JP3851395B2 - Optical member inspection apparatus and optical member inspection method - Google Patents

Description

このように算出される接眼レンズＬ３及び観察者の瞳からなる光学系の物体側有効Ｆナンバー（Ｆ_OA）と撮影レンズＬの物体側有効Ｆナンバー（Ｆ_OB）とが同じであれば、観察者が接眼レンズＬ３を覗いた時と同じ見え方の映像が撮像素子５ａによって撮像される。但し、実際には、人間の眼は無意識のピント調整が可能であるので、ピント面の前後における或る程度の範囲にピントを合わせることができてしまう。また、検査を多少厳しくする必要もある。
【００３０】
従って、撮像レンズＬの絞り１１は、式（２）によって求められたＦ_OAの値（接眼レンズの焦点距離／観察者の瞳径）に対応する絞り値以上（式（2'）が成立する前提下においては、１０以上）となるように、絞り込まれている。その結果、撮像レンズＬの被写界深度は、観察者が接眼レンズＬ３を覗いた時の被写界深度と同じになっている。従って、観察者が接眼レンズＬ３を覗いて物体の実像に目の焦点を合わせた時に見える映像が、そのままの見え方で、撮像素子５ａによって撮像されるのである。つまり、観察者が接眼レンズＬ３を覗いて物体の実像に目の焦点を合わせた時に明瞭に見える第１面Ｐ₁上の欠陥，及び被写界深度（Ｘ）内に含まれるポロプリズムＰの内部の欠陥（図９において網目で示す領域中の欠陥）は、撮像素子５ａによって撮像される映像中においても明瞭に表れる。また、観察者が接眼レンズＬ３を覗いて物体の実像に目の焦点を合わせた時に第１面Ｐ₁からの距離が遠くなるに従ってぼやけて見える被写界深度外の欠陥は、撮像素子５ａによって撮像される映像中においても、第１面Ｐ₁からの距離に応じてぼやけている。
【００３１】
なお、式（１）によると、Ｆナンバーが一定である場合には、撮像レンズＬの焦点距離（ｆ）は、この撮像レンズＬの入射瞳径（Ｄ）に影響を及ぼすことがわかる。そして、この入射瞳径（Ｄ）が小さくなりすぎると、拡散板７の光軸上部分から発散してポロプリズムＰの中心を通って入射瞳に入射する光束量に比べて、拡散板７の周辺部分から発散してポロプリズムＰの側面近傍を通って入射瞳に入射する光束量が極端に少なくなってしまう。その結果、図５に示すように、撮像素子５ａによって撮像される映像中においても、ポロプリズムＰの第１面Ｐ₁に対応した領域の周辺部分（▲２▼）が、中央部分（▲１▼）よりも暗くなってしまう。このような光量差が許容範囲を超えて大きくなると、画像処理において、この暗い部分が欠陥でないにも拘わらず欠陥であるとみなされてしまう恐れがある。
【００３２】
このような光量差の発生は、ポロプリズムＰの第５面Ｐ₅から出射される光線のうち、少なくとも、光軸と平行な全ての光線を撮像レンズＬの入射瞳に入射させることにより、欠陥抽出に影響しない程度に抑えることができる。このようにポロプリズムＰの第５面Ｐ₅から光軸と平行に出射される全ての光線を撮像レンズＬの入射瞳に入射させるには、撮像レンズＬの入射瞳径（Ｄ）をポロプリズムＰの第５面Ｐ₅の対角線長（ｄ）よりも大きくすれば良い。この関係は、下記式（３）によって表される。
【００３３】
Ｄ＞ｄ ……（３）
一方、式（１）において、Ｆを物体側有効Ｆナンバー（Ｆ_O），ｆを物体から主点までの距離ａ（但し、ａ＞０）と置き換えると、下記式（４）が得られる。
【００３４】
Ｆ_O＝ａ／Ｄ
Ｄ＝ａ／Ｆ_O ……（４）
従って、式（４）を式（３）に代入することによって、下記式（５）が得られる。
【００３５】
ａ／Ｆ_O＞ｄ
ａ＞ｄ・Ｆ_O ……（５）
また、主点から像までの距離をｂ（但し、ｂ＞０）とすると、焦点距離ｆ及び像倍率Ｍとの関係は、下記式（６），（７）によって示される。
【００３６】
ｂ／ａ＝Ｍ ……（６）
１／ａ＋１／ｂ＝１／ｆ ……（７）
これら式（６）及び式（７）から焦点距離ｆを解くと、式（８）の通りになる。
【００３７】
ｆ＝ａ（Ｍ／（Ｍ＋１）） ……（８）
式（８）に式（５）を代入すると、下記式（９）が得られる。
ｆ＞ｄ・Ｆ_O・（Ｍ／（Ｍ＋１））
＞ポロプリズムＰの第５面Ｐ₅の対角線長・
接眼レンズＬ３の焦点距離／観察者の瞳径・
（Ｍ／（Ｍ＋１）） ……（９）
この式（９）に従って、撮像レンズＬの焦点距離（ｆ）が設定されているので、光軸と平行な全ての光線が撮像レンズＬの入射瞳に入射する。なお、Ｍは、撮像素子５ａの受光面の大きさに基づいて設定される。
【００３８】
なお、図３に示される補正レンズ１２は、瞳の位置合わせのために挿入されている。
＜画像処理＞
次に、検査対象光学部材（ポロプリズム）が良品であるか不良品であるかの判定を行うために画像処理部８において実行される画像処理の内容を、図４のフローチャートを用いて説明する。
【００３９】
この画像処理は、画像処理部４に接続された図示せぬ検査開始ボタンが押下されることによりスタートする。そして、スタート後最初Ｓ０１では、画像処理部４は、サンプル（検査対象光学部材）のセットを促す指示を、表示装置９上に表示する。
【００４０】
次のＳ０２では、画像処理部４は、撮像素子５ａから入力された画像データのうちの一フレームを、第１メモリ８ａ内に取り込む。
次のＳ０３では、画像処理部４は、第１メモリ８ａに格納されている画像データ（撮像画像）から、検査対象領域（ポロプリズムＰの第５面Ｐ₅に対応する領域）に該当する部分を抽出して、二値化処理を行う（抽出手段に相当）。この二値化処理においては、画像処理部４は、抽出した画像データを構成する各画素の輝度値（０〜２５５）に対して平滑化処理を行うとともに平滑化後の各画素の輝度値（０〜２５５）を一律に所定量減算して閾値データを作成する。そして、検査対象領域から抽出した画像データを構成する各画素の輝度値（０〜２５５）と閾値データ中の対応する画素の輝度値（０〜２５５）とを比較し、前者が後者を下回っている画素については第２メモリ８ｂの対応位置に輝度値２５５を書き込み、前者が後者以上である画素については第２メモリ８ｂの対応位置に輝度値０を書き込む。この二値化処理の完了時点において第２メモリ８ｂに格納されている画像データ（二値化画像）は、観察者が見たときに気になる程濃い暗部に対応する部分のみが白く浮き上がった画像となっている。
【００４１】
次のＳ０４では、画像処理部４は、第２メモリ８ｂから画像データ（二値化画像）を読み出して、読み出した画像データに対して膨張処理を施し、膨張処理がなされた画像データ（膨張画像）を第３メモリ８ｃに書き込む（グループ化手段に相当）。この膨張処理とは、二値化画像中の白い領域の周縁を所定量だけ外側に拡げる処理である。この膨張処理の結果、図８（ａ）のように互いに隣接している領域（但し、図８においては、便宜上、白黒が逆転されて描かれている）は、図８（ｂ）のように、互いに結合して一繋がりの膨張領域となる。
【００４２】
次のＳ０５では、画像処理部４は、第３メモリ８ｃに格納されている画像データ（膨張画像）に対してラベリングを施す（グループ化手段に相当）。即ち、この画像データ（膨張画像）中の各膨張領域に対して、図８（ｂ）に示すように、ユニークな識別番号（ラベル）を付す。
【００４３】
次のＳ０６では、画像処理部４は、Ｓ０５にて膨張画像に付したラベルの個数ｎを求める。
次のＳ０７では、画像処理部４は、Ｓ０６にて求めたラベル個数ｎが０であるかどうかをチェックする。そして、ラベル個数ｎが０である場合には、Ｓ２０において「良である」旨を表示装置９上に表示した後に、検査処理を終了する。
【００４４】
これに対して、ラベル個数ｎが１以上である場合には、画像処理部４は、Ｓ０８において、変数ｉに１を代入した後に、Ｓ０９乃至Ｓ１１のループ処理を実行する。
【００４５】
このループ処理に入って最初のＳ０９では、画像処理部４は、第３メモリ８ｃに格納されている画像データ（膨張画像）中においてｉ番目の識別番号（ラベル）が付されている膨張領域を特定する。そして、特定したｉ番目の膨張領域に対応した二値化画像中の白い領域をサーチし、サーチした全ての白い領域に対してｉ番目の識別番号（ラベル）を付してグループ化する（図８（ａ）参照）（グループ化手段に相当）。なお、ここで、「ｉ番目の膨張領域に対応した二値化画像中の白い領域」とは、膨張画像中におけるｉ番目の膨張領域が占める場所と同じ位置関係にある二値化画像中の場所に包含されている白い領域を、意味している。
【００４６】
次のＳ１０では、画像処理部４は、変数ｉを一つインクリメントする。
次のＳ１１では、画像処理部４は、現時点における変数ｉの値がラベル個数ｎと等しいか否かをチェックする。そして、未だ、変数ｉの値がラベル個数ｎよりも小さい場合には、処理をＳ０９に戻し、次の膨張領域についての処理を実行する。
【００４７】
これに対して、以上のループ処理を繰り返した結果変数ｉの値がラベル個数ｎと一致した場合には、画像処理部４は、Ｓ１２において、第１メモリ８ａに格納されている撮像画像を反転して、第３メモリ８ｃに上書きする。即ち、撮像画像を構成する各画素の輝度値を２５５から減じた計算結果を、第３メモリ８ｃの対応箇所に書き込むのである。
【００４８】
次のＳ１３では、画像処理部４は、第２メモリ８ｂに格納されている二値化画像と第３メモリ８ｃに格納されている画像データ（反転画像）との論理積を算出する（抽出手段に相当）。即ち、二値化画像を構成する画素の値（８ビットパラレルのデジタル値）及び反転画像を構成する画素の値（８ビットパラレルのデジタル値）のうちの対応するもの同士の論理積を算出し、算出結果を第３メモリ８ｃに上書きするのである。この論理積算出の結果、第３メモリ８ｃに格納されている反転画像のうち、二値化画像の白い画素［２５５］の領域に対応する部分のみがそのまま抽出され、他の部分の画素の数値は全て０となる。
【００４９】
次のＳ１４では、画像処理部５は、Ｓ１３の結果抽出された各領域（二値化抽出物）の面積を夫々計測し、各識別番号（ラベル）によってグループ化された領域毎に、その総和を算出する（数値化手段に相当）。
【００５０】
次のＳ１５では、画像処理部５は、Ｓ１３の結果抽出された各領域（二値化抽出物）の平均輝度を、各識別番号（ラベル）によってグループ化された領域毎に、夫々算出する（数値化手段に相当）。
【００５１】
次のＳ１６では、画像処理部５は、各識別番号（ラベル）によってグループ化された領域毎に、Ｓ１４にて算出された面積総和とＳ１５にて算出された平均輝度とを乗算することによって「評価値」を算出する（数値化手段に相当）。
【００５２】
次のＳ１７では、画像処理部５は、Ｓ１６にて算出した「評価値」に基づいて検査対象光学部材の良否判定を行う。即ち、所定の閾値を超えている「評価値」が一つでもあれば不良品であると判定し、所定の閾値を超えている「評価値」が一つもなければ良品であると判定する（判定手段に相当）。
【００５３】
Ｓ１７にて不良品であると判定した場合には（Ｓ１８）、画像処理部５は、Ｓ１９において、「不良である」旨を表示装置９上に表示する。これに対して、Ｓ１７にて良品であると判定した場合には（Ｓ１８）、画像処理部５は、Ｓ２０において、「良である」旨を表示装置９上に表示する。以上により、検査処理が終了する。
＜光学部材検査装置による検査手順＞
本実施形態による光学部材検査装置によって光学部材を検査する時には、検査者は、検査対象光学部材をサンプル保持台６上に固定する。この際、その検査対象光学部材の入射面及び出射面のうち、ファインダー内では接眼レンズＬ３側に位置する面が、撮影レンズＬに対向される。
【００５４】
次に、検査者は、この検査対象光学部材が組み込まれるファインダー中の接眼レンズの焦点距離，人間の瞳径（３ｍｍ），検査対象光学部材の撮影レンズＬに対向している面の最大径（対角線長）及び撮像倍率（Ｍ）から、上述の式（９）に基づいて撮影レンズＬの焦点距離（ｆ）を求める。
【００５５】
次に、検査者は、接眼レンズＬ３の焦点距離及び人間の瞳径（３ｍｍ）から、上述の式（2'）に基づいて撮影レンズＬの物体側有効Ｆナンバー（Ｆ_OB）を決定する。そして、決定された物体側有効Ｆナンバー（Ｆ_OB）に合わせて絞り１１を絞り込む。
【００５６】
次に、検査者は、表示装置９上に表示される検査対象光学部材の映像を見ながらレンズ鏡筒４内の図示せぬ焦点調節機構を操作し、この検査対象光学部材がファインダーに組み込まれた状態において対物レンズ群Ｌ１，Ｌ２による像の結像位置（電子式ファインダーの場合には液晶パネル）により近くなる側の面に、撮像素子５ａに対する撮影レンズＬのピント位置（又は、撮影レンズＬと補正レンズ１２とからなる光学系のピント位置）を合わせる。
【００５７】
この作業が完了した後に、検査者は、光源１０を点灯させるとともに画像処理部４に接続された図示せぬ検査開始ボタンを押下して、画像処理をスタートさせる。検査対象光学素子を通過した光を撮影レンズＬが収束することよって形成される映像が、撮像素子５ａによって撮像される（Ｓ０２）。このようにして撮像される映像の模式図を、図６に示す。図６において、欠陥候補１は、像の結像位置により近くなる側の面上における小さなゴミに対応する暗部であり、欠陥候補２は、像の結像位置により近くなる側の面上における大きなゴミに対応する暗部であり、欠陥候補３は、像の結像位置により近くなる側の面からやや離れた面上におけるゴミに対応する暗部である。
【００５８】
画像処理部８は、画像データ中における所定の閾値よりも暗い領域のみを反転・抽出する（Ｓ０３〜Ｓ１３）。図６の画像に含まれる各暗部（欠陥候補１〜３）は、その輝度断面を表す図７に示されている様に、何れも所定の閾値よりも暗い。従って、これら各欠陥候補１〜３が、夫々反転・抽出されるのである。なお、この時、各領域（二値化領域）に関して、膨張処理（Ｓ０４）によって互いの周縁同士が繋がるか否かに基づいて、相互間の距離が所定範囲内であるかどうかの識別がなされる。そして、相互間の距離が所定範囲内であると識別された各領域（二値化領域）には、夫々同じ識別番号（ラベル）が付される（Ｓ０９）。その結果、図８（ａ）に示すように、見かけ上互いに近接している領域（二値化領域）同士については、それの元になった欠陥が光学部材の同一面上に在るか若しくは別の面上に在るかに拘わらず、同じ識別番号（ラベル）が付される。
【００５９】
そして、画像処理部８は、各識別番号（ラベル）毎に、その識別番号（ラベル）が付された全領域の面積の総和及び平均輝度を乗算することによって「評価値」を算出し、所定の基準値を超えている「評価値」が一つでもあるか否かに基づいて、検査対象光学部材の良否判定を客観的に行う。その結果、面積が非常に大きい領域（図６における欠陥候補２及び欠陥候補３）については、平均輝度の高低に拘わらず「評価値」が大きくなるので、検査対象光学部材が不良品であると判定することができる。また、平均輝度が非常に高い領域（図６における欠陥候補１及び欠陥候補２）については、面積の大小に拘わらず「評価値」が大きくなるので、検査対象光学部材が不良品であると判定することができる。これに対して、面積が小さく且つ平均輝度が低い領域については、「評価値」が小さくなるので、検査対象光学部材が良品であると判定することができる。なお、ある面積及び平均輝度が何れも中頃である領域については、「評価値」と閾値との大小関係如何に依って、良否が判定されることになる。
【００６０】
また、見かけ上互いに近接している領域（二値化領域）同士については、全体として一つの領域（二値化領域）として扱われて、一つの評価値が算出される。その結果、仮に個々の領域（二値化領域）について夫々単独で評価値を算出するのであれば各評価値が夫々判定基準値を下回ってしまう場合であっても、全体として算出した一つの評価値が判定基準値を上回っているのであれば、検査対象の光学部材を不良品として判定することができる。このようにして不良品として判定された光学部材は、仮にファインダー内に組み込まれた場合であっても、その欠陥が観察者によって大きく視認されてしまう。従って、この良否判定結果は、検査対象光学部材を組み込んだ光学機器の性能に対する実際の悪影響に即していると言える。
【００６１】
【発明の効果】
以上のように構成された本発明の光学部材検査装置及び光学部材検査方法によれば、光路長を有する光学部材を撮像して得た映像から欠陥を示す領域を抽出するとともに抽出された領域の図形的特徴量に基づいて良否判定を行うことができる。しかも、このとき、検査対象光学部材を組み込んだ光学機器の性能に対する実際の悪影響に即して、良否判定を行うことができる。
【図面の簡単な説明】
【図１】 本発明の実施の形態による光学部材検査装置の概略側視図
【図２】 図１における撮影レンズの光軸を直線化して示した光学構成図
【図３】 図２において補正レンズ１２を追加した状態を示す光学構成図
【図４】 図１の画像処理部８において実行される画像処理の内容を示すフローチャート
【図５】 周辺光量が低下した状態の画像例を示す模式図
【図６】 欠陥がある場合の画像例を示す模式図
【図７】 図６における輝度断面を示す図
【図８】 膨張処理及びラベリングの説明図
【図９】 検査対象光学部材としてのポロプリズムを含むファインダー光学系の斜視図
【図１０】 ポロプリズムの各面の位置，被写界深度，及び良否判定の厳しさの関係を示す図
【図１１】 薄い濃度の領域が二値化によって複数の領域に分かれてしまう様子を示す説明図
【図１２】 近接している複数の欠陥がファインダー内において一つの大きな欠陥として見えてしまう様子を示す説明図
【符号の説明】
５ａ 撮像素子
７ 拡散板
１０ 光源
１１ 絞り
Ｌ 撮影レンズ
Ｌ３ 接眼レンズ
Ｐ ポロプリズム[0001]
[Technical field to which the invention belongs]
The present invention relates to an optical member inspection apparatus and an optical member inspection method for detecting a defect in an optical member.
[0002]
[Prior art]
FIG. 9 shows an optical member constituting a real-image finder used for a compact camera or a video camera as an example of the optical member. In FIG. 9, the light from the object travels along the optical axis A while being converged by the objective lens groups L1 and L2. After the right and left are reversed by the right-angle prism M, a real image of the object is once formed inside the field frame F formed immediately before the first surface P ₁ of the Porro prism P. After that, the light enters the Porro prism P from the first surface P ₁ while diverging, and is internally reflected in the order of the second surface P ₂ , the third surface P ₃ , and the fourth surface P ₄ . through the fifth surface P ₅ emitted to the outside. Then, after being refracted by the eyepiece lens L3, the light enters the observer's pupil and forms a real image of the object and the field frame F on the observer's retina.
[0003]
Each of such optical members has an optical path length in the inside thereof, and therefore the presence or absence of defects (dirt, dust, scratches, cracks, bubbles, etc.) must be inspected over the entire optical path. For example, in the case of the Porro prism P in FIG. 9, bubbles and dust are mixed in the entire area, the incident surface (first surface P ₁ ), the reflective surface (second surface P ₂ to fourth surface P ₄ ), and the emission surface. The surface (fifth surface P ₅ ) needs to be inspected for the presence or absence of dirt, dust, scratches, or the like.
[0004]
[Problems to be solved by the invention]
In order to meet such a need, an image of the optical member is captured by the imaging device, the image captured by the imaging device is binarized by a predetermined threshold value, and the binarization is determined to be darker than the threshold value. It is conceivable to perform pass / fail judgment based on the graphical feature amount (area x density of each region, etc.).
[0005]
However, when imaging an optical member having an optical path length as described above, there is only one position on the imaging optical axis where the imaging device is in focus, so depending on the position of the defect, the defect is in focus. May or may not match. Even if the defect has the same size and the same property (scratch, dirt, etc.), the defect at the focused position is reflected in the image as a relatively dark area with a relatively small area, A defect at a position out of focus is reflected in the image as a relatively dark area with a relatively large area due to blurring. For this reason, the density of the dark region corresponding to the defect far away from the in-focus position may be as thin as the threshold. When such a dark region is binarized, as shown in FIG. 11, due to slight shading in the dark region, a dark region that should originally be one is a plurality of regions. It will be divided into.
[0006]
However, since such a dark region should have an adverse effect as a defect when the optical member is used, the graphic feature amount of each region divided by binarization is calculated individually. Is irrational. In other words, even when the original area of the defect is large enough to adversely affect its performance when using an optical device incorporating this optical member, the area of each region after binarization becomes extremely small This is because the target feature amount becomes small, so that the optical member is not determined as a defective product.
[0007]
For dark areas with a small area, it is possible to prevent them from being divided into a plurality of regions by devising the preprocessing for binarization to some extent, but this preprocessing itself takes time. There is a problem of end.
[0008]
In addition, when the prism P incorporated in the finder as shown in FIG. 9 is an optical member to be inspected, a plurality of defects are present on the same surface or separate surfaces of the prism P. As shown in FIG. 12 (a), a plurality of independent dark regions may appear adjacent to each other in the image by the image sensor.
[0009]
However, in the finder as shown in FIG. 9, the focus position by the imaging device at the time of inspection and the formation position of the image (real image by the objective lens groups L1 and L2) when the optical member to be inspected is incorporated into the finder. (The position of the field frame F) is different. Accordingly, when the finder is completed, when the observer looks into the finder and focuses the image (real image by the objective lens groups L1 and L2), each defect is blurred due to the focus shift. It is visually recognized as one large defect as shown in b). Therefore, as described above, it is unreasonable to individually calculate each graphic feature amount for a plurality of adjacent dark regions that appear as one defect when the finder is actually used. In other words, even if the area of each region is small and the graphic feature amount is small, if it is visually recognized as a large defect when the observer looks into the finder, the performance is adversely affected.
[0010]
The present invention has been made in view of such a point, and extracts a region indicating a defect from an image obtained by imaging an optical member having an optical path length and is based on a graphic feature amount of the extracted region. Optical member inspection apparatus and optical member inspection method for performing pass / fail determination, and optical member inspection apparatus and optical member capable of performing pass / fail determination in accordance with an actual adverse effect on the performance of an optical device incorporating an inspection target optical member The issue is to provide inspection methods.
[0011]
[Means for Solving the Problems]
The present invention employs the following configuration in order to solve the above problems. That is, according to the first aspect of the present invention, in the optical member inspection apparatus for inspecting an optical member, an imaging lens disposed opposite to the inspection target surface of the optical member and an image formed by the imaging lens are captured. Imaging means, extraction means for extracting a region indicating a defect of the optical member in an image captured by the imaging means, and expansion processing for expanding the periphery of the region indicating the defect extracted by the extraction means together subjected to, and grouping means for grouping the region between indicating the defect results whose periphery led the expansion process, each grouped group by the grouping unit, the defects included in the group all areas the expansion processing the previous state to each digitize to calculate the sum, digitizing means for calculating an evaluation value based on the sum indicated Characterized in that the evaluation value calculated by the digitizing means and a determining means for determining whether more than a predetermined criterion value.
[0012]
According to the first aspect of the present invention, the imaging means captures an image formed by the photographing lens. An image obtained by capturing this image is reflected in a state where defects at respective positions along the optical axis of the finder are superimposed. The extraction unit extracts all regions indicating defects of the optical member from the image captured by the imaging unit. The grouping means groups adjacent areas among the areas extracted by the extracting means. The digitizing means treats the grouped area as one area as a whole, and digitizes each state. The determination unit determines whether or not the optical member to be inspected is acceptable based on whether or not the digitized numerical value exceeds a predetermined determination reference value. Therefore, for a defect that is visually recognized as a single defect by an observer when using an optical apparatus incorporating an inspection target optical member, the areas corresponding to the defect are close to each other. Are treated as a single area. When the entire state is digitized and exceeds the determination reference value, the optical member to be inspected is treated as a defective product.
[0014]
The invention according to claim 2 is specified by the grouping means according to claim 1 adding common identification information to regions indicating the defects connected to the periphery as a result of the expansion process.
[0015]
The expansion invention of claim 3, wherein, grouping means according to claim 2, with subjecting the individual identification information to the husband of the area results became single monolithic expansion processing s, corresponding to the area of each single monolithic The identification information given to the one continuous area is given to each area indicating the defect before processing.
[0016]
According to a fourth aspect of the present invention, for each of the groups grouped by the grouping unit, the numerical unit of the first aspect includes a total sum of the areas before the expansion process of all the regions indicating the defect included in the group. It is specified by calculating the evaluation value based on the luminance.
[0017]
According to a fifth aspect of the present invention, for each of the groups grouped by the grouping unit, the numerical unit of the first aspect includes a sum of areas before the expansion process of all regions indicating the defect included in the group. It is specified by calculating the evaluation value by multiplying the luminance value.
[0018]
The invention according to claim 6 is an optical member inspection method in which an optical member is inspected by an inspection apparatus having an imaging lens and an imaging means for capturing an image formed by the imaging lens. The optical member is arranged on the optical axis of the photographing lens in a state facing the photographing lens, an image formed by the photographing lens is picked up by the image pickup means, and the image picked up by the image pickup means A region indicating the defect of the optical member is extracted, an expansion process is performed to expand the periphery of the extracted region indicating the defect, and the regions indicating the defect connected to the periphery as a result of the expansion process are grouped. However, for each group grouped, and the sum by the respectively digitizing state before expansion processing of all areas indicating the defects contained in the group Out, calculates an evaluation value based on the sum, calculated evaluation value to determine whether more than a predetermined criterion value, characterized.
[0019]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. In this embodiment, the optical member inspection apparatus and the optical member inspection method according to the present invention are used to inspect the Porro prism P included in the viewfinder shown in FIG. 9 that allows an observer to observe an image of an object through an eyepiece lens L3. It is constructed so as to be suitable for.
<Configuration of optical member inspection device>
FIG. 1 is a schematic side view of the optical member inspection apparatus according to the present embodiment. As shown in FIG. 1, a support column 2 is erected in the vicinity of one end edge of a base 1 of the optical member inspection apparatus. A lens barrel 4 is fixed to a side surface in the vicinity of the tip of the column 2 via a stay 3. A television camera 5 incorporating an image pickup element (CCD area sensor) 5a as an image pickup means is attached to the opening end of the lens barrel 4 opposite to the base 1.
[0020]
The lens barrel 4 described above includes a photographing lens L which is composed of a positive lens group as a whole and forms an image on the surface of the image sensor 5a. The optical axis of the photographic lens L faces the direction parallel to the axial direction of the support column 2 and passes through the center of the image sensor 5a. Further, the photographic lens L can be adjusted along the optical axis by a focus adjusting mechanism (not shown) to adjust the focus position with respect to the surface of the image pickup device 5a (a position conjugate with the surface of the image pickup device 5a). it can. Further, in the photographing lens L, an aperture 11 is provided that can freely set an aperture value.
[0021]
On the other hand, a sample holder 6 for fixing the optical member to be inspected is disposed on the extended line of the optical axis of the photographing lens L on the upper surface of the base 1. The Porro prism P as an optical member to be inspected is fixed on the sample holder 6 with the fifth surface P ₅ arranged on the eyepiece L3 side facing the photographing lens L when the finder is completed. More specifically, the fifth surface P ₅ of the Porro prism P is fixed on the sample holder 6 so as to be orthogonal to the optical axis of the photographing lens L at the center thereof. As a result, the optical axis of the photographic lens L is bent by the reflecting surfaces (fourth surface P ₄ , third surface P ₃ , second surface P ₂ ) in the Porro prism P, respectively, and the first surface P ₁ . Orthogonal. As shown in FIG. 9, since the first surface P ₁ is orthogonal to the fifth surface P ₅ , the optical axis of the photographing lens L is finally bent by 90 degrees by the Porro prism P, It extends to the side of the Porro prism P. The optical axis bent by the Porro prism P passes through the center of the diffusion plate 7 and hits the center of the light source 10.
[0022]
The diffuser plate 7 is a diffuser that diffuses illumination light emitted from the light source 10 and introduces it into the first surface P ₁ of the Porro prism P. The diffuser plate 7 constitutes illumination means together with the light source 1.
[0023]
Note that image data output from the image sensor 5 a in the television camera 5 is input to the image processing device 8. The image processing apparatus 8 is a processor as an extracting unit, a grouping unit, a digitizing unit, and a determining unit that determine whether an optical member to be inspected is a non-defective product or a defective product. In other words, the image processing apparatus 8 performs predetermined image processing on the image data input from the image sensor 5a, digitizes the degree of defects of the optical member to be inspected, and converts this numerical value to a certain criterion value ( It is determined whether this numerical value is within the criterion value or exceeded. The image processing apparatus 8 includes a first memory 8a, a second memory 8b, and a third memory 8c in order to execute these processes.
[0024]
The display device 9 displays the image based on the image data from the image sensor 5a (the image captured by the image sensor 5a) as it is, and displays the determination result by the image processing device 8.
[0025]
Next, various conditions for the eyepiece lens L will be described.
2 omits the base 1, the support 2, the stay 3, and the display device 9 from FIG. 1, linearizes the optical axis of the photographing lens L, and moves the Porro prism P along the linearized optical axis. FIG. 2 is an optical configuration diagram expressing the Porro prism P as a quadrangular prism-shaped optical member by developing.
[0026]
In FIG. 2, the focus position of the photographic lens L with respect to the image sensor 5 a is matched with the first surface P ₁ of the Porro prism P. Therefore, when there is a defect such as dust or scratches on the first surface P ₁ is the image of these defects are clearly formed on the surface of the image pickup device 5a.
[0027]
Incidentally, in general, there is a relationship represented by the following formula (1) among the focal length (f), the entrance pupil diameter (D), and the F number of the lens (positive lens system).
F = f / D (1)
Therefore, the object side effective F number (F _OA ) of the optical system including the eyepiece lens L3 of the finder in which the Porro prism P is incorporated and the pupil of the observer is expressed by the following formula if the diopter of the eyepiece lens L3 is 0 diopter. It is represented by (2).
[0028]
F _OA = Focal distance of eyepiece / Observer pupil diameter (2)
Usually, the focal length of the eyepiece lens L3 is less than 30 mm, and the diameter of the human pupil in a bright place is about 3 mm. Therefore, F _OA is expressed by the following formula (2 ′).
[0029]

If the object side effective F number (F _OA ) of the optical system composed of the eyepiece L3 and the observer's pupil calculated in this way is the same as the object side effective F number (F _OB ) of the taking lens L, observation is performed. An image having the same appearance as when the person looks into the eyepiece lens L3 is captured by the image sensor 5a. However, in actuality, since the human eye can unconsciously adjust the focus, it is possible to focus on a certain range before and after the focus surface. Also, it is necessary to make the inspection somewhat strict.
[0030]
Accordingly, the diaphragm 11 of the imaging lens L has the formula (2) by (pupil diameter of the focal length / observer eyepiece) F _OA value obtained corresponding aperture above (formula (2 ') is satisfied It has been narrowed down to 10 or more under the premise. As a result, the depth of field of the imaging lens L is the same as the depth of field when the observer looks into the eyepiece lens L3. Accordingly, the image that is seen when the observer looks into the eyepiece lens L3 and focuses the eyes on the real image of the object is captured by the image sensor 5a in the same way. That is, the defect on the first surface P ₁ clearly seen when the observer looks into the eyepiece lens L3 and focuses the eye on the real image of the object, and the Porro prism P included in the depth of field (X). The internal defects (defects in the area indicated by the mesh in FIG. 9) are clearly shown in the video imaged by the image sensor 5a. Further, the depth of field outside the defects appear blurred with distance from the first surface P ₁ becomes longer when the observer to focus the eye on a real image of the object looking through the ocular lens L3 by the image pickup device 5a Even in the captured image, the image is blurred according to the distance from the first surface P ₁ .
[0031]
In addition, according to Formula (1), when F number is constant, it turns out that the focal distance (f) of the imaging lens L influences the entrance pupil diameter (D) of this imaging lens L. If this entrance pupil diameter (D) becomes too small, the amount of light diffused from the diffuser plate 7 will diverge from the portion on the optical axis of the diffuser plate 7 and enter the entrance pupil through the center of the Porro prism P. The amount of light that diverges from the peripheral portion and enters the entrance pupil through the vicinity of the side surface of the Porro prism P is extremely reduced. As a result, as shown in FIG. 5, even in a video image captured by the imaging device 5a, the peripheral portion of the region corresponding to the first surface P ₁ of the Porro prism P (▲ 2 ▼) is a central portion (▲ 1 It will be darker than ▼). If such a light amount difference becomes larger than the allowable range, there is a possibility that in the image processing, this dark part is regarded as a defect even though it is not a defect.
[0032]
The occurrence of such a light quantity difference is caused by making at least all light rays parallel to the optical axis incident on the entrance pupil of the imaging lens L out of the light rays emitted from the fifth surface P ₅ of the Porro prism P. It can be suppressed to the extent that it does not affect the extraction. Thus, in order to make all the light beams emitted from the fifth surface P ₅ of the Porro prism P parallel to the optical axis enter the entrance pupil of the imaging lens L, the entrance pupil diameter (D) of the imaging lens L is set to the Porro prism. diagonal length of the fifth surface P ₅ of P (d) may be greater than. This relationship is expressed by the following formula (3).
[0033]
D> d (3)
On the other hand, in the equation (1), when F is replaced with the object-side effective F number (F _O ) and f is replaced with the distance a from the object to the principal point (where a> 0), the following equation (4) is obtained.
[0034]
F _O = a / D
D = a / F _O (4)
Therefore, the following formula (5) is obtained by substituting the formula (4) into the formula (3).
[0035]
a / F _O > d
a> d · F _O (5)
If the distance from the principal point to the image is b (where b> 0), the relationship between the focal length f and the image magnification M is expressed by the following equations (6) and (7).
[0036]
b / a = M (6)
1 / a + 1 / b = 1 / f (7)
When the focal length f is solved from these equations (6) and (7), equation (8) is obtained.
[0037]
f = a (M / (M + 1)) (8)
Substituting equation (5) into equation (8) yields equation (9) below.
f> d · F _O · (M / (M + 1))
> Diagonal length of fifth surface P ₅ of Porro prism P
Focal length of eyepiece L3 / observer pupil diameter
(M / (M + 1)) (9)
Since the focal length (f) of the imaging lens L is set according to this equation (9), all the light rays parallel to the optical axis are incident on the entrance pupil of the imaging lens L. Note that M is set based on the size of the light receiving surface of the image sensor 5a.
[0038]
Note that the correction lens 12 shown in FIG. 3 is inserted for pupil alignment.
<Image processing>
Next, the contents of the image processing executed in the image processing unit 8 for determining whether the inspection target optical member (Poro prism) is a non-defective product or a defective product will be described with reference to the flowchart of FIG. .
[0039]
This image processing starts when an unillustrated inspection start button connected to the image processing unit 4 is pressed. Then, in the first step S01 after the start, the image processing unit 4 displays an instruction to set the sample (inspection target optical member) on the display device 9.
[0040]
In the next S02, the image processing unit 4 captures one frame of the image data input from the image sensor 5a into the first memory 8a.
In the next S03, the image processing unit 4 corresponds from the image data stored in the first memory 8a (captured image), the inspection target region (region corresponding to the fifth surface P ₅ Porro prism P) moiety And binarization processing is performed (corresponding to extraction means). In this binarization process, the image processing unit 4 performs a smoothing process on the luminance values (0 to 255) of the respective pixels constituting the extracted image data, and the luminance values ( The threshold value data is created by uniformly subtracting a predetermined amount from 0 to 255). Then, the luminance value (0-255) of each pixel constituting the image data extracted from the inspection target area is compared with the luminance value (0-255) of the corresponding pixel in the threshold data, and the former is lower than the latter. For pixels that are present, the luminance value 255 is written at the corresponding position in the second memory 8b, and for the pixels where the former is greater than or equal to the latter, the luminance value 0 is written at the corresponding position in the second memory 8b. In the image data (binarized image) stored in the second memory 8b at the time of completion of the binarization processing, only the portion corresponding to the dark portion that is dark enough to be noticed when viewed by the observer is whitened. It is an image.
[0041]
In next S04, the image processing unit 4 reads out the image data (binarized image) from the second memory 8b, performs the expansion process on the read image data, and performs the expansion process on the image data (expanded image). ) In the third memory 8c (corresponding to grouping means). This expansion process is a process of expanding the periphery of the white area in the binarized image outward by a predetermined amount. As a result of this expansion processing, areas adjacent to each other as shown in FIG. 8A (however, in FIG. 8, for the sake of convenience, black and white are reversed are drawn) are as shown in FIG. 8B. , Are connected to each other to form a continuous expansion region.
[0042]
In the next S05, the image processing unit 4 labels the image data (expanded image) stored in the third memory 8c (corresponding to grouping means). That is, as shown in FIG. 8B, a unique identification number (label) is assigned to each expansion region in the image data (expansion image).
[0043]
In next S06, the image processing unit 4 obtains the number n of labels attached to the expanded image in S05.
In next S07, the image processing unit 4 checks whether or not the number of labels n obtained in S06 is zero. If the number of labels n is 0, “good” is displayed on the display device 9 in S20, and the inspection process is terminated.
[0044]
On the other hand, when the number of labels n is 1 or more, the image processing unit 4 executes the loop processing of S09 to S11 after substituting 1 for the variable i in S08.
[0045]
In the first step S09 after entering this loop process, the image processing unit 4 displays the expansion area to which the i-th identification number (label) is attached in the image data (expansion image) stored in the third memory 8c. Identify. Then, a white area in the binarized image corresponding to the identified i-th expanded area is searched, and the i-th identification number (label) is assigned to all the searched white areas to be grouped (see FIG. 8 (a)) (corresponding to grouping means). Here, the “white region in the binarized image corresponding to the i-th dilated region” is the binarized image in the same positional relationship as the location occupied by the i-th dilated region in the dilated image. It means the white area included in the place.
[0046]
In next S10, the image processing unit 4 increments the variable i by one.
In next step S11, the image processing unit 4 checks whether or not the current value of the variable i is equal to the number of labels n. If the value of the variable i is still smaller than the label number n, the process returns to S09, and the process for the next expansion region is executed.
[0047]
On the other hand, if the value of the variable i matches the number of labels n as a result of repeating the above loop processing, the image processing unit 4 inverts the captured image stored in the first memory 8a in S12. Then, the third memory 8c is overwritten. That is, the calculation result obtained by subtracting the luminance value of each pixel constituting the captured image from 255 is written in the corresponding portion of the third memory 8c.
[0048]
In the next S13, the image processing unit 4 calculates a logical product of the binarized image stored in the second memory 8b and the image data (inverted image) stored in the third memory 8c (extraction means). Equivalent). In other words, the logical product of the corresponding values of the pixel value (8-bit parallel digital value) constituting the binarized image and the pixel value (8-bit parallel digital value) constituting the inverted image is calculated. The calculation result is overwritten in the third memory 8c. As a result of the logical product calculation, out of the inverted image stored in the third memory 8c, only the portion corresponding to the white pixel [255] region of the binarized image is extracted as it is, and the numerical values of the pixels of the other portions are extracted. All become 0.
[0049]
In the next S14, the image processing unit 5 measures the area of each region (binarized extract) extracted as a result of S13, and sums up each region grouped by each identification number (label). Is calculated (equivalent to a numerical means).
[0050]
In next S15, the image processing unit 5 calculates the average luminance of each region (binarized extract) extracted as a result of S13 for each region grouped by each identification number (label) ( Equivalent to numerical means).
[0051]
In next S16, the image processing unit 5 multiplies the total area calculated in S14 by the average luminance calculated in S15 for each region grouped by each identification number (label). "Evaluation value" is calculated (corresponding to numerical means).
[0052]
In the next S17, the image processing unit 5 determines the quality of the optical member to be inspected based on the “evaluation value” calculated in S16. That is, if there is even one “evaluation value” exceeding a predetermined threshold, it is determined as a defective product, and if there is no “evaluation value” exceeding a predetermined threshold, it is determined as a non-defective product ( Equivalent to determination means).
[0053]
If it is determined in S17 that the product is a defective product (S18), the image processing unit 5 displays “not good” on the display device 9 in S19. On the other hand, if it is determined that the product is non-defective in S17 (S18), the image processing unit 5 displays “good” on the display device 9 in S20. Thus, the inspection process ends.
<Inspection procedure by optical member inspection device>
When inspecting the optical member by the optical member inspection apparatus according to the present embodiment, the inspector fixes the optical member to be inspected on the sample holder 6. At this time, of the entrance surface and the exit surface of the optical member to be inspected, the surface located on the eyepiece L3 side in the viewfinder faces the photographing lens L.
[0054]
Next, the examiner sets the focal length of the eyepiece lens in the finder in which the inspection target optical member is incorporated, the human pupil diameter (3 mm), and the maximum diameter of the surface of the inspection target optical member facing the photographing lens L ( From the diagonal length) and the imaging magnification (M), the focal length (f) of the photographic lens L is obtained based on the above equation (9).
[0055]
Next, the examiner determines the object-side effective F number (F _OB ) of the photographic lens L from the focal length of the eyepiece lens L3 and the human pupil diameter (3 mm) based on the above equation (2 ′). Then, the diaphragm 11 is narrowed down in accordance with the determined object-side effective F number (F _OB ).
[0056]
Next, the inspector operates a focus adjustment mechanism (not shown) in the lens barrel 4 while viewing the image of the inspection target optical member displayed on the display device 9, and the inspection target optical member is incorporated into the viewfinder. In this state, the focus position of the photographic lens L with respect to the image sensor 5a (or the photographic lens L) is closer to the image forming position (liquid crystal panel in the case of an electronic viewfinder) of the image by the objective lens groups L1 and L2. And the focus position of the optical system consisting of the correction lens 12.
[0057]
After this operation is completed, the inspector turns on the light source 10 and presses an unillustrated inspection start button connected to the image processing unit 4 to start image processing. An image formed by the photographing lens L converging light that has passed through the inspection target optical element is imaged by the imaging element 5a (S02). A schematic diagram of the video imaged in this way is shown in FIG. In FIG. 6, defect candidate 1 is a dark part corresponding to small dust on the surface closer to the image formation position, and defect candidate 2 is larger on the surface closer to the image formation position. A dark part corresponding to dust, and the defect candidate 3 is a dark part corresponding to dust on a surface slightly away from the surface closer to the image formation position.
[0058]
The image processing unit 8 inverts and extracts only a region darker than a predetermined threshold in the image data (S03 to S13). Each dark part (defect candidates 1 to 3) included in the image of FIG. 6 is darker than a predetermined threshold as shown in FIG. Therefore, these defect candidates 1 to 3 are inverted and extracted, respectively. At this time, for each region (binarized region), whether or not the distance between each other is within a predetermined range is determined based on whether or not the peripheral edges are connected by the expansion process (S04). The The same identification number (label) is assigned to each region (binarized region) identified as having a distance between each other within a predetermined range (S09). As a result, as shown in FIG. 8A, for the regions (binarized regions) that are apparently close to each other, the defect that is the basis thereof is on the same surface of the optical member or The same identification number (label) is given regardless of whether it is on another surface.
[0059]
Then, for each identification number (label), the image processing unit 8 calculates an “evaluation value” by multiplying the sum of the areas of all the areas to which the identification number (label) is attached and the average luminance, Based on whether or not there is at least one “evaluation value” exceeding the reference value, the quality of the optical member to be inspected is objectively determined. As a result, since the “evaluation value” is large regardless of the average luminance level in the region having a very large area (defect candidate 2 and defect candidate 3 in FIG. 6), the optical member to be inspected is a defective product. Can be determined. In addition, in the region where the average luminance is very high (defect candidate 1 and defect candidate 2 in FIG. 6), the “evaluation value” becomes large regardless of the size of the area, so it is determined that the optical member to be inspected is defective. can do. On the other hand, since the “evaluation value” is small for a region having a small area and a low average luminance, it can be determined that the inspection target optical member is a non-defective product. In addition, for a region where both a certain area and average luminance are medium, pass / fail is determined depending on the magnitude relationship between the “evaluation value” and the threshold value.
[0060]
In addition, regions (binarized regions) that are apparently close to each other are treated as one region (binarized region) as a whole, and one evaluation value is calculated. As a result, if evaluation values are calculated individually for each region (binarized region), even if each evaluation value falls below the criterion value, one evaluation calculated as a whole If the value exceeds the determination reference value, the optical member to be inspected can be determined as a defective product. Thus, even if the optical member determined to be defective is incorporated into the finder, the defect is visually recognized by an observer. Therefore, it can be said that this pass / fail judgment result is in line with the actual adverse effect on the performance of the optical device incorporating the optical member to be inspected.
[0061]
【The invention's effect】
According to the optical member inspection apparatus and the optical member inspection method of the present invention configured as described above, an area indicating a defect is extracted from an image obtained by imaging an optical member having an optical path length, and the extracted area is Pass / fail judgment can be made based on the graphical feature amount. In addition, at this time, it is possible to make a pass / fail judgment in accordance with an actual adverse effect on the performance of the optical apparatus in which the optical member to be inspected is incorporated.
[Brief description of the drawings]
FIG. 1 is a schematic side view of an optical member inspection apparatus according to an embodiment of the present invention. FIG. 2 is an optical configuration diagram in which an optical axis of a photographing lens in FIG. 1 is linearized. FIG. 4 is a flowchart showing the contents of image processing executed in the image processing unit 8 of FIG. 1. FIG. 5 is a schematic diagram showing an example of an image in a state where the peripheral light amount is reduced. 6 is a schematic diagram showing an example of an image when there is a defect. FIG. 7 is a diagram showing a luminance cross section in FIG. 6. FIG. 8 is an explanatory diagram of expansion processing and labeling. FIG. 9 is a Porro prism as an optical member to be inspected. FIG. 10 is a perspective view of a finder optical system including a diagram. FIG. 10 is a diagram showing the relationship between the position of each surface of a Porro prism, the depth of field, and the severity of pass / fail judgment. FIG. Like being divided into areas Explanatory view showing a state where become visible as a large defect of one the illustration [12] in a plurality of defects are close to each finder that indicates the EXPLANATION OF REFERENCE NUMERALS
5a Image sensor 7 Diffuser 10 Light source 11 Aperture L Photo lens L3 Eyepiece P Porro prism

Claims (6)

In an optical member inspection apparatus for inspecting an optical member,
A photographic lens disposed opposite to the inspection target surface of the optical member;
Imaging means for imaging an image formed by the photographic lens;
Extracting means for extracting a region indicating a defect of the optical member in an image captured by the imaging means;
Grouping means for grouping the areas indicating the defects connected to the periphery as a result of the expansion process, while performing expansion processing to expand the periphery of the area indicating the defect extracted by the extraction means;
Each grouped group by the grouping unit, a state before the expansion process of the entire region showing the defects contained in the group and each digitized to calculate the total evaluation value based on the sum A numerical means for calculating
An optical member inspection apparatus comprising: determination means for determining whether or not the evaluation value calculated by the numerical means exceeds a predetermined determination reference value. The grouping means includes
An optical member inspection apparatus according to claim 1, wherein the subjecting the common identification information in the region between indicating the defect results whose periphery led the expansion process. The grouping means includes
With subjecting the individual identification information to the husband of the area results became single monolithic expansion processing people in each area showing the defect before the expansion process corresponding to the region of the individual single monolithic, in the region of the single monolithic The optical member inspection apparatus according to claim 2, wherein identification information is attached. It said digitizing means, for each grouped group by the grouping unit calculates the evaluation value based on the sum of the area before the expansion process of the entire region showing the defect in that group and brightness The optical member inspection apparatus according to claim 1. It said digitizing means, for each grouped group by the grouping unit, wherein the evaluation value by multiplying the sum and the luminance value of the area before the expansion process of the entire region showing the defects contained in the group The optical member inspection apparatus according to claim 1, wherein: In an optical member inspection method in which an optical member is inspected by an inspection apparatus having an imaging lens and an imaging unit that captures an image formed by the imaging lens.
In a state where the inspection target surface of the optical member is directed to the photographing lens, the optical member is disposed on the optical axis of the photographing lens,
The image formed by the photographing lens is imaged by the imaging means,
Extracting an area indicating a defect of the optical member in an image captured by the imaging unit;
Applying an expansion process to expand the periphery of the extracted region indicating the defect,
As a result of this expansion process, the areas showing the defects connected to the periphery are grouped together,
For each group grouped, the state before the expansion process of the entire region showing the defects contained in the group and each digitized to calculate the total sum, and calculates an evaluation value based on the sum,
An optical member inspection method, comprising: determining whether the calculated evaluation value exceeds a predetermined determination reference value.

Images