JP4087146B2

JP4087146B2 - Shape measuring method and shape measuring apparatus

Info

Publication number: JP4087146B2
Application number: JP2002127753A
Authority: JP
Inventors: 照己鎌田; 隆二崎田; 攻中山
Original assignee: Ricoh Co Ltd
Current assignee: Ricoh Co Ltd
Priority date: 2002-04-30
Filing date: 2002-04-30
Publication date: 2008-05-21
Anticipated expiration: 2022-04-30
Also published as: JP2003322516A

Description

【０００１】
【発明の属する技術分野】
この発明は、例えば円筒状被検物等の表面形状や、キズや膨らみ，うねり，へこみ等の欠陥を検出する形状測定方法及び形状測定装置に関するものである。
【０００２】
【従来の技術】
感光体ドラムなどの円筒状被検物の欠陥検査方法として、例えば特開平２−201142号公報及び特開平４−169840号公報が開示されている。特開平２−201142号公報に示された異常検出方法は、図１７に示すように、光源３１からのレーザ光ビーム３２を回転多面鏡３６を介して感光体ドラム３３の軸方向に走査するように照射し、この走査光はドラム３０の感光層表面にて反射し、正常な表面からの反射光は、ほぼ受光器３５に進入して反射光の強度が検出され、受光器５の出力は所定の演算処理部等に入力される。ここでの処理で検出値が異常に低下したときに、表面状態の異常として検出する。また、特開平４−169840号公報に示された円周表面傷検査方法は、図１８に示すように、ハロゲン光源等を備えた投光器４１から感光体ドラム４３へ向けてスリット光４２を投射し、感光体ドラム４３の表面欠陥によって散乱された散乱光をレンズ４４によって集光してラインセンサ４５で受光して欠陥による散乱光により異常を検出している。
【０００３】
【発明が解決しようとする課題】
感光体にはピンホールや打痕，擦り傷，気泡の巻き込み，クラック，ゴミ等の付着による欠陥及び感光層の膜厚のムラや液ダレ，支持体の傷等多種多様な欠陥が生ずる可能性がある。上記のような光学式検査装置による場合は、ピンホールや打痕，擦り傷，ゴミ等の付着による欠陥のように、表面の凹凸の変化率の大きな欠陥は精度良く検出できるが、感光層の膜厚ムラ等の如く凹凸の変化率の小さい欠陥あるいは支持体の傷のように感光体表面に凹凸の変化のない欠陥に対しては検出精度に問題があった。
【０００４】
一方、三次元測定法の１手法としてモアレ法が挙げられる。モアレ法には、実体格子型と格子投影型があり、様々な分野において広く利用されている。格子投影型のモアレ法とは、図１９に示すように、投影用と観察用とに、それぞれ小さな格子Ｇ１，Ｇ２を配置し、Ｇ１をレンズＬ１により物体に投影し、物体形状に応じて変形した格子線をレンズＬ２を通じてもう一つの格子Ｇ２上に結像させ、縞等高線を基準面から所定距離のところに生じさせるようにしたものである。実体格子型のモアレ法は、図２０に示すように、基準面に一つの格子Ｇを設置し、図２１に示すように、レンズＬ１の位置に点光源Ｓ１を置き、レンズＬ２の位置に観察眼ｅを置いて、格子Ｇの光源Ｓ１による影を物体上に落し、物体形状に応じて変形した格子Ｇの影を形成させ、これを格子Ｇを通して観察し、格子Ｇと変形した格子の影とによって生じるモアレ縞を観測する方法である。
【０００５】
この実体格子型のモアレ法をさらに詳しく説明する。図２１に示すように、光源Ｓ１及び観測点Ｓ２と物体Ｏとの間の同一平面に格子Ｇ１,Ｇ２を配置し、光源Ｓ１と観察点Ｓ２の距離をｄ、光源Ｓ１及び観察点Ｓ２から格子Ｇ１，Ｇ１までの距離をＬ、格子Ｇ１，Ｇ２から物体Ｏまでの距離をｈ、格子Ｇ１，Ｇ２はいずれもピッチｓをもつが、格子Ｇ１、Ｇ２は面内で互いにεだけずれている（格子ピッチの位相でいえば２πε／ｓ）とすると、下記（１）式で表せる。
cos(２π/ｓ)・[{ｄｈ−ε(ｈ＋Ｌ)}／（ｈ＋Ｌ）] （１）
形成されるモアレ縞（等高線）は、格子面を基準（０次）として、格子面から離れるに従い、順に１次、２次とカウントされる次数を持つ。そこで縞次数Ｎのモアレ縞をcos２πＮと置くことによって得られる。その結果、第Ｎ次のモアレ等高線は基準面からｈ_Ｎだけ離れた下記（２）式で示される位置に形成される。
ｈ_Ｎ＝｛（Ｎｓ＋ε）Ｌ｝／（ｄ−Ｎｓ−ε）（２）
これは位置の座標ｘを含んでおらず、縞次数Ｎによって定める固有の値となっている。すなわち等高線が形成されていることを示す。
【０００６】
図２２に示す実体格子型のモアレ法は、図２１に示す格子Ｇ１，Ｇ２を１枚の連続した格子Ｇとしたものに相当し、ε＝０となる。（２）式から下記（３）式が成り立つ。
ｈ_Ｎ＝（ＮｓＬ）／（ｄ−Ｎｓ）（３）
但し、等高線といいながら、その間隔Δｈ_N＝ｈ_N+1−ｈ_Nは一定ではなく、次数Ｎによって異なってしまう。
【０００７】
従来、モアレ法による三次元形状測定法は対象物を直観的に把握することはできるが、（１）凹凸の判定がし難い、（２）高感度の三次元測定には不向き（現時点ではモアレ縞等高線の間隔は10μｍ程度が限界とされている）、（３）モアレ縞のビジビリティーが縞ごとに均一でないためモアレ像を画像処理の対象として扱いにくい等々の問題が指摘されている。この問題は、格子投影型の場合、２枚の格子を利用しているために、その一方を移動させて縞走査すなわちモアレ縞の位相をシフトさせることによって、等高線間隔を等価的に細かく分割するとともに対象の凹凸判定や測定感度の向上が可能である。
【０００８】
この位相シフト法の原理を説明する。位相変調された縞画像は、図２３に示すように、バイアスをａ、振幅をｂ、操作可能な位相をθ、高さに相当する位相値をΦとすると、下記式で表せる。
Ｉ＝Ｉ(θ)＝ａ(ｘ，ｙ)＋ｂ(ｘ，ｙ)cos{Φ(ｘ、ｙ)＋θ}
ここで求めたいのは各点（ｘ、ｙ）における位相Φ（ｘ、ｙ）である。バイアスａや振幅ｂは、表面の反射率や汚れなどで変化する未知数成分なので、位相θを０、π／２、πと変化させた下記式で示される３つの縞画像を生成する。
Ｉ＝Ｉ(０)＝ａ(ｘ，ｙ)＋ｂ(ｘ，ｙ)cos{Φ(ｘ、ｙ)＋θ}
Ｉ＝Ｉ(π/２)＝ａ(ｘ，ｙ)＋ｂ(ｘ，ｙ)cos{Φ(ｘ、ｙ)＋θ}
Ｉ＝Ｉ(π)＝ａ(ｘ，ｙ)＋ｂ(ｘ，ｙ)cos{Φ(ｘ、ｙ)＋θ}
そして下記（４）式で位相Φ（ｘ、ｙ）を算出すれば反射率や汚れ成分を除去して、各点の位相Φ（ｘ、ｙ）を求めることができる。
Φ(ｘ、ｙ)＝tan^−１｛(Ｉ３−Ｉ２)／(Ｉ１−Ｉ２)｝＋π／４（４）
【０００９】
しかし、実体格子型の場合には格子Ｇが一枚であるため、格子投影型のモアレ法のような位相シフトを行っても、すべての次数の縞等高線の位相を揃えながら位相を変えることはできない。このような問題点に対して、例えば特許第2887517号公報に示す高感度三次元測定法では、格子面の垂直移動と光源又は観察点の水平移動を同時に行うことにより、各次数のモアレ縞の位相にほぼ大きな変化をきたすことなく、各次数の縞の位相がほぼ揃った状態で測定対象に対する縞位相のシフトができるので、複数枚の縞画像から位相シフト法の原理に基いて処理することができ、これによって測定対象に対するモアレ縞による測定点の密度が増大するとともに、モアレ縞１周期について約1/40〜1/100程度の物理的な分割が可能となり、実体格子型のモアレ法では困難とされていた面の凹凸の判定や測定感度の向上を図ることができる。
【００１０】
このように位相シフト法を適用して円筒状被検物等の全面測定を行う場合、少なくとも被検物を３回転以上させて位相シフトさせるために格子移動とモアレ縞の撮像を繰り返す必要があり測定に時間がかかる。また、格子を複数方向（平行と回転）に移動させる必要があるため、装置構成が複雑になる等の問題がある。
【００１１】
特開平７−332956号公報に示された表面形状測定装置や文献「位相シフトによる実体格子型モアレ法」（1991年度精密工学会秋季大会学術講演会講演論文集）、「液晶ガラスのフラットネス計測」（O plus E 1996年9月）では、平行光を与えることにより、縞次数による縞間隔の違いをなくしているため、全ての縞の位相を揃えながらシフトさせている。これらの方法では格子運動のみにより位相シフトさせることが可能である。
【００１２】
しかし依然として円筒状被検物等の全面測定を行う場合、位相シフトした画像を得るために、格子移動と撮像という動作を繰り返し被検物を３回転以上させる必要があるため測定時間の増大を招く。また、特開平10−54711号公報に示された表面形状測定方法では、被検物の高さを変えることにより位相シフトさせている。この場合においても被検物の移動と撮像を複数回繰り返す必要があるため測定時間の増大を招く。また、凹凸形状の定量化に関しては明確な方法が充分に説明されてない。
【００１３】
この発明は上述した問題点を解消し、ローラ部品等の円柱状被検物を対象とし、実体格子型のモアレ法に位相シフト法を適用し、１系統の画像撮像装置による１回の１連の撮像により位相シフトした画像を得ることにより、高速に形状測定を行い、その定量的な形状データから被検物表面の検査を行うとともに被検物の移動量誤差を計測することができる形状測定方法及び形状測定装置を提供することを目的とするものである。
【００１４】
【課題を解決するための手段】
この発明に係る形状測定方法は、被検物を回転させ、２次元エリアセンサを有する実体格子型のモアレ光学系により前記被検物にモアレ縞を投影し、前記被検物に投影されたモアレ縞が形成される領域とモアレ縞が形成される領域以外の領域を前記２次元エリアセンサで撮像し、前記２次元エリアセンサによって撮像された画像におけるモアレ縞が形成される領域の特定の縞次数のモアレ縞を所望の位相だけシフトさせてモアレ縞データを生成し、生成したモアレ縞データから前記被検物の形状測定を行い、前記２次元エリアセンサによって撮像された画像におけるモアレ縞が形成される領域以外の領域の画像から被検物の位置変動を検出することを特徴とする。
【００１５】
前記形状測定中に位相シフトしたモアレ縞データと同一画像のモアレ縞の発生していない領域に規則性のあるパターン光を投影し、その位置又はパターン形状から被検物の位置変動を検出したり、位相シフトしたモアレ縞データと同一画像のモアレ縞の発生していない領域に形成されたモアレ縞を形成するための格子パターン端部のモアレを形成しない部分のパターンの変化から被検物の位置変動を検出して、位置変動検出用のセンサを使用せずに、位相シフトしたモアレ縞データと同一画像を画像処理することにより、形状測定中の位置変動を検出する。
【００１６】
また、形状測定中に位相シフトしないモアレ縞画像から被検物の位置変動を検出しても良い。
【００１７】
さらに、被検物の位置変動検出データに応じて形状測定に利用する位相シフト算出領域を変更して位相シフト演算を行い、位相シフト誤差を低減させる。
【００１８】
また、被検物の位置変動に応じて測定ヘッドと被検物の位置関係が一定になるように制御して、回転振れの影響を抑制し位相シフト誤差を低減させる。
【００１９】
この発明に係る形状測定装置は、２次元エリアセンサを有する実体格子型のモアレ光学系を有する測定ヘッドと、被検物を保持して回転させる把持回転機構とを有し、前記被検物を回転させ、前記実体格子型のモアレ光学系により前記被検物にモアレ縞を投影し、前記被検物に投影されたモアレ縞が形成される領域とモアレ縞が形成される領域以外の領域を前記２次元エリアセンサで撮像し、前記２次元エリアセンサによって撮像された画像におけるモアレ縞が形成される領域の特定の縞次数のモアレ縞を所望の位相だけシフトさせてモアレ縞データを生成し、生成したモアレ縞データから前記被検物の形状測定を行い、前記２次元エリアセンサによって撮像された画像におけるモアレ縞が形成される領域以外の領域の画像から被検物の位置変動を検出することを特徴とする。
【００２０】
【発明の実施の形態】
図１はこの発明の形状測定装置の構成を示す斜視図である。図に示すように、円筒状被検物１の形状や欠陥を検出する形状測定装置２は、例えば電子写真方式の画像形成に使用する感光体等の円筒状被検物１を固定する把持冶具３と、把持冶具３を回転する回転モータ４と、自動ステージ５に設けられた測定ヘッド６を有する。把持冶具３は、例えば三つ爪チャック等を有し、円筒状被検物１を芯出しして固定する。自動ステージ５は、円筒状被検物１の軸方向であるｘ方向と、ｘ方向と直交するｙ方向及びｚ方向に移動自在である。そして円筒状被検物１を把持冶具３で固定し、回転モータ４により把持冶具３を回転しながら、測定ヘッド６を円筒状被検物１の軸方向であるｘ方向に移動させて、円筒状被検物１の全面の測定を行う。
【００２１】
測定ヘッド６は、図２の斜視図に示すように、光源１０と、光源１０より円筒状被検物１側に設けられた格子パターン１１と、格子パターン１１に対して円筒状被検物１と反対側で、格子パターン１１に対して光源１０と同じ距離を隔てて設けられたレンズ１２と、レンズ１２に対して格子パターン１１とは反対側に設けられた２次元エリアセンサ１３を有する。２次元エリアセンサ１３は、光源１０から格子パターン１１を介して円筒状被検物１に投影されたモアレ縞を撮像するモアレ縞撮像領域１３１と、それ以外の部分を撮像する背景撮像領域１３２を有する。モアレ縞撮像領域１３１は少なくとも任意の３ラインの画素列でモアレ縞を撮像する。ここでは３ラインの場合に関して説明する。モアレ縞撮像領域１３１の各ラインを、図３に示すように、Ａ列とＢ列とＣ列とし、モアレ縞撮像領域１３１と格子パターン１１と円筒状被検物１の位置関係は、図３及び図４に示すように被検物１が円筒状であることを利用して、画素列Ａ，Ｂ，Ｃの視野に対応する高さを変化させてやる。また、画素列Ａ，Ｂ，Ｃはｙ方向に並んでいる。ここで所望のステップ量が与えられるように、円筒状被検物１の回転スピードと２次元エリアセンサ１３の走査周期と撮像倍率と画素列Ａ，Ｂ，Ｃ間の距離を調節してやる。
【００２２】
このモアレ縞撮像領域１３１の画素列Ａ，Ｂ，Ｃで円筒状被検物１を撮像するときは、まず、図４に示すように、時刻ｔ_１においてＡ列で円筒状被検物1の領域３（ステップ０面）を、Ｂ列で領域２（ステップ１面）を、Ｃ列で領域１（ステップ２面）撮像する。次に時刻ｔ_２においては、Ａ列で領域４（ステップ０面）を、Ｂ列で領域３（ステップ1面）を、Ｃ列で領域２（ステップ２面）を撮像する。さらに、時刻ｔ_３ではＡ列で領域５（ステップ０面）を、Ｂ列で領域４（ステップ１面）を、Ｃ列で領域３（ステップ２面）を撮像する。これを繰り返すことにより、画像メモリ上に、図５に示すように、各時刻ごとの画素列Ａ，Ｂ，Ｃによる検出データが得られる。そこで、時刻ｔ₁のＡ列のデータと時刻ｔ_２のＢ列のデータと時刻ｔ_３のＣ列のデータを下記（４）式
Φ(ｘ、ｙ)＝tan^−１｛(Ｉ３−Ｉ２)／(Ｉ１−Ｉ２)｝＋π／４（４）
から領域３の形状測定を行うことができる。この定量的な形状データをもとに円筒状被検物１の表面に生じるうねりやへこみ等の欠陥検査や平坦度の検査を行う。
【００２３】
厳密には、縞次数により縞間隔が異なるので、測定する縞次数によりステップ量が異なり測定誤差が生じるが、レンズ１２から格子パターン１１までの距離Ｌ＝200ｍｍ、光源１０とレンズ１２の距離70ｍｍ、格子パターン１１の格子間隔ｓ＝83.3μｍ（12本／ｍｍ）とした場合、下記（３）式
ｈ_Ｎ＝（ＮｓＬ）／（ｄ−Ｎｓ）（３）
によりモアレ等高線縞ｈ_Ｎは、図６に示すようになる。ここで、円筒状被検物１の基準高さを縞次数ｎ＝３の位置に、測定範囲をｎ＝２〜４の約480μｍの範囲に設定したとすると、Δｈ_２＝239.423μｍ、Δｈ_３＝239.995μmの差は0.572μｍとわずかであり、高低差が数μｍ程度のうねりやへこみを測定するには問題のないレベルである。
【００２４】
以上の方法においては、円筒状被検物１と２次元エリアセンサ１３の画素列Ａ，Ｂ，Ｃの位置関係が非常に重要となってくる。すなわち、円筒状被検物１が回転振れの影響でｙ方向に移動した場合、所望のステップ量からずれてしまうので形状データに誤差が生じる。一方、円筒状被検物１が２次元エリアセンサ１３の光軸方向ｚに振れた場合は、ステップ量は変わらないため回転振れを含めた形状データが得られるが、円筒状被検物１の１周に対応する周波数成分を取り除く等の周波数処理により取り除くことができる。すなわち、位相シフトモアレ法を用いた円筒状被検物１の形状測定において、常に測定ヘッド６と円筒状被検物１の位置関係がｙ方向において一定となるように制御し、位相シフト誤差量を低減させることが必要である。
【００２５】
そこで、図２に示すように、モアレ縞撮像領域１３１とそれ以外の部分の背景撮像領域１３２を有する２次元エリアセンサ１３で撮像した同一画像から形状情報と位置情報を得る。この２次元エリアセンサ１３で円筒状被被検物１に投影したモアレ縞を撮像した画像の典型的な例を図７に示す。図７に示す画像２０の外側矩形が２次元エリアセンサ１３の撮像領域を示す。この画像２０の中心領域２１には円筒状被検物１の被測定部の縞状の像があり、それ以外の無地部は被測定部が存在しない背景部分のため反射光がなく一様な画像２２となる。ここで円筒状被検物１の回転に伴い、図７に示すy方向のずれは一様な画像２２の距離ｄｙに相当し、z方向のずれは撮像倍率の変動となるので中心領域２１の距離ｗｙに相当する。このため距離ｄｙと距離ｗｙを求めることにより、２次元エリアセンサ１３で撮像した同一画像２０を画像処理することによりずれ量を求めることができ、同一画像２０からずれ量と表面形状を同時に検出することができる。
【００２６】
ここで位相シフトモアレにより測定する領域が円筒状被検物１の中心軸近辺であることを考慮すると、ｙ方向のずれとしては（ｄｙ＋ｗｙ）／２を利用することもできる。さらに、よりコントラストの大きな画像とするために、円筒状被検物１の下方からの透過光を２次元エリアセンサ１３に入射しても良い。
【００２７】
このようにして検出したずれ量により測定ヘッド６の位置を自動ステージ５で制御することにより、円筒状被検物１に対する測定ヘッド６の位置を一定に保つことができ、位相シフト誤差を低減することができる。また、検出したｙ方向のずれ量に応じて２次元エリアセンサ１３の位相シフトに使用するラインを選択することにより、正確な位相シフトを行うことができる。
【００２８】
また、高精度な測定を行うときは、高倍率でモアレ縞を撮像する必要があり、図７に示すように背景部分を取り込むことは難しい。このため測定ヘッド６に、図８に示すように、レーザポインタ等のスポット光源１４を設け、このスポット光源１４から、図９に示すように、円筒状被検物１の表面にスポット光２３を投射し、２次元エリアセンサ１３はモアレ縞撮像領域１３１でモレア縞を撮像するとともに背景撮像領域１３２でスポット光２３によりスポット像２４を撮像する。この２次元エリアセンサ１３で撮像した画像２０を図１０に示す。この画像２０のスポット像２４の位置ｄｙから三角測量の原理で円筒状被検物１の位置を検出することができる。このようにして円筒状被検物１の外形を利用して距離センサでｙ方向のずれ量を測定することなしにｙ方向のずれ量を検出することができる。
【００２９】
ここで円筒状被検物１に投射するスポット光２３は一点に限らず、複数の点であっても良い。また、スポット光２３の代わりに円や矩形等の形状を画像として投影し、その像を２次元エリアセンサ１３で撮像して、その形状変化や面積変化あるいは重心等を用いることにより円筒状被検物１の傾きや位置等も検出することができる。
【００３０】
また、図１１の模式図に示すように、格子パターン１１の２次元エリアセンサ１３の観測点側の格子端部においては、円筒状被検物１からの反射光が再度通過しない端部パターン部分１１１があり、この端部パターン部分１１１においてはモアレ縞は形成されず、２次元エリアセンサ１３で格子パターン１１がそのまま撮像され、撮像された画像２０には、図１２に示すように、円弧状の格子模様２５が形成される。この円弧状の格子模様２５から円筒状被検物１の中心軸を求めることにより、円弧状被検物１の回転に伴うずれ量を検出することができる。
【００３１】
また、格子パターン１１の端部パターン部分１１１を２次元エリアセンサ１３に直接取り込まないで、モアレ縞のみを撮像するように、２次元エリアセンサ１３の解像度を格子パターン１１の幅に対して充分細かくない条件で画像取り込む場合は、端部パターン部分１１１で生ずる円弧状の格子模様２５は細かすぎて観測できなくなる。このような場合は、図１３（ａ）に示すように、端部パターン部分１１１だけ周期を大きくしたり、（ｂ）に示すように、格子パターン１１を形成しているガラス等の格子パターン基板１１２の端面に光が通過しないようにしたずれ量測定パターン１１３を形成することにより円弧状の格子模様２５を有する画像２０を得ることができる。
【００３２】
また、位相シフトを行うためには、少なくとも同一個所において３枚の画像データを取得する必要がある。そこで通常のモアレ法を用い円筒状被検物１を撮像し、図１４に示すように、位相をシフトした３枚の形状測定用の画像２６ａ，２６ｂ，２６ｃを撮像して円筒状被検物１の凹凸情報を得て、この凹凸情報から画像縦方向である円周方向の列毎に位相シフト処理を行い最も凸な部分を求め、直線近似を行い位相つなぎを行った画像２７ａ〜２７ｃから円筒状被検物１の軸２８を検出しても良い。この場合は位相シフトを行わないため、円筒状被検物１の形状等の精度は劣るが、軸２８を表すラインからy方向の位置変動を検出することができる。この際、モアレ法から求まる凹凸データを円筒状被検物１の円筒形状にフィッティングを行うことによって、より高精度な位置を検出することができる。また、図１４に示すように、位相とびが発生している部分をつないで、円周方向により広い範囲でフィッティングを行うことにより高精度化を図ったり、中心部分の１周期のみを用いてフィッティングを行うことにより、位相つなぎ処理を省き演算時間を短縮することもできる。さらに、モアレ処理を図１５に示すように、限定した領域２９においてのみ行い、直線近似を行うことによって、円筒状被検物１の中心軸位置算出を行うための処理量を減らすことができ、計算コストを低減することができる。
【００３３】
また、図４に示すように、順次移動する円筒状被検物１と同期することによって位相シフトを行うような場合も、形状測定を行う処理は異なる時間で取り込んだデータから位相シフト処理を行うが、円筒状被検物１の軸位置を算出するためのモアレ法において使用する画像データは同一時間の画像データを用いて、時間ごとに求めることができる。さらに、このように順次移動する被被検物表面と同期することによって位相シフトを行うような場合においては、検出した円筒被検物１の軸の位置から、図１６に示すように、位相シフトを行うラインを変更することによって、軸ずれに伴うｙ方向の位置変動の誤差を低減することができる。図１６においては、位相シフトを行うラインとして、軸２８の最も凸なラインから上下にそれぞれ一定幅のラインを示しているが、実際は上下ではなく検出した中心からの相対位置として規定することになるので、求めた軸をそのまま利用するとは限らないし、位相シフトを行うそれぞれのラインの間隔も、円筒状被検物１の径にばらつきがある場合などには適宜変更を行うことにより、より精度を向上することができる。
【００３４】
また、軸の直線近似により求まる２次元エリアセンサ１３となす角度を用い、２次元カメラ・光源・格子パターンからなる光学系と、被被検物の軸方向を一定に制御して位相シフトを行うラインデータを正確に軸方向と平行にすることにより、回転に伴い測定個所ごとに異なってしまう位相シフト量を一定にすることができる。
【００３５】
さらに、図１４に示す１枚の位相シフトを行わないモアレ縞画像において、円周方向、すなわち画像２６の高さ変化を円筒状被検物１の円弧にフィッティングすることによって求めた軸位置情報によって、位相シフトして求めた形状データを補正することができる。なお、位相シフト画像が格子投影条件を変化させるようにして行う従来の方法では、各画像においてπ／２のオフセットが順次生ずるので、これを合わせて補正する必要があるが、図４に示すように、回転に同期して位置で位相シフトを行う方法においては、この補正をしないですむ。
【００３６】
【発明の効果】
この発明は以上説明したように、２次元エリアセンサによって得られる位相シフトしたモアレ縞データから形状測定中に、位相シフトしたモアレ縞データと同一画像中から被検物の位置変動検出を行うようにしたから、位置変動検出用のセンサを使用せずに、形状測定中の位置変動を検出することができ、位置変動検出用のセンサのキャリブレーション等を不要にすることができ、形状測定の簡易化を図るとともに、測定精度を向上させることができる。
【００３７】
形状測定中に位相シフトしたモアレ縞データと同一画像のモアレ縞の発生していない領域に規則性のあるパターン光を投影し、その位置又はパターン形状から被検物の位置変動を検出することにより、形状測定中の位置変動を容易に検出することができる。
【００３８】
また、位相シフトしたモアレ縞データと同一画像のモアレ縞の発生していない領域に形成されたモアレ縞を形成するための格子パターン端部のモアレを形成しない部分のパターンの変化から被検物の位置変動を検出することにより、より簡単な構成で形状測定中の位置変動を検出することができる。
【００３９】
また、形状測定中に位相シフトしないモアレ縞画像から被検物の位置変動を検出することにより、一度に検出可能な形状の領域を大きくできるとともに被検物の回転振れも検出することができる。
【００４０】
さらに、被検物の位置変動検出データに応じて形状測定に利用する位相シフト算出領域を変更して位相シフト演算を行うことにより、位相シフト誤差を低減させることができる。
【００４１】
また、被検物の位置変動に応じて測定ヘッドと被検物の位置関係が一定になるように制御することにより、回転振れの影響を抑制して位相シフト誤差を低減させることができる。
【図面の簡単な説明】
【図１】この発明の形状測定装置の構成を示す斜視図である。
【図２】測定ヘッドの構成を示す斜視図である。
【図３】２次元エリアセンサと格子パターンと円筒状被検物の位置関係を示す配置図である。
【図４】２次元エリアセンサのモアレ縞撮像領域の画素列の測定領域を示す配置図である。
【図５】モアレ縞撮像領域の画素列の時刻毎の測定データを示す説明図である。
【図６】モアレ縞撮像領域の画素列の測定データを示す説明図である。
【図７】モアレ縞を撮像した画像を示す図である。
【図８】スポット光源を有する測定ヘッドの構成を示す斜視図である。
【図９】２次元エリアセンサでモアレ縞とスポット像を撮像するときの配置図である。
【図１０】モアレ縞とスポット像の画像を示す図である。
【図１１】格子パターンの円筒状被検物からの反射光が再度通過しない端部パターン部分を示す模式図である。
【図１２】端部パターン部分による円弧状の格子模様を含む画像を示す図である。
【図１３】円弧状の格子模様を含む画像を形成する他の格子パターンの構成図である。
【図１４】通常のモアレ法を用い円筒状被検物を撮像した画像を示す図である。
【図１５】モアレ処理を行う限定した領域を示す模式図である。
【図１６】位相シフトを行うラインを示す模式図である。
【図１７】従来の構成を示す配置図である。
【図１８】他の従来例の構成を示す配置図である。
【図１９】格子投影型のモアレ法の説明図である。
【図２０】実体格子型のモアレ法の説明図である。
【図２１】モアレ法による光源と観測点と格子及び物体の配置図である。
【図２２】実体格子型のモアレ法の光源と観測点と格子及び物体の配置図である。
【図２３】位相変調された縞画像の光強度特性図である。
【符号の説明】
１；円筒状被検物、２；形状測定装置、３；把持冶具、４；回転モータ、
５；自動ステージ、６；測定ヘッド、１０；光源、１１；格子パターン、
１２；レンズ、１３；２次元エリアセンサ、１４；スポット光源。[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a shape measuring method and a shape measuring apparatus for detecting, for example, a surface shape of a cylindrical specimen or the like, and defects such as scratches, bulges, waviness and dents.
[0002]
[Prior art]
For example, JP-A-2-201142 and JP-A-4-169840 are disclosed as defect inspection methods for cylindrical specimens such as photosensitive drums. In the abnormality detection method disclosed in Japanese Patent Laid-Open No. 2-201142, as shown in FIG. 17, the laser beam 32 from the light source 31 is scanned in the axial direction of the photosensitive drum 33 through the rotary polygon mirror 36. The scanning light is reflected on the surface of the photosensitive layer of the drum 30, and the reflected light from the normal surface almost enters the light receiver 35 and the intensity of the reflected light is detected, and the output of the light receiver 5 is Input to a predetermined arithmetic processing unit or the like. When the detected value is abnormally lowered by the processing here, it is detected as an abnormality of the surface state. Further, the circumferential surface flaw inspection method disclosed in Japanese Patent Application Laid-Open No. 4-169840 projects slit light 42 from a projector 41 having a halogen light source or the like toward a photosensitive drum 43 as shown in FIG. The scattered light scattered by the surface defect of the photosensitive drum 43 is collected by the lens 44, received by the line sensor 45, and the abnormality is detected by the scattered light due to the defect.
[0003]
[Problems to be solved by the invention]
There is a possibility that a variety of defects such as pinholes, dents, scratches, entrainment of bubbles, cracks, dust, etc., and unevenness of the film thickness of the photosensitive layer, dripping, and scratches on the support may occur on the photoreceptor. is there. In the case of the optical inspection apparatus as described above, a defect having a large change rate of surface irregularities, such as a defect due to adhesion of pinholes, dents, scratches, dust, etc., can be detected with high accuracy. There is a problem in detection accuracy for a defect having a small unevenness change rate such as uneven thickness or a defect having no unevenness on the surface of the photosensitive member such as a scratch on the support.
[0004]
On the other hand, a moire method is one method of the three-dimensional measurement method. The Moire method has a solid lattice type and a lattice projection type, and is widely used in various fields. As shown in FIG. 19, the grid projection type moire method is such that small gratings G1 and G2 are arranged for projection and observation, respectively, and G1 is projected onto an object by a lens L1 and deformed according to the object shape. The lattice line thus formed is imaged on another grating G2 through the lens L2, and a fringe contour line is generated at a predetermined distance from the reference plane. In the actual lattice type moire method, as shown in FIG. 20, one lattice G is installed on the reference plane, and as shown in FIG. 21, the point light source S1 is placed at the position of the lens L1, and the observation is made at the position of the lens L2. The eye e is placed, the shadow of the light source S1 of the lattice G is dropped on the object, a shadow of the lattice G deformed according to the shape of the object is formed, and this is observed through the lattice G. This is a method of observing moire fringes caused by the above.
[0005]
This solid lattice moire method will be described in more detail. As shown in FIG. 21, gratings G1 and G2 are arranged on the same plane between the light source S1 and the observation point S2 and the object O, the distance between the light source S1 and the observation point S2 is d, and the grating from the light source S1 and the observation point S2 The distance from G1 and G1 is L, the distance from the lattices G1 and G2 to the object O is h, and the lattices G1 and G2 both have a pitch s, but the lattices G1 and G2 are shifted from each other by ε in the plane ( If the phase of the lattice pitch is 2πε / s), it can be expressed by the following equation (1).
cos (2π / s) · [{dh−ε (h + L)} / (h + L)] (1)
The moire fringes (contour lines) that are formed have orders that are counted as first order and second order as they move away from the lattice plane with the lattice plane as a reference (0th order). Therefore, the moire fringes of the fringe order N are obtained by placing cos 2πN. As a result, the Nth moire contour line is h from the reference plane._NIt is formed at a position indicated by the following formula (2) that is separated by
h_N= {(Ns + ε) L} / (d−Ns−ε) (2)
This does not include the position coordinate x, and is a unique value determined by the fringe order N. That is, the contour lines are formed.
[0006]
The real lattice type moire method shown in FIG. 22 corresponds to a case where the lattices G1 and G2 shown in FIG. 21 are formed as one continuous lattice G, and ε = 0. The following formula (3) is established from the formula (2).
h_N= (NsL) / (d-Ns) (3)
However, although it is called a contour line, the interval Δh_N= H_{N + 1}-H_NIs not constant and varies depending on the order N.
[0007]
Conventionally, the three-dimensional shape measurement method based on the moire method can intuitively grasp the object, but (1) it is difficult to determine unevenness, and (2) is not suitable for high-sensitivity three-dimensional measurement (currently moire). The interval between the fringe contour lines is limited to about 10 μm.) (3) Since the visibility of moire fringes is not uniform for each fringe, it is difficult to handle the moire image as an object of image processing. This problem is that, in the case of the grating projection type, two gratings are used, so that one of the two gratings is moved to shift the fringe scanning, that is, the phase of the moire fringes, so that the contour line interval is equally finely divided. At the same time, it is possible to determine the unevenness of the object and improve the measurement sensitivity.
[0008]
The principle of this phase shift method will be described. As shown in FIG. 23, the phase-modulated fringe image can be expressed by the following equation, where a is a bias, b is an amplitude, θ is an operable phase, and Φ is a phase value corresponding to a height.
I = I (θ) = a (x, y) + b (x, y) cos {Φ (x, y) + θ}
What is desired here is the phase Φ (x, y) at each point (x, y). Since the bias a and the amplitude b are unknown components that change due to the reflectance and dirt on the surface, three fringe images represented by the following equations are generated by changing the phase θ to 0, π / 2, and π.
I = I (0) = a (x, y) + b (x, y) cos {Φ (x, y) + θ}
I = I (π / 2) = a (x, y) + b (x, y) cos {Φ (x, y) + θ}
I = I (π) = a (x, y) + b (x, y) cos {Φ (x, y) + θ}
If the phase Φ (x, y) is calculated by the following equation (4), the reflectance and dirt components can be removed and the phase Φ (x, y) at each point can be obtained.
Φ (x, y) = tan^-1{(I3-I2) / (I1-I2)} + π / 4 (4)
[0009]
However, in the case of the real lattice type, since there is only one lattice G, even if a phase shift like the lattice projection type moire method is performed, it is not possible to change the phase while aligning the phases of the fringe contour lines of all orders. Can not. For such a problem, for example, in the high-sensitivity three-dimensional measurement method disclosed in Japanese Patent No. 2887517, the vertical movement of the lattice plane and the horizontal movement of the light source or the observation point are performed at the same time. Since the fringe phase can be shifted with respect to the measurement target with almost the same phase of the fringes of each order without causing any substantial change in the phase, processing from multiple fringe images based on the principle of the phase shift method As a result, the density of measurement points due to moire fringes on the object to be measured is increased, and a physical division of about 1/40 to 1/100 is possible for one cycle of moire fringes. It is possible to determine the unevenness of the surface which has been difficult and to improve the measurement sensitivity.
[0010]
When the entire surface of a cylindrical specimen or the like is measured by applying the phase shift method as described above, it is necessary to repeat grating movement and moire fringe imaging in order to shift the specimen by at least three rotations. It takes time to measure. Moreover, since it is necessary to move the grating in a plurality of directions (parallel and rotation), there is a problem that the apparatus configuration is complicated.
[0011]
Surface shape measuring device disclosed in Japanese Patent Application Laid-Open No. 7-332956 and the literature “Substance Lattice Moire Method by Phase Shift” (Proceedings of the 1991 Precision Engineering Fall Meeting), “Measurement of Flatness of Liquid Crystal Glass "(O plus E September 1996) eliminates the difference in fringe spacing due to the fringe order by applying parallel light, so that all fringe phases are shifted while being aligned. In these methods, it is possible to shift the phase only by the grating motion.
[0012]
However, when still measuring the whole surface of a cylindrical specimen or the like, it is necessary to repeat the movement of the grating and the imaging and to rotate the specimen three times or more in order to obtain a phase-shifted image, thereby increasing the measurement time. . In the surface shape measuring method disclosed in Japanese Patent Laid-Open No. 10-54711, the phase is shifted by changing the height of the test object. Even in this case, it is necessary to repeat the movement and imaging of the test object a plurality of times, resulting in an increase in measurement time. In addition, a clear method for quantifying the uneven shape is not sufficiently explained.
[0013]
The present invention solves the above-described problems, applies to a cylindrical specimen such as a roller part, applies the phase shift method to the tangential grid type moire method, and performs one series of one image pickup device. Shape measurement that can measure the shape of the object at high speed by measuring the phase-shifted image by imaging, and inspect the surface of the object from the quantitative shape data and measure the movement error of the object It is an object of the present invention to provide a method and a shape measuring apparatus.
[0014]
[Means for Solving the Problems]
  The shape measuring method according to the present invention rotates a test object,A moiré fringe is projected onto the test object by a body grating type moire optical system having a two-dimensional area sensor, and the area other than the area where the moire fringe projected on the test object is formed and the area where the moire fringe is formed A region is imaged by the two-dimensional area sensor, and moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase. Then, the shape of the test object is measured from the generated moire fringe data, and the position variation of the test object is detected from the image of the area other than the area where the moire fringe is formed in the image captured by the two-dimensional area sensor. It is characterized by doing.
[0015]
A pattern light having regularity is projected onto a region where no moire fringes are generated in the same image as the moire fringe data phase-shifted during the shape measurement, and the position variation of the test object is detected from the position or pattern shape. The position of the test object from the change in the pattern of the portion where the moire fringes formed in the area where the moire fringes of the same image as the phase-shifted moire fringes are not formed are formed. By detecting a change and performing image processing on the same image as the phase-shifted moire fringe data without using a position change detection sensor, a position change during shape measurement is detected.
[0016]
Further, the position variation of the test object may be detected from a moire fringe image that is not phase-shifted during shape measurement.
[0017]
Furthermore, the phase shift calculation region is changed by changing the phase shift calculation area used for the shape measurement according to the position variation detection data of the test object, and the phase shift error is reduced.
[0018]
Further, the positional relationship between the measurement head and the test object is controlled to be constant according to the position variation of the test object, thereby suppressing the influence of rotational shake and reducing the phase shift error.
[0019]
  The shape measuring apparatus according to the present invention includes a measuring head having a gantry optical system of a solid lattice type having a two-dimensional area sensor, and a gripping rotation mechanism for holding and rotating a test object,The test object is rotated, and moire fringes are projected onto the test object by the solid lattice type moire optical system, and a moire fringe projected on the test object and a moire fringe are formed. A region other than a region is imaged by the two-dimensional area sensor, and the moire fringes of a specific fringe order in the region where the moire fringes are formed in the image captured by the two-dimensional area sensor are shifted by a desired phase. Data is generated, the shape of the test object is measured from the generated moire fringe data, and the position of the test object is determined from an image in a region other than a region where moire fringes are formed in the image captured by the two-dimensional area sensor. It is characterized by detecting fluctuations.
[0020]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a perspective view showing a configuration of a shape measuring apparatus according to the present invention. As shown in the figure, a shape measuring apparatus 2 for detecting the shape and defect of a cylindrical specimen 1 is a gripping jig for fixing the cylindrical specimen 1 such as a photosensitive member used for electrophotographic image formation, for example. 3, a rotation motor 4 that rotates the gripping jig 3, and a measurement head 6 provided on the automatic stage 5. The gripping jig 3 has, for example, a three-claw chuck or the like, and centers and fixes the cylindrical specimen 1. The automatic stage 5 is movable in the x direction, which is the axial direction of the cylindrical test object 1, and in the y direction and the z direction orthogonal to the x direction. Then, the cylindrical test object 1 is fixed by the gripping jig 3, and the measurement head 6 is moved in the x direction which is the axial direction of the cylindrical test object 1 while rotating the gripping jig 3 by the rotary motor 4. The entire surface of the test object 1 is measured.
[0021]
As shown in the perspective view of FIG. 2, the measuring head 6 includes a light source 10, a grid pattern 11 provided on the cylindrical test object 1 side from the light source 10, and the cylindrical test object 1 with respect to the grid pattern 11. And a two-dimensional area sensor 13 provided on the opposite side of the lens 12 from the lattice pattern 11. The two-dimensional area sensor 13 includes a moire fringe imaging region 131 that images a moire fringe projected from the light source 10 through the lattice pattern 11 onto the cylindrical test object 1 and a background imaging region 132 that images other portions. Have. The moiré fringe imaging region 131 images moiré fringes with at least three arbitrary pixel rows. Here, the case of three lines will be described. As shown in FIG. 3, the lines of the moiré fringe imaging region 131 are arranged in rows A, B, and C, and the positional relationship between the moire fringe imaging region 131, the lattice pattern 11, and the cylindrical test object 1 is shown in FIG. As shown in FIG. 4, the height corresponding to the visual field of the pixel rows A, B, and C is changed using the fact that the test object 1 is cylindrical. The pixel columns A, B, and C are arranged in the y direction. Here, the rotational speed of the cylindrical test object 1, the scanning period of the two-dimensional area sensor 13, the imaging magnification, and the distance between the pixel rows A, B, and C are adjusted so that a desired step amount is given.
[0022]
When imaging the cylindrical test object 1 with the pixel rows A, B, and C of the moire fringe imaging region 131, first, as shown in FIG.₁In FIG. 5, the region 3 (step 0 surface) of the cylindrical test object 1 is imaged in the A row, the region 2 (step 1 surface) is imaged in the B row, and the region 1 (step 2 surface) is imaged in the C row. Next time t₂In FIG. 4, the region A (step 0 surface) is imaged in the A column, the region 3 (step 1 surface) is imaged in the B column, and the region 2 (step 2 surface) is imaged in the C column. Furthermore, time t₃Then, the region A (Step 0 surface) is imaged in the A column, the region 4 (Step 1 surface) is imaged in the B column, and the region 3 (Step 2 surface) is imaged in the C column. By repeating this, detection data by pixel columns A, B, and C for each time is obtained on the image memory as shown in FIG. Therefore, time t₁Column A data and time t₂B column data and time t₃The data in column C of the following formula (4)
Φ (x, y) = tan^-1{(I3-I2) / (I1-I2)} + π / 4 (4)
Thus, the shape of the region 3 can be measured. Based on this quantitative shape data, a defect inspection such as swell and dent generated on the surface of the cylindrical object 1 and a flatness inspection are performed.
[0023]
Strictly speaking, since the fringe interval varies depending on the fringe order, the step amount varies depending on the fringe order to be measured, and a measurement error occurs. However, the distance L from the lens 12 to the lattice pattern 11 is 200 mm, the distance between the light source 10 and the lens 12 is 70 mm, When the lattice interval s of the lattice pattern 11 is 83.3 μm (12 pieces / mm), the following equation (3)
h_N= (NsL) / (d-Ns) (3)
Due to the moire contour stripes h_NIs as shown in FIG. Here, assuming that the reference height of the cylindrical specimen 1 is set at the position of the fringe order n = 3 and the measurement range is set at a range of about 480 μm where n = 2 to 4, Δh₂= 239.423 μm, Δh₃The difference of 239.995 μm is as small as 0.572 μm, which is a level that does not cause any problem in measuring undulations and dents with a height difference of about several μm.
[0024]
In the above method, the positional relationship between the cylindrical test object 1 and the pixel columns A, B, and C of the two-dimensional area sensor 13 is very important. That is, when the cylindrical test object 1 moves in the y direction due to the influence of the rotational shake, an error occurs in the shape data because it deviates from the desired step amount. On the other hand, when the cylindrical specimen 1 shakes in the optical axis direction z of the two-dimensional area sensor 13, the step amount does not change, and thus shape data including rotational shake can be obtained. It can be removed by frequency processing such as removing a frequency component corresponding to one round. That is, in measuring the shape of the cylindrical specimen 1 using the phase shift moire method, the positional relationship between the measurement head 6 and the cylindrical specimen 1 is always controlled to be constant in the y direction, and the phase shift error amount is set. It is necessary to reduce it.
[0025]
Therefore, as shown in FIG. 2, shape information and position information are obtained from the same image captured by the two-dimensional area sensor 13 having the moire fringe imaging region 131 and the background imaging region 132 of other parts. FIG. 7 shows a typical example of an image obtained by imaging the moire fringes projected on the cylindrical object 1 by the two-dimensional area sensor 13. An outer rectangle of the image 20 illustrated in FIG. 7 indicates an imaging region of the two-dimensional area sensor 13. In the central region 21 of this image 20, there is a striped image of the measured portion of the cylindrical test object 1, and the other plain portion is a background portion where the measured portion does not exist and is uniform with no reflected light. An image 22 is obtained. Here, as the cylindrical test object 1 rotates, the displacement in the y direction shown in FIG. 7 corresponds to the distance dy of the uniform image 22, and the displacement in the z direction results in fluctuations in the imaging magnification. It corresponds to the distance wy. Therefore, by obtaining the distance dy and the distance wy, the amount of deviation can be obtained by performing image processing on the same image 20 captured by the two-dimensional area sensor 13, and the amount of deviation and the surface shape are detected simultaneously from the same image 20. be able to.
[0026]
Here, considering that the region to be measured by the phase shift moire is in the vicinity of the central axis of the cylindrical specimen 1, (dy + wy) / 2 can also be used as the displacement in the y direction. Furthermore, in order to obtain an image having a higher contrast, transmitted light from below the cylindrical test object 1 may be incident on the two-dimensional area sensor 13.
[0027]
By controlling the position of the measuring head 6 by the automatic stage 5 based on the detected deviation amount, the position of the measuring head 6 with respect to the cylindrical object 1 can be kept constant, and the phase shift error is reduced. be able to. In addition, an accurate phase shift can be performed by selecting a line used for the phase shift of the two-dimensional area sensor 13 in accordance with the detected displacement amount in the y direction.
[0028]
Further, when performing highly accurate measurement, it is necessary to capture moire fringes at a high magnification, and it is difficult to capture the background portion as shown in FIG. Therefore, the measuring head 6 is provided with a spot light source 14 such as a laser pointer as shown in FIG. 8, and the spot light 23 is emitted from the spot light source 14 onto the surface of the cylindrical object 1 as shown in FIG. The two-dimensional area sensor 13 images a moire fringe in the moire fringe imaging region 131 and images a spot image 24 with the spot light 23 in the background imaging region 132. An image 20 captured by the two-dimensional area sensor 13 is shown in FIG. The position of the cylindrical test object 1 can be detected from the position dy of the spot image 24 of the image 20 by the principle of triangulation. In this way, the amount of deviation in the y direction can be detected without measuring the amount of deviation in the y direction with the distance sensor using the outer shape of the cylindrical specimen 1.
[0029]
Here, the spot light 23 projected onto the cylindrical test object 1 is not limited to one point, and may be a plurality of points. Further, instead of the spot light 23, a shape such as a circle or a rectangle is projected as an image, the image is picked up by the two-dimensional area sensor 13, and the shape change, the area change, the center of gravity, or the like is used to detect the cylindrical inspection. The inclination and position of the object 1 can also be detected.
[0030]
Further, as shown in the schematic diagram of FIG. 11, at the lattice end portion on the observation point side of the two-dimensional area sensor 13 of the lattice pattern 11, the end pattern portion where the reflected light from the cylindrical test object 1 does not pass again. In this end pattern portion 111, moire fringes are not formed, and the lattice pattern 11 is directly captured by the two-dimensional area sensor 13, and the captured image 20 has an arc shape as shown in FIG. The lattice pattern 25 is formed. By obtaining the central axis of the cylindrical test object 1 from the arc-shaped lattice pattern 25, it is possible to detect the amount of deviation accompanying the rotation of the arc-shaped test object 1.
[0031]
Further, the resolution of the two-dimensional area sensor 13 is sufficiently fine with respect to the width of the lattice pattern 11 so that only the moire fringes are captured without directly capturing the end pattern portion 111 of the lattice pattern 11 into the two-dimensional area sensor 13. When the image is captured under such a condition, the arc-shaped lattice pattern 25 generated in the end pattern portion 111 is too fine to be observed. In such a case, as shown in FIG. 13A, the period is increased only by the end pattern portion 111, or as shown in FIG. 13B, a lattice pattern substrate such as glass forming the lattice pattern 11 is formed. An image 20 having an arc-shaped lattice pattern 25 can be obtained by forming a shift amount measurement pattern 113 that prevents light from passing through the end face of 112.
[0032]
In order to perform phase shift, it is necessary to acquire three pieces of image data at least at the same location. Therefore, the cylindrical specimen 1 is imaged using a normal moire method, and as shown in FIG. 14, three phase-measurement images 26a, 26b, and 26c are imaged to obtain a cylindrical specimen. 1 is obtained, and the phase shift process is performed for each column in the circumferential direction, which is the vertical direction of the image, to obtain the most convex portion, and from the images 27a to 27c obtained by performing linear approximation and phase linking. The shaft 28 of the cylindrical test object 1 may be detected. In this case, since the phase shift is not performed, the accuracy of the shape or the like of the cylindrical test object 1 is inferior, but the position fluctuation in the y direction can be detected from the line representing the axis 28. At this time, by fitting the unevenness data obtained from the moire method to the cylindrical shape of the cylindrical test object 1, a more accurate position can be detected. Further, as shown in FIG. 14, by connecting portions where phase jumps are generated and fitting is performed in a wider range in the circumferential direction, high accuracy is achieved, or fitting is performed using only one cycle of the central portion. By performing the above, it is possible to omit the phase linking process and shorten the calculation time. Furthermore, as shown in FIG. 15, the moire processing is performed only in the limited region 29, and by performing linear approximation, the processing amount for calculating the center axis position of the cylindrical test object 1 can be reduced. Calculation cost can be reduced.
[0033]
Further, as shown in FIG. 4, even when the phase shift is performed by synchronizing with the cylindrical specimen 1 that moves sequentially, the shape measurement processing is performed from the data acquired at different times. However, the image data used in the moire method for calculating the axial position of the cylindrical test object 1 can be obtained for each time using image data of the same time. Further, in the case where the phase shift is performed by synchronizing with the surface of the specimen to be sequentially moved in this way, the phase shift is performed from the detected position of the axis of the cylindrical specimen 1 as shown in FIG. By changing the line for performing y, it is possible to reduce the error of the position fluctuation in the y direction due to the axis deviation. In FIG. 16, lines having a constant width are respectively shown above and below the most convex line of the axis 28 as lines for performing phase shift, but in actuality, they are defined as relative positions from the detected center, not up and down. Therefore, it is not always possible to use the obtained axis as it is, and the interval between the lines for performing the phase shift is also appropriately changed when the diameter of the cylindrical test object 1 is varied, thereby further improving the accuracy. Can be improved.
[0034]
In addition, using the angle formed with the two-dimensional area sensor 13 obtained by linear approximation of the axis, the optical system composed of a two-dimensional camera, a light source, and a lattice pattern, and the axial direction of the test object are controlled to perform phase shift. By making the line data exactly parallel to the axial direction, it is possible to make the phase shift amount that varies depending on the measurement location with rotation constant.
[0035]
  further,As shown in FIG.In a single moire fringe image without phase shift, it is obtained by phase shifting according to the axial position information obtained by fitting the circumferential direction, that is, the height change of the image 26 to the arc of the cylindrical test object 1. The corrected shape data can be corrected. Note that in the conventional method in which the phase shift image is changed so as to change the grating projection conditions, an offset of π / 2 is sequentially generated in each image, and this needs to be corrected together. However, as shown in FIG. In addition, this correction is not necessary in the method of performing the phase shift at the position in synchronization with the rotation.
[0036]
【The invention's effect】
  As described above, according to the present invention, during the shape measurement from the phase-shifted moire fringe data obtained by the two-dimensional area sensor, the phase-shifted moire fringe data is obtained from the same image.Since the position variation of the test object is detected,Position fluctuation during shape measurement can be detected without using a sensor for position fluctuation detection, calibration of the sensor for position fluctuation detection can be made unnecessary, and shape measurement can be simplified. In addition, the measurement accuracy can be improved.
[0037]
By projecting regular pattern light to the area where no moire fringe is generated in the same image as the moire fringe data phase-shifted during shape measurement, and detecting the position fluctuation of the test object from the position or pattern shape The position fluctuation during the shape measurement can be easily detected.
[0038]
In addition, the phase change of the moire fringe data in the same image as the moire fringes formed in the region where the moire fringes are not generated causes the change in the pattern of the portion where the moire fringes at the end of the lattice pattern are not formed to change the pattern of the test object. By detecting the position variation, it is possible to detect the position variation during shape measurement with a simpler configuration.
[0039]
Further, by detecting the position variation of the test object from the moire fringe image that is not phase-shifted during the shape measurement, it is possible to enlarge the region of the shape that can be detected at one time and to detect the rotational shake of the test object.
[0040]
Furthermore, the phase shift calculation can be performed by changing the phase shift calculation region used for shape measurement according to the position variation detection data of the test object, thereby reducing the phase shift error.
[0041]
Further, by controlling so that the positional relationship between the measurement head and the test object becomes constant according to the position variation of the test object, it is possible to reduce the phase shift error by suppressing the influence of the rotational shake.
[Brief description of the drawings]
FIG. 1 is a perspective view showing a configuration of a shape measuring apparatus according to the present invention.
FIG. 2 is a perspective view showing a configuration of a measurement head.
FIG. 3 is a layout diagram showing a positional relationship among a two-dimensional area sensor, a lattice pattern, and a cylindrical test object.
FIG. 4 is a layout diagram illustrating a measurement region of a pixel column in a moire fringe imaging region of a two-dimensional area sensor.
FIG. 5 is an explanatory diagram showing measurement data for each pixel row in a moiré fringe imaging region.
FIG. 6 is an explanatory diagram showing measurement data of a pixel column in a moire fringe imaging region.
FIG. 7 is a diagram illustrating an image obtained by capturing moire fringes.
FIG. 8 is a perspective view showing a configuration of a measurement head having a spot light source.
FIG. 9 is a layout diagram when a moire fringe and a spot image are captured by a two-dimensional area sensor.
FIG. 10 is a diagram illustrating an image of moire fringes and a spot image.
FIG. 11 is a schematic diagram showing an end pattern portion where reflected light from a cylindrical test object having a lattice pattern does not pass again.
FIG. 12 is a diagram illustrating an image including an arc-shaped lattice pattern by end pattern portions.
FIG. 13 is a configuration diagram of another lattice pattern for forming an image including an arc-shaped lattice pattern.
FIG. 14 is a diagram illustrating an image obtained by capturing an image of a cylindrical test object using a normal moire method.
FIG. 15 is a schematic diagram illustrating a limited region where moire processing is performed.
FIG. 16 is a schematic diagram showing lines for performing phase shift.
FIG. 17 is a layout diagram showing a conventional configuration.
FIG. 18 is a layout view showing the configuration of another conventional example.
FIG. 19 is an explanatory diagram of a lattice projection type moire method.
FIG. 20 is an explanatory diagram of an actual lattice type moire method.
FIG. 21 is a layout diagram of a light source, observation points, a grid, and an object by a moire method.
FIG. 22 is a layout diagram of a light source, observation points, a grid, and an object of an actual grid type moire method.
FIG. 23 is a light intensity characteristic diagram of a phase-modulated fringe image.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1; Cylindrical test object, 2; Shape measuring apparatus, 3; Grasping jig, 4; Rotation motor,
5; automatic stage, 6; measuring head, 10; light source, 11; lattice pattern,
12; Lens, 13; Two-dimensional area sensor, 14; Spot light source.

Claims

The object is rotated , and the moire fringes are projected onto the test object by the moire optical system having a two-dimensional area sensor, and the moire fringes projected on the test object are formed. An area other than the area where the image is formed is imaged by the two-dimensional area sensor,
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
A shape measurement method, comprising: detecting a change in position of a test object from an image of a region other than a region where moire fringes are formed in an image captured by the two-dimensional area sensor.

The test object is rotated, and a moire fringe is projected onto the test object by means of a real lattice type moire optical system having a two-dimensional area sensor, and regular pattern light is projected onto the test object surface. A region other than the region where the moire fringes projected on the specimen are formed and the region where the moire fringes are formed are imaged by the two-dimensional area sensor,
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
A change in the position of the test object is detected from the position or pattern shape of the pattern light image formed in an area other than the area where moire fringes are formed in the image captured by the two-dimensional area sensor. Surface shape measurement method.

Rotating the object to be tested and projecting moire fringes onto the object by a moiré optical system having a two-dimensional area sensor and forming the moire fringes projected on the object and the moire fringes An area other than the area where the image is formed is imaged by the two-dimensional area sensor,
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
An object to be detected from a change in the pattern of a portion that is formed in a region other than a region where moire fringes are formed in an image picked up by the two-dimensional area sensor and does not form moire at the end of the lattice pattern for forming moire fringes. A surface shape measuring method characterized by detecting a position variation of the surface.

The object is rotated , and the moire fringes are projected onto the test object by the moire optical system having a two-dimensional area sensor, and the moire fringes projected on the test object are formed. An area other than the area where the image is formed is imaged by the two-dimensional area sensor,
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
A method for measuring a surface shape, comprising: detecting a change in position of a test object from a moire fringe image that is not phase-shifted in an image captured by the two-dimensional area sensor.

The surface shape measurement method according to claim 1, wherein a phase shift calculation is performed by changing a phase shift calculation region used for shape measurement in accordance with position variation detection data of the test object.

The shape measuring method according to claim 1, wherein control is performed so that a positional relationship between the measuring head and the test object is constant according to a change in position of the test object.

A measurement head having a solid lattice type moire optical system having a two-dimensional area sensor, and a gripping rotation mechanism for holding and rotating a test object,
The test object is rotated, and moire fringes are projected onto the test object by the solid lattice type moire optical system, and a moire fringe projected on the test object and a moire fringe are formed. An area other than the area is imaged by the two-dimensional area sensor,
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
A shape measuring apparatus for detecting a position variation of a test object from an image of an area other than an area where moire fringes are formed in an image taken by the two-dimensional area sensor.

A measurement head having a solid lattice moire optical system having a two-dimensional area sensor and a spot light source for projecting regular pattern light; and a gripping rotation mechanism for holding and rotating a test object,
The test object is rotated, and moire fringes are projected onto the test object by the solid lattice type moire optical system, and a moire fringe projected on the test object and a moire fringe are formed. An area other than the area is imaged by the two-dimensional area sensor ,
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
A change in the position of the test object is detected from the position or pattern shape of the pattern light image formed in an area other than the area where moire fringes are formed in the image captured by the two-dimensional area sensor. Surface shape measuring device.

A measurement head having a solid lattice type moire optical system having a two-dimensional area sensor, and a gripping rotation mechanism for holding and rotating a test object ,
The test object is rotated, and moire fringes are projected onto the test object by the solid lattice type moire optical system, and a moire fringe projected on the test object and a moire fringe are formed. An area other than the area is imaged by the two-dimensional area sensor,
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
An object to be detected from a change in the pattern of a portion that is formed in a region other than a region where moire fringes are formed in an image picked up by the two-dimensional area sensor and does not form moire at the end of the lattice pattern for forming moire fringes. A surface shape measuring apparatus for detecting a positional variation of the surface.

A measurement head having a solid lattice type moire optical system having a two-dimensional area sensor, and a gripping rotation mechanism for holding and rotating a test object,
The test object is rotated, the moire fringes are projected onto the test object by the solid lattice moire optical system, and the area where the moire fringes projected on the test object are formed is imaged by the two-dimensional area sensor. And
Moire fringe data is generated by shifting a moire fringe of a specific fringe order in a region where moire fringes are formed in an image captured by the two-dimensional area sensor by a desired phase, and the test is performed from the generated moire fringe data. Measure the shape of the object,
A surface shape measuring apparatus for detecting a position variation of an object to be detected from a moire fringe image that is not phase-shifted in an image picked up by the two-dimensional area sensor .