JP3973024B2

JP3973024B2 - Defect detection method and defect detection system using 8-neighbor point adjacent comparison method in imaging inspection apparatus

Info

Publication number: JP3973024B2
Application number: JP2002186739A
Authority: JP
Inventors: 康一梶山
Original assignee: V Technology Co Ltd
Current assignee: V Technology Co Ltd
Priority date: 2002-06-26
Filing date: 2002-06-26
Publication date: 2007-09-05
Anticipated expiration: 2022-06-26
Also published as: JP2004028836A

Description

【０００１】
【発明の属する技術分野】
本発明は、撮像検査装置における８近傍点隣接比較方式（又は８点近傍隣接比較方式）による斬新な欠陥検出方法、欠陥検出システムに関する。
【０００２】
【従来の技術】
従来、例えば撮像検査装置を使用して液晶板等の液晶パターン（例えばＴＦＴ表示部）のように繰り返し同一のパターンが存在する被検査体の欠陥検出を行う場合、撮像装置にて当該パターン群を撮像し、その画像データから欠陥検出を行うようにしているが、このような欠陥検出方法は大きく分けるとパターンの絶対的な位置情報を基に欠陥を検出する方法（Pattern Matching等）と、パターンの位置とは独立に欠陥を検出する方法（隣接比較、ＤＲＣ（Design Rule Check）等）に分けられる。
ＬＣＤやＰＤＰ検査に適用する場合には、後者の方法が位置合わせの誤差がないため信頼性が高く一般的に採用されている方法である。
【０００３】
従来の撮像検査装置における近傍点隣接比較方式による欠陥検出方法について以下に説明する。
従来における欠陥検出方法は、撮像検査装置の撮像系にてパターン（例えばＴＦＴ表示部パターン）を撮像し、図１３に示すような多数のパターン像を含む画像データ５０を収集して、検査対象点Ａを挟んで隣接する４点（１），（２），（３），（４）のうち、左右又は上下の隣接点（２），（３）又は隣接点（１），（４）の平均値（輝度データ平均値）のどちらか一方を選定し検査対象点Ａとの比較点として検査対象点Ａの欠陥検出を行っていた。
又は、左右、上下の４個の隣接点（２），（３）、隣接点（１），（４）を使用しても、その４点の平均値と検査対象点Ａとの比較を行うだけであった。
【０００４】
【発明が解決しようとする課題】
しかしながら、上述した従来の欠陥検出方法の場合には、基板の隅部側に存在するパターンの欠陥検出において、左右方向（又は上下方向）の隣接２点の比較を行うと、一方の隣接点は基板上となるため、左右（又は上下）２点の輝度データが大きく異なることになり、このため、擬似欠陥が多発するという不都合があった。
本発明は、上記事情に鑑みてなされたものであり、欠陥検出過程を改良し、基板等の隅部領域等においても擬似欠陥が発生することがなく、高精度で欠陥検出を行うことができる撮像検査装置における８近傍点隣接比較方式による欠陥検出方法、欠陥検出システムを提供することを目的とするものである。
【０００５】
【課題を解決するための手段】
上記課題を解決するために、請求項１記載の発明は、被検査体を撮像して得られる同一繰り返しパターンの検査部位の画像データを基に、８近傍点隣接比較方式により個々の検査部位の欠陥検出を行う撮像検査装置における８近傍点隣接比較方式による欠陥検出方法において、検査対象点を挟んで隣接する８点のうち、左右、上下又は斜め方向に隣接する３種の２点同士を優先順位を付けて順に比較し比較対象の適否判定を行う予備判定過程と、予備判定過程の判定結果に応じて検査対象点との比較に用いる最適な比較方向の２点の選定を行う選定過程と、選定過程にて選定した最適な比較方向の２点の平均値と検査対象点とを比較し、当該検査対象点の欠陥の有無の検出を行う過程と、を含むことを特徴とするものである。
【０００６】
請求項２記載の発明は、請求項１記載の撮像検査装置における８近傍点隣接比較方式による欠陥検出方法において、前記予備判定過程は、検査対象点を挟んで隣接する左右、上下又は斜め方向の各２点を任意の順序で優先させて比較対象の適否判定を行うことを特徴とするものである。
【０００７】
請求項３記載の発明は、請求項１又は２記載の撮像検査装置における８近傍点隣接比較方式による欠陥検出方法において、前記予備判定過程は、コンピュータ制御におけるパイプライン処理により左右、上下又は斜め方向に隣接する３種の２点同士を優先順位を付けて順に比較することを特徴とするものである。
本発明の欠陥検出方法によれば、欠陥検出過程を改良し、検査対象点を挟んで隣接する８点のうち、左右、上下又は斜め方向に隣接する３種の２点同士を優先順位を付けて順に比較し比較対象の適否判定を行う予備判定過程を設け、予備判定過程の判定結果に応じて検査対象点との比較に用いる最適な比較方向の２点の選定を行い、選定した最適な比較方向の２点の平均値と検査対象点とを比較し、当該検査対象点の欠陥の有無の検出を行うものであるから、特に基板等の隅部領域等においても擬似欠陥が発生することがなく、高精度で検査対象点の欠陥検出を行うことができる。
【０００８】
請求項４記載の発明は、被検査体を撮像して得られる同一繰り返しパターンの検査部位の画像データを基に、８近傍点隣接比較方式により個々の検査部位の欠陥検出を行う撮像検査装置における８近傍点隣接比較方式による欠陥検出システムにおいて、被検査体を撮像し、撮像素子に結像する光源、レンズを含む撮像系と、撮像素子から出力される光電変換された画像データを画像処理し欠陥検出用の画像データを生成する画像処理部と、全体の制御を行う制御部と、制御部の制御の基に前記画像データにおける検査対象点を挟んで左右、上下又は斜め方向に隣接する８点のうち、いずれかの方向の隣接２点の輝度データの各々の優先順位を付けた比較演算、平均値演算等の各種の演算処理を行う演算処理部と、演算処理部の演算結果から検査対象点との比較に用いる最適な比較方向の２点の選定を行う選定部と、選定した最適な比較方向の２点の平均値と検査対象点の輝度データとを比較し、当該検査対象点の欠陥の有無の検出を行う欠陥検出部とを有することを特徴とするものである。
本発明の欠陥検出システムによれば、前記撮像系と、画像処理部と、制御部と、画像処理部と、演算処理部と、選定部と、欠陥検出部とを有する構成で、上述した欠陥検出方法を実現し、特に基板等の隅部領域等においても擬似欠陥が発生することがなく、高精度で検査対象点の欠陥検出を行うことができる。
【０００９】
【発明の実施の形態】
以下に本発明の実施の形態について詳細に説明する。
（実施の形態１）
図１は本発明の実施の形態１の撮像検査装置の全体構成を示す概略ブロック図であり、この撮像検査装置は、例えば多数の同一パターン、すなわちＴＦＴ表示部、配線部のパターンが表面に上下左右に列設された液晶板２０を撮像し、撮像素子（ＣＣＤ素子）４に結像する光源２、レンズ３を含む撮像系１と、撮像素子４から出力される光電変換された画像データを画像処理し欠陥検出用の画像データを生成する画像処理部５と、全体の制御を行う制御部（ＣＰＵ）６と、制御部６の制御の基に前記画像データにおける検査対象点Ａを挟んで左右、上下又は斜め方向に隣接する８点のうち、いずれかの方向の隣接２点の輝度データの各々の任意の優先順位を付けた比較演算、平均値演算等の各種の演算処理を行う演算処理部７と、演算処理部７の演算結果から検査対象点Ａとの比較に用いる最適な比較方向の２点の選定を行う選定部８と、選定した最適な比較方向の２点の平均値と検査対象点Ａの輝度データとを比較し、当該検査対象点Ａの欠陥の有無の検出を行う欠陥検出部９と、欠陥検出用の画像の表示を行う表示部１０とを有している。
【００１０】
次に、上述した構成の撮像検査装置を使用した８近傍点隣接比較方式による欠陥検出方法を以下に詳述する。
本実施の形態１の８近傍点隣接比較方式による欠陥検出方法は、撮像系１により液晶板２０を撮像し、画像処理部５により欠陥検出用の画像データを生成して図２に原理的に示すような８近傍点隣接比較による検査対象点Ａの欠陥検出を行う。
すなわち、まず、演算処理部７は、制御部６の制御の基に、検査対象点Ａを挟んで隣接する８点のうち、左右、上下又は斜め方向に隣接する３種の２点同士を優先順位（例えば左右、上下、斜めの順）を付けて順に比較演算し、比較対象の適否判定を行う予備判定を実行する。
【００１１】
具体的には、検査対象点Ａの左右の２点（２），（６）の比較演算を行い、これら２点（２），（６）の輝度データが等しい場合には両側パターン同士は同様の形状が繰り返すパターンであると又は横方向の直線上にあると判定され、次に欠陥検出部９は２点（２），（６）の輝度データの平均値に明欠陥検出用の閾値ＴｈＢ（１００％以上）又は暗欠陥検出用の閾値ＴｈＤ（１００％以下）を乗じて、検査対象点Ａの輝度データとの比較演算を行い、当該検査対象点Ａの欠陥の有無の検出を行う。
すなわち、欠陥検出部９は検査対象点Ａの輝度データが、２点（２），（６）の輝度データの平均値に閾値ＴｈＢを乗じた値より大きい場合には、検出結果は明欠陥となる。また、検査対象点Ａの輝度データが、２点（２），（６）の輝度データの平均値に閾値ＴｈＤを乗じた値より小さい場合には、検出結果は暗欠陥となる。
【００１２】
次に、検査対象点Ａの左右の２点（２），（６）の比較演算を行い、これら２点（２），（６）の輝度データが等しくないと判定した場合には、選定部８は検査対象点Ａとの比較に用いる最適な比較方向の２点の選定を縦方向に切り替える。これにより、演算処理部７は、自動的に上下の２点（４），（８）の比較演算を行い、これら２点（４），（８）の輝度データが等しい場合には、次に欠陥検出部９は２点（４），（８）の輝度データの平均値に明欠陥検出用の閾値ＴｈＢ（１００％以上）又は暗欠陥検出用の閾値ＴｈＤ（１００％以下）を乗じて、検査対象点Ａの輝度データとの比較演算を行い、当該検査対象点Ａの欠陥の有無の検出を行う。
すなわち、欠陥検出部９は検査対象点Ａの輝度データが、２点（４），（８）の輝度データの平均値に閾値ＴｈＢを乗じた値より大きい場合には、明欠陥となる。また、検査対象点Ａの輝度データが、２点（４），（８）の輝度データの平均値に閾値ＴｈＤを乗じた値より小さい場合には、暗欠陥となる。
【００１３】
更に、検査対象点Ａの上下の２点（４），（８）の比較演算を行い、これら２点（４），（８）の輝度データが等しくないと判定された場合には、選定部８は検査対象点Ａとの比較に用いる最適な比較方向の２点の選定を斜め方向に切り替える。
これにより、演算処理部７が自動的に検査対象点Ａの斜めの２点（１），（５）又は（３），（７）の比較演算を行い、以降は上述した場合と同様に欠陥検出部９による明欠陥又は暗欠陥の検出が行われる。
【００１４】
このようにして比較される点を含むパターンが例えば基板の隅に存在するような場合でも、例えば自動的に横方向の比較検査を停止し、比較方向を縦方向に切り替えることによって従来のような擬似欠陥発生を無くすことができる。
また、上述した本実施の形態１では、演算処理部７と、選定部８、欠陥検出部９を用いたハードウェア構成にて２点比較処理、欠陥検出処理を行う場合を説明したが、図３に示すような制御部（ＣＰＵ）６によるパイプライン処理（複数の制御命令を順に出力する処理）にて一連の横、縦、斜めの優先順位を付けた２点比較処理を行うことが可能である。
【００１５】
次に、図４を参照して、例えばＴＦＴ表示部のパターン像３１、配線部のパターン像３２が表面に上下左右に列設され、隅部に基板端部３３が存在する液晶板の画像データ３０に基づく８近傍点隣接比較方式による欠陥検出方法について説明する。
この場合には、検査対象点Ａ’を含むＴＦＴ表示部のパターン像３１に横方向に隣り合う点（６）と基板端部３３上の点（２）との比較では、これら両者の輝度データは大きく異なるため、前記選定部８は２点比較方向を縦方向の２点（４）’，（８）’に切り替える。
【００１６】
これにより、演算処理部７は、自動的に上下の２点（４）’，（８）’の比較演算を行い、これら２点（４）’，（８）’の輝度データが等しい場合には、次に欠陥検出部９は２点（４）’，（８）’の輝度データの平均値と検査対象点Ａ’の輝度データとの比較演算を行い、当該検査対象点Ａ’の欠陥の有無の検出を行う。検査対象点Ａ’の輝度データが２点（４）’，（８）’の輝度データの平均値より小さい場合には、検査対象点Ａ’は暗欠陥となる。
この場合も、２点比較方向を横方向から縦方向に切り替えることによって従来のような擬似欠陥発生を無くすことができる。
【００１７】
次に、検査領域別の欠陥判定について図５乃至図８を参照して説明する。
図５はＴＦＴ表示部のパターン像３１内同士の８近傍点隣接比較方式による欠陥検出方法を、図６は配線部のパターン像３２の縦配線における８近傍点隣接比較方式による欠陥検出方法を示す。
また、図７は配線部のパターン像３２の横配線における８近傍点隣接比較方式による欠陥検出方法を示し、図８は配線部のパターン像３２の配線の交差点上の８近傍点隣接比較方式による欠陥検出方法を示す。
【００１８】
図５乃至図８において、検査対象点Ａの周辺領域を明暗に分けて８ビットのコードで表すと、下記表１のようになる。
【００１９】
【表１】

【００２０】
この結果、検査対象点Ａの周辺領域の値により各々独立の閾値と欠陥判定基準を与え、ＴＦＴ表示部と配線部との検査条件を変えることで、ＴＦＴ表示部、配線部の欠陥検査を同時に行う可能となる。
また、図８に示す交差点（横配線、縦配線の交差部分）における検査対象点Ａの場合、交差点を基準に特定の位置を検査対象点Ａとして特定でき、これにより、任意の位置にグレイレベルによらずに検査対象点Ａを指定して検査条件を設定できる。
【００２１】
（実施の形態２）
次に、図９乃至図１２を参照して本発明の実施の形態２について説明する。
図９は、ＤＲＣ（Design Rule Check）法を適用する方向変化の少ないランダムパターンを示すものである。ＤＲＣ法は微小欠陥検出に有効である。
【００２２】
図９に示すランダムパターンの検査対象点Ａに対して、横（左右）方向の隣接２点を比較し欠陥検出を行う場合には、図１０左欄に示すように、横方向の隣接２点（２），（６）に対する既述した場合と同様な比較演算が行われ、これら２点（２），（６）の輝度データが等しい場合には両側パターン同士は同様の形状が繰り返すパターンであると又は横方向の直線上にあると判定され、次に欠陥検出部９は２点（２），（６）の輝度データの平均値と検査対象点Ａとの比較を行い、検査対象点Ａの輝度データが２点（２），（６）の輝度データの平均値より小さいので検査対象点Ａは暗欠陥であるとする。
逆に、図１０右欄に示す場合には、検査対象点Ａの輝度データが２点（２），（６）の輝度データの平均値より大きいので検査対象点Ａは明欠陥であるとする。
【００２３】
次に、図９に示すランダムパターンの検査対象点Ａに対して、縦（上下）方向の隣接２点（４），（８）を比較し欠陥検出を行う場合には、図１１に示すように、縦方向の隣接２点（４），（８）に対する既述した場合と同様な比較演算が行われ、これら２点（４），（８）の輝度データが等しい場合には両側パターン同士は同様の形状が繰り返すパターンであると又は縦方向の直線上にあると判定され、次に欠陥検出部９は２点（４），（８）の輝度データの平均値と検査対象点Ａとの比較を行い、検査対象点Ａの輝度データが２点（４），（８）の輝度データの平均値より小さい図１１の上側に示す例の場合には検査対象点Ａは暗欠陥であるとする。
また、検査対象点Ａの輝度データが２点（４），（８）の輝度データの平均値より大きい図１１の下側に示す例の場合には検査対象点Ａは明欠陥であるとする。
【００２４】
次に、図９に示すランダムパターンの検査対象点Ａに対して、斜め方向の隣接２点（３），（７）を比較し欠陥検出を行う場合には、図１２に示すように、斜め方向の隣接２点（３），（７）に対する既述した場合と同様な比較演算が行われ、これら２点（３），（７）の輝度データが等しい場合には両側パターン同士は同様の形状が繰り返すパターンであると又は斜め方向の直線上にあると判定され、次に欠陥検出部９は２点（３），（７）の輝度データの平均値と検査対象点Ａとの比較を行い、検査対象点Ａの輝度データが２点（３），（７）の輝度データの平均値より小さい場合には検査対象点Ａは暗欠陥であるとし、また検査対象点Ａの輝度データが２点（３），（７）の輝度データの平均値より大きい場合には検査対象点Ａは明欠陥であるとする。
【００２５】
このように、本実施の形態１、２によれば、検査対象点Ａと隣接点との比較を行う前に、まず優先順位を付けつつ左右、上下又は斜め方向の隣接する比較点２点が同じレベルの輝度を有しているか否かを判断する予備判定を行い、予備判定を行った後、初めてこの両点が比較対象として適切であるとしてこの両点の平均値と検査対象点Ａとを比較し明、暗の判定により欠陥検出を行うものである。
【００２６】
この結果、特に液晶板等のパターンエッジ部での擬似欠陥を発生を防止し、高精度に検査対象点Ａの欠陥検出を行うことができる。
なお、本発明は面積をもったパターン上の欠陥検出の他、直線状のパターンの欠陥検出にも適用可能である。
【００２７】
【発明の効果】
本発明によれば、特に基板等の隅部領域等においても擬似欠陥が発生することがなく、高精度で検査対象点の欠陥検出を行うことができる撮像検査装置における８近傍点隣接比較方式による欠陥検出方法を提供できる。
また本発明によれば、撮像系と、画像処理部と、制御部と、画像処理部と、演算処理部と、選定部と、欠陥検出部とを有する構成で、本発明の欠陥検出方法を実現し、特に基板等の隅部領域等においても擬似欠陥が発生することがなく、高精度で検査対象点の欠陥検出を行うことができる撮像検査装置における８近傍点隣接比較方式による欠陥検出システムを提供できる。
【図面の簡単な説明】
【図１】本発明の実施の形態１の撮像検査装置の概略ブロック図である。
【図２】本実施の形態１の８近傍点隣接比較方式による欠陥検出方法の原理的説明図である。
【図３】本実施の形態１の欠陥検出方法におけるパイプライン処理の説明図である。
【図４】本実施の形態１の８近傍点隣接比較方式による基板上のパターンの欠陥検出方法の説明図である。
【図５】本実施の形態１の８近傍点隣接比較方式による表示部内の欠陥検出方法の説明図である。
【図６】本実施の形態１の８近傍点隣接比較方式による縦配線上の欠陥検出方法の説明図である。
【図７】本実施の形態１の８近傍点隣接比較方式による横配線上の欠陥検出方法の説明図である。
【図８】本実施の形態１の８近傍点隣接比較方式による交差点上の欠陥検出方法の説明図である。
【図９】本発明の実施の形態２のランダムパターンの説明図である。
【図１０】本実施の形態２の８近傍点隣接比較方式によるランダムパターン上の横方向の欠陥検出方法の説明図である。
【図１１】本実施の形態２の８近傍点隣接比較方式によるランダムパターン上の縦方向の欠陥検出方法の説明図である。
【図１２】本実施の形態２の８近傍点隣接比較方式によるランダムパターン上の斜め方向の欠陥検出方法の説明図である。
【図１３】従来の欠陥検出方法の概略説明図である。
【符号の説明】
１撮像系
２光源
３レンズ
４撮像素子
５画像処理部
６制御部
７演算処理部
８選定部
９欠陥検出部
１０表示部
２０液晶板
３０画像データ
３１パターン像
３２パターン像
３３基板端部
Ａ検査対象点
Ａ’検査対象点[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a novel defect detection method and defect detection system based on an 8-neighbor point adjacent comparison method (or 8-point neighborhood adjacent comparison method) in an imaging inspection apparatus.
[0002]
[Prior art]
Conventionally, for example, when a defect detection is performed on an object to be inspected in which an identical pattern exists repeatedly, such as a liquid crystal pattern (for example, a TFT display unit) such as a liquid crystal plate, using the imaging inspection apparatus, the pattern group is stored in the imaging apparatus. Image detection and defect detection are performed from the image data. Such defect detection methods can be broadly divided into methods for detecting defects based on absolute pattern position information (Pattern Matching, etc.), and patterns. It is divided into a method of detecting a defect independently of the position (adjacent comparison, DRC (Design Rule Check), etc.).
When applied to LCD or PDP inspection, the latter method is generally employed because it has no alignment error and has high reliability.
[0003]
Described below is a defect detection method using a neighboring point adjacent comparison method in a conventional imaging inspection apparatus.
In the conventional defect detection method, a pattern (for example, TFT display portion pattern) is imaged by an imaging system of an imaging inspection apparatus, and image data 50 including a large number of pattern images as shown in FIG. Of the four points (1), (2), (3), (4) that are adjacent to each other across A, the left and right or upper and lower adjacent points (2), (3) or the adjacent points (1), (4) Either one of the average values (luminance data average value) is selected, and the defect of the inspection target point A is detected as a comparison point with the inspection target point A.
Alternatively, even if four adjacent points (2), (3), and adjacent points (1), (4) on the left, right, and upper and lower sides are used, the average value of the four points is compared with the inspection target point A. It was only.
[0004]
[Problems to be solved by the invention]
However, in the case of the above-described conventional defect detection method, when comparing two adjacent points in the horizontal direction (or vertical direction) in the defect detection of the pattern existing on the corner side of the substrate, one adjacent point is Since it is on the substrate, the luminance data at the two left and right (or top and bottom) points are greatly different, and this has the disadvantage that pseudo defects frequently occur.
The present invention has been made in view of the above circumstances, improves the defect detection process, and does not generate a pseudo defect even in a corner region of a substrate or the like, and can detect a defect with high accuracy. It is an object of the present invention to provide a defect detection method and a defect detection system based on an 8-neighbor point adjacent comparison method in an imaging inspection apparatus.
[0005]
[Means for Solving the Problems]
In order to solve the above-mentioned problem, the invention described in claim 1 is based on the image data of the inspection part of the same repeating pattern obtained by imaging the object to be inspected. In the defect detection method based on the 8-neighbor point adjacent comparison method in the imaging inspection apparatus that performs defect detection, priority is given to three types of two points adjacent in the left, right, up, down, or diagonal directions among the 8 points adjacent to each other with the inspection target point in between. A preliminary determination process in which rankings are compared in order and the suitability of the comparison target is determined, and a selection process in which two points in the optimal comparison direction used for comparison with the inspection target point are selected according to the determination result of the preliminary determination process; And comparing the average value of the two points in the optimum comparison direction selected in the selection process with the inspection target point, and detecting the presence or absence of a defect at the inspection target point. is there.
[0006]
According to a second aspect of the present invention, in the defect detection method according to the 8-neighbor point adjacent comparison method in the imaging inspection apparatus according to the first aspect, the preliminary determination process is performed in the left, right, up, down, or diagonal directions adjacent to each other with the inspection target point interposed therebetween. Each of the two points is prioritized in an arbitrary order and the suitability of the comparison target is determined.
[0007]
According to a third aspect of the present invention, in the defect detection method according to the 8-neighbor point adjacent comparison method in the imaging inspection apparatus according to the first or second aspect, the preliminary determination process is performed in a horizontal, vertical, or diagonal direction by pipeline processing in computer control. The two types of two points adjacent to each other are given priority and compared in order.
According to the defect detection method of the present invention, the defect detection process is improved, and among the eight points adjacent to each other with the inspection target point interposed, three types of two points adjacent in the left, right, up, down, or diagonal directions are prioritized. A preliminary determination process is performed to compare and determine the suitability of the comparison target in order, and two points in the optimal comparison direction used for comparison with the inspection target point are selected according to the determination result of the preliminary determination process. Since the average value of the two points in the comparison direction is compared with the inspection target point and the presence or absence of a defect at the inspection target point is detected, a pseudo defect occurs particularly in a corner region of the substrate or the like. Therefore, the defect detection of the inspection target point can be performed with high accuracy.
[0008]
According to a fourth aspect of the present invention, there is provided an imaging inspection apparatus for detecting defects in individual inspection parts by an 8-neighbor point adjacent comparison method based on image data of inspection parts having the same repeated pattern obtained by imaging the inspection object. In a defect detection system based on an 8-neighbor point adjacent comparison method, an image of an object to be inspected and an image forming system including a light source and a lens for forming an image on the image sensor, and photoelectrically converted image data output from the image sensor are subjected to image processing. An image processing unit that generates image data for defect detection, a control unit that performs overall control, and adjacent to each other in the left, right, up, down, or diagonal directions across the inspection target point in the image data under the control of the control unit 8 Among the points, an arithmetic processing unit for performing various arithmetic processing such as comparison calculation and average value calculation for each of the luminance data of two adjacent points in either direction, and the calculation result of the arithmetic processing unit. The selection unit that selects two points in the optimum comparison direction used for comparison with the target point, and the average value of the two selected points in the optimum comparison direction are compared with the luminance data of the point to be inspected. And a defect detector for detecting the presence or absence of the defect.
According to the defect detection system of the present invention, the defect described above has a configuration including the imaging system, the image processing unit, the control unit, the image processing unit, the arithmetic processing unit, the selection unit, and the defect detection unit. A detection method is realized, and a defect at an inspection target point can be detected with high accuracy without generating a pseudo defect particularly in a corner region of a substrate or the like.
[0009]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail.
(Embodiment 1)
FIG. 1 is a schematic block diagram showing the overall configuration of the imaging inspection apparatus according to Embodiment 1 of the present invention. This imaging inspection apparatus has, for example, a large number of identical patterns, that is, TFT display portions and wiring portion patterns vertically on the surface. The imaging system 1 including the light source 2 and the lens 3 that images the liquid crystal plates 20 arranged on the left and right and forms an image on the imaging device (CCD device) 4, and the photoelectrically converted image data output from the imaging device 4 An image processing unit 5 that performs image processing to generate image data for defect detection, a control unit (CPU) 6 that performs overall control, and an inspection target point A in the image data under the control of the control unit 6 Arithmetic that performs various arithmetic processing such as comparison calculation and average value calculation with arbitrary priority of luminance data of two adjacent points in any direction among eight points adjacent in the left, right, up, down, or diagonal directions Processing unit 7 and arithmetic processing unit The selection unit 8 that selects two points in the optimum comparison direction used for comparison with the inspection target point A from the calculation result of the above, the average value of the selected two points in the optimum comparison direction, and the luminance data of the inspection point A And a defect detection unit 9 that detects the presence or absence of a defect at the inspection target point A and a display unit 10 that displays an image for defect detection.
[0010]
Next, a defect detection method based on the 8-neighbor point adjacent comparison method using the imaging inspection apparatus having the above-described configuration will be described in detail below.
In the defect detection method by the 8-neighbor point adjacent comparison method of the first embodiment, the liquid crystal plate 20 is imaged by the imaging system 1, and image data for defect detection is generated by the image processing unit 5, and the principle is shown in FIG. The defect detection of the inspection target point A is performed by the 8-neighbor point adjacent comparison as shown.
That is, first, the arithmetic processing unit 7 gives priority to two types of two points that are adjacent in the left, right, up, down, or diagonal directions among the eight points that are adjacent to each other with the inspection target point A sandwiched between them under the control of the control unit 6. Preliminary determination is performed in which ranks (for example, left, right, up, down, and diagonal order) are added and compared in order, and whether the comparison target is appropriate is determined.
[0011]
Specifically, the two left and right points (2) and (6) of the inspection target point A are compared, and when the luminance data of these two points (2) and (6) are equal, the patterns on both sides are the same. It is determined that the shape is a pattern that repeats or is on a straight line in the horizontal direction, and then the defect detection unit 9 sets the threshold value ThB for bright defect detection to the average value of the luminance data of two points (2) and (6). (100% or more) or a threshold value ThD (100% or less) for dark defect detection is multiplied and compared with the luminance data of the inspection target point A to detect the presence or absence of the defect at the inspection target point A.
That is, when the luminance data of the inspection target point A is larger than the value obtained by multiplying the average value of the luminance data of the two points (2) and (6) by the threshold value ThB, the defect detection unit 9 determines that the detection result is a bright defect. Become. Further, when the luminance data of the inspection target point A is smaller than the value obtained by multiplying the average value of the luminance data of the two points (2) and (6) by the threshold value ThD, the detection result is a dark defect.
[0012]
Next, a comparison operation is performed on the two left and right points (2) and (6) of the inspection target point A, and when it is determined that the luminance data of these two points (2) and (6) are not equal, the selection unit 8 switches the selection of two optimal comparison directions used for comparison with the inspection target point A in the vertical direction. Thereby, the arithmetic processing unit 7 automatically performs the comparison operation of the upper and lower two points (4) and (8), and when the luminance data of these two points (4) and (8) are equal, The defect detection unit 9 multiplies the average value of the luminance data of the two points (4) and (8) by a threshold ThB (100% or more) for detecting a bright defect or a threshold ThD (100% or less) for detecting a dark defect, A comparison operation with the luminance data of the inspection target point A is performed, and the presence or absence of a defect at the inspection target point A is detected.
That is, the defect detection unit 9 becomes a bright defect when the luminance data of the inspection target point A is larger than the value obtained by multiplying the average value of the luminance data of the two points (4) and (8) by the threshold value ThB. Further, when the luminance data of the inspection target point A is smaller than the value obtained by multiplying the average value of the luminance data of the two points (4) and (8) by the threshold value ThD, a dark defect occurs.
[0013]
Further, the comparison operation of the two points (4) and (8) above and below the inspection target point A is performed, and when it is determined that the luminance data of these two points (4) and (8) are not equal, the selection unit 8 switches the selection of two optimal comparison directions used for comparison with the inspection target point A in an oblique direction.
As a result, the arithmetic processing unit 7 automatically performs the comparison operation of the two oblique points (1), (5) or (3), (7) of the inspection target point A, and thereafter the defect is the same as in the case described above. The detection unit 9 detects a bright defect or a dark defect.
[0014]
Even when the pattern including the points to be compared exists in the corners of the substrate, for example, the comparison inspection in the horizontal direction is automatically stopped and the comparison direction is switched to the vertical direction as in the conventional case. Pseudo defects can be eliminated.
In the first embodiment described above, the case where the two-point comparison process and the defect detection process are performed in the hardware configuration using the arithmetic processing unit 7, the selection unit 8, and the defect detection unit 9 has been described. It is possible to perform a two-point comparison process with a series of horizontal, vertical, and diagonal priorities by pipeline processing (processing for sequentially outputting a plurality of control instructions) by the control unit (CPU) 6 as shown in FIG. It is.
[0015]
Next, referring to FIG. 4, for example, a pattern image 31 of the TFT display portion and a pattern image 32 of the wiring portion are arranged on the surface vertically and horizontally, and the image data of the liquid crystal plate having the substrate end portion 33 at the corner portion. A defect detection method based on the 8-neighbor point adjacent comparison method based on the number 30 will be described.
In this case, in comparison between the point (6) laterally adjacent to the pattern image 31 of the TFT display portion including the inspection target point A ′ and the point (2) on the substrate end 33, luminance data of both of them is obtained. The selection unit 8 switches the two-point comparison direction to two longitudinal points (4) ′ and (8) ′.
[0016]
Thereby, the arithmetic processing unit 7 automatically performs the comparison operation of the upper and lower two points (4) ′ and (8) ′, and the luminance data of these two points (4) ′ and (8) ′ are equal. Next, the defect detection unit 9 performs a comparison operation between the average value of the luminance data of the two points (4) ′ and (8) ′ and the luminance data of the inspection target point A ′, and the defect of the inspection target point A ′. Detect the presence or absence of. When the luminance data of the inspection target point A ′ is smaller than the average value of the luminance data of the two points (4) ′ and (8) ′, the inspection target point A ′ becomes a dark defect.
In this case as well, the occurrence of pseudo defects as in the prior art can be eliminated by switching the two-point comparison direction from the horizontal direction to the vertical direction.
[0017]
Next, defect determination for each inspection area will be described with reference to FIGS.
FIG. 5 shows a defect detection method using the 8-neighbor point adjacent comparison method in the pattern image 31 of the TFT display unit, and FIG. 6 shows a defect detection method using the 8-neighbor point adjacent comparison method in the vertical wiring of the pattern image 32 of the wiring unit. .
FIG. 7 shows a defect detection method by the 8-neighbor point adjacent comparison method in the horizontal wiring of the pattern image 32 of the wiring portion, and FIG. 8 shows an 8-neighbor point adjacent comparison method at the wiring intersection of the pattern image 32 of the wiring portion. Defect detection method is shown.
[0018]
In FIG. 5 to FIG. 8, when the peripheral area of the inspection target point A is divided into light and dark and represented by an 8-bit code, the following Table 1 is obtained.
[0019]
[Table 1]

[0020]
As a result, independent threshold values and defect determination criteria are given according to the values in the peripheral region of the inspection target point A, and the defect inspection of the TFT display unit and the wiring unit is performed simultaneously by changing the inspection conditions between the TFT display unit and the wiring unit. Can be done.
In addition, in the case of the inspection target point A at the intersection shown in FIG. 8 (intersection portion of the horizontal wiring and the vertical wiring), a specific position can be specified as the inspection target point A with reference to the intersection. The inspection condition can be set by specifying the inspection target point A.
[0021]
(Embodiment 2)
Next, a second embodiment of the present invention will be described with reference to FIGS.
FIG. 9 shows a random pattern with little direction change to which the DRC (Design Rule Check) method is applied. The DRC method is effective for detecting minute defects.
[0022]
In the case of performing defect detection by comparing two adjacent points in the horizontal (left and right) direction with respect to the inspection target point A of the random pattern shown in FIG. 9, as shown in the left column of FIG. A comparison operation similar to that described above for (2) and (6) is performed, and when the luminance data of these two points (2) and (6) are equal, both side patterns are patterns in which the same shape repeats. If it is determined that it is on the straight line in the horizontal direction, then the defect detection unit 9 compares the average value of the luminance data of the two points (2) and (6) with the inspection target point A to determine the inspection target point. Since the luminance data of A is smaller than the average value of the luminance data of two points (2) and (6), it is assumed that the inspection target point A is a dark defect.
Conversely, in the case shown in the right column of FIG. 10, since the luminance data of the inspection target point A is larger than the average value of the luminance data of the two points (2) and (6), the inspection target point A is assumed to be a bright defect. .
[0023]
Next, in the case where defect detection is performed by comparing two points (4) and (8) in the vertical (vertical) direction with respect to the inspection target point A of the random pattern shown in FIG. 9, as shown in FIG. In addition, a comparison operation similar to that described above for the two adjacent points (4) and (8) in the vertical direction is performed, and when the luminance data of these two points (4) and (8) are equal, Is determined to be a pattern in which the same shape is repeated or on a straight line in the vertical direction, and then the defect detection unit 9 determines the average value of the luminance data of two points (4) and (8) and the inspection target point A In the example shown on the upper side of FIG. 11 in which the luminance data of the inspection target point A is smaller than the average value of the luminance data of the two points (4) and (8), the inspection target point A is a dark defect. And
Further, in the case of the example shown on the lower side of FIG. 11 in which the luminance data of the inspection target point A is larger than the average value of the luminance data of the two points (4) and (8), it is assumed that the inspection target point A is a bright defect. .
[0024]
Next, when the defect detection is performed by comparing the adjacent two points (3) and (7) in the oblique direction with respect to the inspection target point A of the random pattern shown in FIG. 9, as shown in FIG. A comparison operation similar to that described above for the two adjacent points (3) and (7) in the direction is performed, and when the luminance data of these two points (3) and (7) are equal, the patterns on both sides are the same. It is determined that the shape is a repeated pattern or on a straight line in the oblique direction, and then the defect detection unit 9 compares the average value of the luminance data of the two points (3) and (7) with the inspection target point A. When the luminance data of the inspection target point A is smaller than the average value of the luminance data of the two points (3) and (7), it is determined that the inspection target point A is a dark defect, and the luminance data of the inspection target point A is If the average value of the luminance data of two points (3) and (7) is larger, the inspection target point A is not clearly defined. And it is.
[0025]
As described above, according to the first and second embodiments, before comparing the inspection target point A with the adjacent points, first, prioritizing priorities, two adjacent comparison points in the left, right, up, down, or diagonal directions are provided. Preliminary determination is performed to determine whether or not they have the same level of brightness, and after performing the preliminary determination, the average value of both points and the inspection target point A are considered to be appropriate for comparison for the first time. Are detected by light and dark judgments.
[0026]
As a result, it is possible to prevent the occurrence of a pseudo defect particularly at a pattern edge portion of a liquid crystal plate or the like, and to detect a defect at the inspection target point A with high accuracy.
Note that the present invention is applicable not only to detecting defects on a pattern having an area but also to detecting defects on a linear pattern.
[0027]
【The invention's effect】
According to the present invention, a pseudo defect does not occur particularly in a corner region of a substrate or the like, and the 8-neighbor point adjacent comparison method is used in an imaging inspection apparatus capable of detecting a defect of an inspection target point with high accuracy. A defect detection method can be provided.
According to the present invention, the defect detection method of the present invention is configured with an imaging system, an image processing unit, a control unit, an image processing unit, an arithmetic processing unit, a selection unit, and a defect detection unit. A defect detection system using an 8-neighbor point adjacent comparison method in an imaging inspection apparatus that can realize a defect detection of a point to be inspected with high accuracy without causing a pseudo defect even in a corner region of a substrate or the like. Can provide.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram of an imaging inspection apparatus according to a first embodiment of the present invention.
FIG. 2 is a principle explanatory diagram of a defect detection method based on an 8-neighbor point adjacent comparison method according to the first embodiment;
FIG. 3 is an explanatory diagram of pipeline processing in the defect detection method according to the first embodiment;
FIG. 4 is an explanatory diagram of a defect detection method for a pattern on a substrate according to the 8-neighbor point adjacent comparison method according to the first embodiment;
FIG. 5 is an explanatory diagram of a defect detection method in a display unit using an 8-neighbor point adjacent comparison method according to the first embodiment;
FIG. 6 is an explanatory diagram of a defect detection method on vertical wiring by the 8-neighbor point adjacent comparison method according to the first embodiment;
7 is an explanatory diagram of a defect detection method on a horizontal wiring by an 8-neighbor point adjacent comparison method according to the first embodiment; FIG.
FIG. 8 is an explanatory diagram of a defect detection method on an intersection according to the 8-neighbor point adjacent comparison method according to the first embodiment;
FIG. 9 is an explanatory diagram of a random pattern according to the second embodiment of the present invention.
10 is an explanatory diagram of a lateral defect detection method on a random pattern according to an 8-neighbor point adjacent comparison method according to the second embodiment; FIG.
FIG. 11 is an explanatory diagram of a vertical defect detection method on a random pattern according to the 8-neighbor point adjacent comparison method according to the second embodiment;
FIG. 12 is an explanatory diagram of a defect detection method in a diagonal direction on a random pattern by an 8-neighbor point adjacent comparison method according to the second embodiment;
FIG. 13 is a schematic explanatory diagram of a conventional defect detection method.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Image pick-up system 2 Light source 3 Lens 4 Image pick-up element 5 Image processing part 6 Control part 7 Arithmetic processing part 8 Selection part 9 Defect detection part 10 Display part 20 Liquid crystal board 30 Image data 31 Pattern image 32 Pattern image 33 Substrate edge part A Inspection object Point A 'inspection target point

Claims

被検査体を撮像して得られる同一繰り返しパターンの検査部位の画像データを基に、８近傍点隣接比較方式により個々の検査部位の欠陥検出を行う撮像検査装置における８近傍点隣接比較方式による欠陥検出方法において、
検査対象点を挟んで隣接する８点のうち、左右、上下又は斜め方向に隣接する３種の２点同士を優先順位を付けて順に比較し比較対象の適否判定を行う予備判定過程と、
予備判定過程の判定結果に応じて検査対象点との比較に用いる最適な比較方向の２点の選定を行う選定過程と、
選定過程にて選定した最適な比較方向の２点の平均値と検査対象点とを比較し、当該検査対象点の欠陥の有無の検出を行う過程と、
を含むことを特徴とする撮像検査装置における８近傍点隣接比較方式による欠陥検出方法。Defects by an 8-neighbor point adjacent comparison method in an imaging inspection apparatus that detects defects of individual inspection sites by an 8-neighbor point adjacent comparison method based on image data of an inspection site of the same repetitive pattern obtained by imaging the object to be inspected In the detection method,
Preliminary determination process in which three types of two points adjacent in the left, right, up, down, or diagonal direction among the eight points adjacent to each other with respect to the inspection target point are compared in order and the suitability of the comparison target is determined.
A selection process for selecting two points in the optimal comparison direction to be used for comparison with the inspection target point according to the determination result of the preliminary determination process;
Comparing the average value of the two points in the optimal comparison direction selected in the selection process with the inspection target point, and detecting the presence or absence of a defect at the inspection target point;
A defect detection method using an 8-neighbor point adjacent comparison method in an imaging inspection apparatus.

前記予備判定過程は、検査対象点を挟んで隣接する左右、上下又は斜め方向の各２点を任意の順序で優先させて比較対象の適否判定を行うことを特徴とする請求項１記載の撮像検査装置における８近傍点隣接比較方式による欠陥検出方法。2. The imaging according to claim 1, wherein the preliminary determination process determines whether or not a comparison target is appropriate by prioritizing two points in the left, right, top, bottom, and diagonal directions adjacent to each other with respect to the inspection target point in an arbitrary order. A defect detection method using an 8-neighbor point adjacent comparison method in an inspection apparatus.

前記予備判定過程は、コンピュータ制御におけるパイプライン処理により左右、上下又は斜め方向に隣接する３種の２点同士を優先順位を付けて順に比較することを特徴とする請求項１又は２記載の撮像検査装置における８近傍点隣接比較方式による欠陥検出方法。3. The imaging according to claim 1, wherein in the preliminary determination process, two types of two points adjacent in the left, right, up, down, or diagonal directions are prioritized and compared in order by pipeline processing in computer control. A defect detection method using an 8-neighbor point adjacent comparison method in an inspection apparatus.

被検査体を撮像して得られる同一繰り返しパターンの検査部位の画像データを基に、８近傍点隣接比較方式により個々の検査部位の欠陥検出を行う撮像検査装置における８近傍点隣接比較方式による欠陥検出システムにおいて、
被検査体を撮像し、撮像素子に結像する光源、レンズを含む撮像系と、撮像素子から出力される光電変換された画像データを画像処理し欠陥検出用の画像データを生成する画像処理部と、全体の制御を行う制御部と、制御部の制御の基に前記画像データにおける検査対象点を挟んで左右、上下又は斜め方向に隣接する８点のうち、いずれかの方向の隣接２点の輝度データの各々の優先順位を付けた比較演算、平均値演算等の各種の演算処理を行う演算処理部と、演算処理部の演算結果から検査対象点との比較に用いる最適な比較方向の２点の選定を行う選定部と、選定した最適な比較方向の２点の平均値と検査対象点の輝度データとを比較し、当該検査対象点の欠陥の有無の検出を行う欠陥検出部とを有することを特徴とする撮像検査装置における８近傍点隣接比較方式による欠陥検出システム。Defects by an 8-neighbor point adjacent comparison method in an imaging inspection apparatus that detects defects of individual inspection sites by an 8-neighbor point adjacent comparison method based on image data of an inspection site of the same repetitive pattern obtained by imaging the object to be inspected In the detection system,
An imaging system including a light source and a lens that images an object to be inspected and forms an image on the imaging device, and an image processing unit that performs image processing on photoelectrically converted image data output from the imaging device to generate image data for defect detection A control unit that performs overall control, and two adjacent points in any direction among eight points that are adjacent in the left, right, up, down, or diagonal directions across the inspection target point in the image data based on the control of the control unit The calculation processing unit for performing various types of calculation processing such as comparison calculation and average value calculation with each priority of luminance data, and the optimal comparison direction used for comparison with the inspection target point from the calculation result of the calculation processing unit A selection unit that selects two points, and a defect detection unit that compares the average value of the two selected points in the optimum comparison direction with the luminance data of the inspection target point and detects the presence or absence of a defect at the inspection target point; In an imaging inspection apparatus characterized by having Defect detection system according takes 8 neighboring points adjacent comparison method.