and xy-axis on multiple reference plane positions: = light image. First, the correction method of the correction method, the photographing grid = /, and the correction method of the reference point projection image will be described. (Photographing processing of the image for correction) Quasi: The grid projection image and the Z-axis are moved to the predetermined position by the X-ray 1st electric dry 〇 3, and then lit as the Z-axis light source by the first 1 Camera 6 and 28th 201224392 2 Camera 7 reduces the mirror image of the grid pattern projected on the reference plane u and reflected in the mirror. Here, the image is taken not only once, but by the number of times of reduction, for example, five times, the average of the silk image data is used as the time: #料, and can also be reduced with the influence of __. After the shot is taken, the z-axis light source 1 is subtracted, and the captured image is saved in the sub-processing unit 15 t. The series of processes from the movement of the piezoelectric stage 14 to the image is repeated until the number of phase shifts set in advance is asif. Image capture without changing the reference dimension The grid image is captured and referenced. Immediately following the Z coordinate of the z-axis grid projection surface, the xy-axis two-point image is continuously captured. That is, first, the axis light source 12 is redundantly used, and the i-th camera 6 and the camera 7 capture a mirror image of a two-dimensional lattice pattern projected on the reference plane mountain and a lattice pattern of light=sub. After the end of the shooting, the xy axis is turned off, and the captured image is stored in the analysis processing unit 15. And the m'7th point punctual projection light 8 is used to capture the reference point and the mirror image projected from the mountain by the first camera 6 and the second camera 7. Here, in the same manner as in the case of the correction processing, 3 is performed once, for example, by a predetermined number of times, for example, a random: random data of the machine, and also by using the lightning: device 8, the captured image is saved in " ΐΐίίΤ! The first position of the image plane 11a can be corrected. Then, the error is repeated by the z-axis moving platform 14 and the reference plane 11a is only moved by the size specified by 201224392. Repeat all the above processing. <1 The processing is repeated only for a predetermined number of times (the number of planes of the predetermined reference plane), and the image processing required for the correction is made. (Correction processing) FIG. 23 is not shown by FIG. A flowchart of processing for creating a spatial coordinate table by processing the captured image line. First, by the phase shift method described above, the number of phase shifts corresponding to the number of phase shifts captured in the processing shown in FIG. The phase of the pixel. Before the phase analysis of the Yamaguchi joint L and the xy axis coordinate, in the processing of Fig. 23, the image is taken according to the image taken by the laser light at the lighting reference point. The position of the fiducial point in the image. Then, by way of example The Fourier transform lattice method 'determines the phase values of the X-axis and the y-axis based on the captured two-dimensional lattice image. Then, the phase values of the obtained X-axis and y-axis are phase-connected, and The direction is converted to the same phase. At this time, the searched reference point is used for the center point of the phase connection. All the above processes repeat only the number of times corresponding to the number of faces of the specified reference plane 'by this, for all pixels' The phase values in the χ direction, the y direction, and the ζ direction at each position of the reference plane are obtained. After obtaining the phase values at the positions of the reference planes, the space coordinates (x, the standard, the y coordinate, and the z coordinate) are calculated. First, the z coordinate can be obtained by determining which of the plurality of phases obtained by the above-described processing is obtained by the phase value of the paraxial axis of the desired table position in each pixel. 201224392 = 2 processing The obtained phase value and =r:•axis is positive. *11 is used as the space coordinate table of the indicated number of impurities to complete the calibration method. A process ^24 indicates the measurement of the shape of the object 0 to be measured. First of all, the object to be measured The 〇 is placed on the reference plane 11a, and the bean 4 is moved to an arbitrary position by the Z-axis moving platform 14 and the reference plane 11a corresponding to the height of the measuring object 0. Photograph 19 (Heart ~ Figure 19 (6) Similarly, the two-axis projection grid is used to move the piezoelectric stage 3 to a predetermined position, and the H-frame 2 only moves the movement amount corresponding to the number of phase shifts as a z-axis light source! (4) The grid image projected onto the object to be measured by the i-th cameras 6 and 7 is stored in the analysis processing unit 15. The "continuous processing" from the above-mentioned compression === is repeated until reaching In the following, according to the image corresponding to the number of phase shifts taken, by example 31 201224392, the phase shift method is used to analyze her, and the second processing is performed for each nt FIG. 22 and FIG. 4 The spatial coordinates are obtained based on the obtained phase values, so that finally the 'synthesis process shown in Fig. 25 is performed on the plurality of shape measurement results', whereby a shape measurement result can be obtained. (Synthesis processing) In order to synthesize (4) images, it is necessary to search for a point which becomes the same -= in each image. Therefore, in the case of the correction processing, the first ‘9 points of the reference point projections are shot with f!! 8. The image of the reference point projected on the reference surface 11a is imaged next, and then the reference point light projecting laser 8 is turned off and the captured image is stored in the analysis processing unit 15. Second, search for the reference point on the captured image. Then, the phase evaluation value (PEV) of all the captured images is calculated. Thereafter, a determination is made as to whether or not the captured image is an image obtained by capturing an image. The image is inverted by the image in which the image is inverted. Then, the determination of whether or not the captured image is the image captured by the first camera is performed is rotated by 90 when the image is not the same. Processing. Here, rotate 9〇. It is for the shape measuring device 1GG towel' as shown! As shown in FIG. 4, the first camera and the second camera 'when the arrangement relationship between the first camera and the second camera is different from that of the shape measuring device f_', according to the arrangement relationship, the image of FIG. 32 201224392 is appropriately performed. Rotate and rotate. Then, for all images, the Chengsen 4 xy coordinates are again placed in the same root system, & H is the quantity::==: is lower than the prescribed _ discard; = so 'by the shape of the present invention According to the measurement method, the shape of the measurement object (four) can be measured from the multi-directional direction at the same time. Therefore, the shape of the back side of the measurement object that cannot be obtained from the one-shot imaging can be measured. In addition, when the image captured by the plurality of cameras is combined, even when the measurement target contains a metal, or when the measurement object has a curved surface, the surface of the measurement object can also be highly accurate. The shape of the object to be measured is calculated. The present invention has been described in detail above with reference to specific examples. However, it is understood by those skilled in the art that any changes or modifications may be made without departing from the scope of the invention. For example, it is also possible to configure. Therefore, the present invention is not limited to the above embodiment. [Industrial Applicability] According to the present invention, the shape of the object to be measured can be measured from a plurality of directions at the same time, and the influence of the halation can be reduced, and the shape of the object to be measured can be measured with high precision. It is useful for inspection of electronic parts, anthropometrics, medical measurement, and stereoscopic observation or stereo measurement of small creatures. 0 33 201224392 - although the present invention has been disclosed above, it is not the invention ♦ anyone skilled in the art' The scope of the patent application is defined by the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a perspective view of a shape measuring apparatus of an example of the present invention. Figure 2 is a front elevational view of the shape measuring device of the present invention. DETAILED DESCRIPTION OF THE INVENTION Fig. 4 of the present invention is a plan view of a shape measuring apparatus of an example of the present invention. Ceremony 2 5 (heart ~ graph Liu is a diagram showing a sample in which a lattice pattern is projected and a mirror thereof. Fig. 6 is a view showing a mirror provided with a stray light preventing wall. Fig. 7 is a view showing a projection for shape measurement. Fig. 8 is a view showing a reference plane on which a two-dimensional lattice pattern used in the full-space tabular method is projected and a mirror image thereof. Fig. 9 is a view of the use of a camera and one side. Fig. 10 is a simplified perspective view of a shape measuring device for a case where a camera and a double mirror are used. Fig. 11 (a) is a view showing a luminance distribution, and Fig. 11 (b) is a view Fig. 2 is a diagram showing the relationship between the phase shift amount and the luminance. 34 201224392 Figs. 13(a) to 13(b) are diagrams showing the principle of shape measurement using the full space table method. 14(a) to 14(c) are diagrams showing the correspondence relationship between the phase value of a certain pixel and the 座 coordinate, the y coordinate, and the z coordinate. Fig. 15 (a) shows the phase distribution data before resampling. Figure of measurement points, Figure 15 (b) shows re-sampling Fig. 16(a) is a view showing an average value of measurement results, and Fig. 16(b) is a view showing a z coordinate distribution by the synthesis processing of the present invention, Fig. 16 (c) is a view showing the z coordinate distribution at each z coordinate position of the present invention. Fig. 17 is a view showing a sample as a measurement object and its surface shape. Fig. 18 (a) to Fig. 18 (d) A diagram showing a real image and a mirror image of a sample taken by two cameras, and Fig. 18(e) is a view showing a synthesized image. Fig. 19 (a) to Fig. 19 (d) show FIG. 19(e) is a view showing a height distribution of a synthesized image, and FIG. 19(e) is a view showing a height distribution of a real image and a mirror image of a sample taken by two cameras. FIG. An image of the object to be measured at the time of halation. Fig. 21 (a) to Fig. 21 (e) are views showing a shape measurement result obtained by the shape measuring device of an example of the present invention and a synthesized image. It is a diagram of a flowchart of a captured image in the process of being processed. Fig. 23 is a diagram showing the coordinates of the production space in the process of the correction process. Fig. 24 is a view showing a flowchart of a shape measuring method according to the present invention. Fig. 25 is a view showing a flowchart of a method of synthesizing an image in the shape measuring method of the present invention. Description of component symbols: I: Light source for z-axis 2: Lattice glass for 2-axis, 2-axis lattice glass · 3: First lens 4: Piezo platform 5: Peltier cooling device 6 : First camera 7 : Second camera 8: Reference point projection laser 9: First mirror 10: Second mirror II: Reference plate 11a: Reference surface 12: xy-axis light source / xy-axis projector 12a: LED point light source 12b: pinhole plate 12c: 2nd lens 13: Lattice glass 13a for xy axis: Lattice glass/x-axis glass 13b for X-axis: Lattice glass/y-axis glass 36 for y-axis 201224392 14: Z-axis moving platform 15: Analysis processing unit 16: Miscellaneous Astigmatism preventing walls 26, 36: Camera 30: Mirrors 100, 200, 300: Shape measuring device I, Ι〇, Ιι, I2, I3, Ih, II: Party value K: Interval L: Straight line/photographic line Ο: Measurement Object 0': image of the object to be measured P, P〇, Pi, P2, ... PN: point
Pa 、 Pb 、 Pc :點Pa, Pb, Pc: point
Ps :抽樣點 R〇、R4、R2、…Rn :基準面的位置/基準面 X、y、z :軸 x〇、χι、x2、·，·χΝ : X 座標 y〇、yi、y2、…yN : y 座標 Z〇、Z!、Z2、一ZN : Z 座標 Θ :初始相位/相位值 θ〇、、02、…θρ :相位值 37Ps : sampling point R 〇, R4, R2, ... Rn : position of the reference plane / reference plane X, y, z: axis x 〇, χι, x2, ·, · χΝ : X coordinates y 〇, yi, y2, ... yN : y coordinates Z〇, Z!, Z2, a ZN: Z coordinate Θ: initial phase/phase value θ〇, 02, ... θρ: phase value 37