JP3591161B2 - Surface inspection equipment - Google Patents

Description

【００１９】
ここでα_１＝π／４にしたとき、α_２＝０の検光子９ａを通った光強度Ｉ_１は、Ｉ_１＝Ｉ_０ρ^２となり、α_２＝π／４の検光子９ｂを通った光強度Ｉ_２は、Ｉ_２＝Ｉ_０（１＋ρ^２＋２ρｃｏｓΔ）／２、α_２＝−π／４の検光子９ｃを通った光強度Ｉ_３は、Ｉ_２＝Ｉ_０（１＋ρ^２−２ρｃｏｓΔ）／２となる。この光強度Ｉ_１，Ｉ_２，Ｉ_３からｔａｎΨとｃｏｓΔ及び表面反射強度Ｉ_０は次式で得られる。
【００２０】
【数２】

【００２１】
次に、上記原理を使用した表面検査装置の動作を説明する。投光部２から出射されて一定速度で移動している鋼板４の表面で反射した偏光は検光子９ａ，９ｂ，９ｃを通ってリニアアレイカメラ８ａ，８ｂ，８ｃに入射する。このリニアアレイカメラ８ａ〜８ｃで反射光の光強度を検出するときに、リニアアレイカメラ８ａは前面にα_２＝０の検光子９ａが設けられているから前記光強度Ｉ_１を検出し、リニアアレイカメラ８ｂは前面にα_２＝π／４の検光子９ｂが設けられているから光強度Ｉ_２を検出し、リニアアレイカメラ８ｃは前面にα_２＝−π／４の検光子９ｃが設けられているから光強度Ｉ_３を検出する。なお、リニアアレイカメラ８ａ〜８ｃが出力する偏光画像信号の出力レベルを同じようなレベルにするためにリニアアレイカメラ８ａのゲインをリニアアレイカメラ８ｂ，８７ｃのゲインの１／２にしておくと前記式における光強度Ｉ_１はリニアアレイカメラ８ａで検出した光強度の２倍になる。
【００２２】
リニアアレイカメラ８ａ，８ｂ，８ｃで検出した光強度Ｉ_１，Ｉ_２，Ｉ_３を示す偏光画像信号はそれぞれ信号変換部１１ａ，１１ｂ，１１ｃに送られる。信号変換部１１ａ〜１１ｃは送られた光強度Ｉ_１，Ｉ_２，Ｉ_３を示す偏光画像信号を、正常部の画像を信号の基準レベルとし、正常部の画像信号が全階調の中心輝度になるようにして正規化して平坦化し、正常部に対する相対的な変化を示す画像信号に変換する。すなわち、例えば図４（ａ）に示すように、リニアアレイカメラ８ａ，８ｂ，８ｃで検出した光強度Ｉ_１，Ｉ_２，Ｉ_３を示す偏光画像信号を正常部の画像を信号の基準レベルとし、図４（ｂ）に示すように、正常部の画像信号が２５５階調の中心輝度である１２８階調になるようにして、光強度Ｉ_１，Ｉ_２，Ｉ_３を幅方向に所定の移動平均幅で平均輝度を求めて正規化して平坦化する。例えばリニアアレイカメラ８ａにｊ番目の画素の光強度をＩａｊ、所定の幅で移動平均処理を行った後の光強度をＩａｊｍとすると、〔Ｇ（Ｉａｊ−Ｉａｊｍ）／Ｉａｊｍ〕＋１２８なる計算を行う。なお、Ｇは各カメラ，画素に共通な定数である。この平坦化した正常部に対する相対的な変化を示す画像信号をそれぞれフレ−ムメモリ１２ａ，１２ｂ，１２ｃに格納する。このように正規化する前後の光強度Ｉ_１，Ｉ_２，Ｉ_３の画像を図５（ａ），（ｂ）に示す。図５（ａ）に示すように、リニアアレイカメラ８ａ，８ｂ，８ｃで検出した光強度Ｉ_１，Ｉ_２，Ｉ_３を示す偏光画像は光強度Ｉ_１，Ｉ_２，Ｉ_３の輝度に応じて濃淡が異なる画像を形成しているが、図５（ｂ）に示すように、正規化した相対的な変化を示す画像は、同じ濃度を示す正常部を基準として異常部は明るくなったり暗くなったりして濃度変化を示す。このようにリニアアレイカメラ８ａ〜８ｃで検出した光強度Ｉ_１，Ｉ_２，Ｉ_３を正常部を基準として正規化することにより、リニアアレイカメラ８ａ〜８ｃの個々のドリフト等は問題にならずに光強度Ｉ_１，Ｉ_２，Ｉ_３に応じた画像を得ることができる。
【００２３】
演算手段１３はフレ−ムメモリ１２ａ〜１２ｃに展開された光強度Ｉ_１，Ｉ_２，Ｉ_３の相対的な変化を示す画像を各画素毎に読み出して、各画素毎に振幅反射率比ｔａｎΨとｃｏｓΔと表面反射強度Ｉ_０を逐次演算して、ｔａｎΨ，ｃｏｓΔ，Ｉ_０の画像デ−タとしてｔａｎΨ記憶手段１４ａとｃｏｓΔ記憶手段１４ｂとＩ_０記憶手段１４ｃに格納する。この演算によって得られるｔａｎΨ，ｃｏｓΔ，Ｉ_０の画像は、図６（ａ），（ｂ），（ｃ）に示すように、正常部を基準とした相対値画像となり、例えば、ｃｏｓΔについてはΔ＝９０度の正常部を基準にして生成され、異常部は明るくなるか暗くなる。このｔａｎΨ，ｃｏｓΔ，Ｉ_０の相対値画像を不図示の表示装置に表示する。
【００２４】
パラメ−タ比較手段１５はｔａｎΨ記憶手段１４ａとｃｏｓΔ記憶手段１４ｂとＩ_０記憶手段１４ｃに記憶されたｔａｎΨとｃｏｓΔ及び表面反射強度Ｉ_０のレベル変化から、鋼板４の表面の模様状疵や凹凸状疵の有無とその種別を判定して出力手段１６から記録装置や表示装置に出力する。すなわち、鋼板４の異常部は、図５（ｃ）のｔａｎΨ，ｃｏｓΔ，Ｉ_０の画像に示すように、異常すなわち疵等の種別によって正常部を基準にした極性が異なり、例えば鋼板４の内部介在物に起因した疵２１はｔａｎΨ画面とｃｏｓΔ画面で正極性となり、Ｉ_０画面では負極性を示す。また、油しみの場合はｔａｎΨ画面とｃｏｓΔ画面とＩ_０画面で正，負，正極性あるいはすべて正極性となり、ｔａｎΨ，ｃｏｓΔ，Ｉ_０のレベルにより異常部の種別を弁別することができる。
【００２５】
上に述べた疵も含めて冷延鋼板における異なる疵種Ｓ，Ｔ，Ｕ，Ｖ，Ｗ，Ｘに対するｔａｎΨ，ｃｏｓΔ，Ｉ_０の極性変化を調べた結果を図７に示す。図中、プラス（＋）は正極性、マイナス（−）は負極性、０は変化無しを意味する。図７から明らかなように、ｔａｎΨ，ｃｏｓΔ，Ｉ_０の極性の１つまたは２つの組合せでは疵種を判別することはできないが、ｔａｎΨ，ｃｏｓΔ，Ｉ_０の３つの極性を組み合わせることにより疵種を明確に区分けすることができる。
【００２６】
また、冷延鋼板，鍍金鋼板における種々の疵についてｔａｎΨ，ｃｏｓΔ，Ｉ_０の絶対値を演算した後に極性を調べたところ、相対値の演算の場合と同じ結果になった。
【００２７】
このように、正常部の画像信号を２５５階調の中心輝度である１２８階調になるようにして正規化して相対的画像を生成することにより、エリプソパラメ−タｔａｎΨやｃｏｓΔの異常部における変化を明確にすることができる。すなわち、一般的な画像処理の階調は２５５階調であるが、例えばリニアアレイカメラ８ｃで検出した光強度Ｉ_３の正常部の階調度が小さくなると、図５（ａ）のＩ_３画像に示すように画像全体が非常に暗くなる。この画像デ−タを用いてエリプソパラメ−タである振幅反射率比ｔａｎΨやｃｏｓΔの絶対値を演算すると、エリプソパラメ−タｔａｎΨやｃｏｓΔの異常部における変化が明確にならない場合も生じ、疵検出に不都合が生じることもあるが、このような不都合を解消することができる。
【００２８】
また、各パラメ−タは理論的には、０≦ｔａｎΨ≦∞，−１≦ｃｏｓΔ≦１，０≦Ｉ_０≦５１０という値をとり得るが、鋼板４の表面疵検査では、例えば０．４≦ｔａｎΨ≦１．６，−０．２８≦ｃｏｓΔ≦０．２８，３２≦Ｉ_０≦２２４のように、とり得る値が限られることがある。このような場合は、現実にとり得る値が０から２５５階調に対応するようにレンジの拡大を行うことで、フレ−ムメモリ１２ａ〜１２ｃの階調を有効に使い、例えば疵検出のしきい値設定をより細かくできるようになる。
【００２９】
なお、上記各実施例は鋼板４の反射光をビ−ムスプリッタ７ａ〜７ｃで３本のビ−ムに分離してリニアアレイカメラ８ａ，８ｂ，８ｃで受光した場合について説明したが、図８に示すように、リニアアレイカメラ８ａ，８ｂ，８ｃを鋼板４の移動方向に対して位置をずらして配置して、設定位置のずれ量Ｌを考慮して鋼板４の同一位置からの反射光を検出しても良い。また、図９の上面図と図１０の側面図に示すように、リニアアレイカメラ８ａ，８ｂ，８ｃを鋼板４の移動方向と直交する同一ライン上で同じ高さの位置に設け、鋼板４の同一位置からの反射光をで同時に検出するようにしても良い。
【００３０】
また、上記各実施例は投光部２から出射した偏光を鋼板４の表面に直接入射し、その反射光をリニアアレイカメラ８ａ，８ｂ，８ｃで直接受光する場合について説明したが、図１１に示すように、光源５から偏光子６を通して所定の入射角で鋼板４の表面に入射し、その反射光をミラ−１７で反射して鋼板４の表面に対して直交するように設けたリニアアレイカメラ８ａ，８ｂ，８ｃで受光するようにしても良い。このように投光部２と受光部３を構成することにより、光学系１の設置スペ−スを小さくすることができ、オンラインにおける設置の自由度を改善することができる。この場合、ｔａｎΨ，ｃｏｓΔ，Ｉ_０の絶対的な変化を問題にしないで相対的な変化を問題にするから、アルミニュ−ム等のミラ−１７による偏光の影響を問題にしないで済み、光学系１の構成や設置の自由度を高めることができる。
【００３１】
また、上記実施例はｔａｎΨ，ｃｏｓΔ，Ｉ_０の各パラメ−タを演算により求める場合について説明したが、光強度Ｉ_１，Ｉ_２，Ｉ_３の信号レベルの組合せに対応した各パラメ−タの値を納めたテ−ブルをあらかじめ作成しておき、そのデ−ブルを参照することによりパラメ−タを求めることもできる。この場合、Ｉ_１，Ｉ_２，Ｉ_３の取り得る値が制限されていると、テ−ブルの容量が小さく、処理時間が短くてすむ。冷延鋼板，鍍金鋼板について種々の疵について調べたところ、Ｉ_１，Ｉ_２，Ｉ_３を２５６階調で処理する場合、すべての疵を対象にした場合１２８±６４の範囲の値、また、最低の等級の一部の疵を除けば１２８±３２の範囲内の値を想定しておけば良いことが確認された。また、その一部の疵についても１２８±３２を超えた値を、１２８＋３２または１２８−３２に置き換えて計算してもパラメ−タの極性は同じで、等級判定に用いるパラメ−タの値も大きく変わらないことを確認した。
【００３２】
【発明の効果】
この発明は以上説明したように、被検査面に対して偏光を入射し、その反射光を異なる方位角を有する検光子を通して受光し、異なる偏光の光強度分布を測定し、測定した光強度分布を正常部が全階調の中心輝度になるように正規化して平坦化し、正常部に対する相対的な変化を示す画像信号に変換するようにしたから、受光部の各カメラのドリフト等を問題とせずに光強度に応じた画像を生成することができる。
【００３３】
また、生成した相対的画像からエリプソパラメ−タｔａｎΨやｃｏｓΔ及び反射光強度Ｉ_０を演算するから、エリプソパラメ−タｔａｎΨやｃｏｓΔ及び反射光強度Ｉ_０の異常部における変化を明確にすることができ、シ−ト状製品の表面にある異常部の有無や種別をオンラインで精度良く検出することができる。
【００３４】
さらに、エリプソパラメ−タの絶対値を測定しないで相対値を測定するから、光学系や信号処理部の厳密さが大幅に緩和され、装置全体を低価格にすることができるとともに設置や調整も容易にすることができる。
【図面の簡単な説明】
【図１】この発明の実施例の光学系を示す構成図である。
【図２】上記実施例の信号処理部を示すブロック図である。
【図３】光学系の動作を示す配置説明図である。
【図４】上記実施例の正規化処理を示す画像濃度特性図である。
【図５】正規化処理した画像を示す画面図である。
【図６】ｔａｎΨ，ｃｏｓΔ，Ｉ_０の画像を示す画面図である。
【図７】冷延鋼板における疵種の極性特性図である。
【図８】第２の実施例の光学系を示す側面図である。
【図９】第３の実施例の光学系を示す上面図である。
【図１０】第３の実施例の光学系を示す側面図である。
【図１１】第４の実施例の光学系を示す側面図である。
【符号の説明】
１ 光学系
２ 投光部
３ 受光部
４ 鋼板
５ 光源
６ 偏光子
７ ビ−ムスプリッタ
８ リニアアレイカメラ
９ 検光子
１０ 信号処理部
１１ 信号変換部
１２ フレ−ムメモリ
１３ 演算手段
１４ａ ｔａｎΨ記憶手段
１４ｂ ｃｏｓΔ記憶手段
１４ｃ Ｉ_０記憶手段
１５ パラメ−タ比較手段[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a surface inspection device for optically detecting a surface flaw or the like of a thin steel plate or the like.
[0002]
[Prior art]
For example, as a device for optically detecting a surface flaw of a steel sheet, a method of detecting a flaw using scattering of laser light or a change in diffraction pattern is often used. This method is an effective method for detecting flaws having obvious irregularities on the surface of a steel sheet.
[0003]
On the other hand, the flaws of the steel sheet have no irregularities on the surface, and are said to be pattern flaws such as unevenness of physical properties, unevenness of micro roughness, local existence of a thin oxide film or thickness unevenness of a coating film thickness. There is something. Such a pattern-like flaw is difficult to detect by scattering of laser light or change in diffraction pattern. For example, if there is an abnormal part locally having a thick oxide film of about 400 mm on the surface of a steel sheet having an oxide film of about 100 mm in a normal part, such an abnormal part area may have poor coating in the surface treatment process. Therefore, there is a demand to detect and remove as a flaw. However, the difference in the oxide film thickness between the abnormal part and the normal part is buried in the roughness of the steel sheet surface, and cannot be detected at all by a method utilizing light scattering or diffraction.
[0004]
In order to detect a flaw that cannot be detected by the method using light scattering or diffraction, a flaw inspection method using polarized light is disclosed in, for example, JP-A-52-138183 and JP-A-58-204356. It has been disclosed. The inspection method disclosed in Japanese Patent Application Laid-Open No. 52-138183 detects the presence or absence of a defect based on whether or not the ratio of P-polarized light to S-polarized light reflected from the surface of the test object is higher than a predetermined comparison level. is there. Further, the detection method disclosed in Japanese Patent Application Laid-Open No. 58-204356 improves the S / N ratio when a surface defect is detected by irradiating the surface of the test object with light at a specific angle of incidence. It was made. Although not a flaw inspection method, a method for measuring film thickness and surface properties using polarized light is disclosed in, for example, JP-A-62-293104 and JP-A-4-58138. In the inspection method disclosed in Japanese Patent Application Laid-Open No. Sho 62-293104, the polarized light reflected from the sample is received through three analyzers having different azimuth angles, and the ellipsometric parameters at each position are determined from the light intensities of the three different polarized lights. That is, of the electric vector of the reflected light, the amplitude reflectance ratio tanΨ and the phase difference Δ between the P-polarized light component in the direction of the incident surface and the S-polarized light component in the direction perpendicular to the incident surface are calculated on the surface to be inspected. This is a method for accurately measuring the oxide film, coating thickness, or physical properties. The detection method disclosed in Japanese Patent Application Laid-Open No. 4-58138 discloses a method for detecting polarized light reflected from a sample when the polarized light reflected from the sample is guided to an image sensor via a phase shifter comprising a quarter-wave plate and an analyzer. In this method, the position of the transmission axis is changed by a predetermined angle, and the analyzer is rotated at each angle to obtain a polarization parameter for each pixel of the image sensor to accurately measure the birefringence distribution.
[0005]
[Problems to be solved by the invention]
In the inspection method disclosed in JP-A-52-138183 and JP-A-58-204356, a normal part and an abnormal part are discriminated by using polarized light, but the amplitude reflectance which is an ellipsometric parameter is used. The flaw is detected without determining the ratio tanΨ and the phase difference Δ. The flaws on the surface of a steel plate or the like are often different in optical properties from the normal part, and it can be said that such a part has a different complex refractive index from the normal part. In such a case, unless both the amplitude reflectance ratio tanΨ of the ellipsometer and the phase difference Δ are taken into account, only a part of the change in the ellipsometer can be captured, and, for example, an abnormal part is detected as an inspection result. Even if it could be done, it was not possible to distinguish whether it was oil stains, oxide film unevenness, or any abnormal deposits, etc., and determine the type and degree of abnormal parts Was difficult.
[0006]
On the other hand, the inspection method disclosed in Japanese Patent Application Laid-Open No. 62-293104 uses the amplitude reflectance ratio tan Ψ and the phase difference Δ of the ellipsometric parameter, so that oil stains, unevenness of the oxide film, May be able to discriminate the adhesion of However, this method is basically a point measurement and is not suitable for inspecting the entire surface of a steel plate or the like. For example, a large number of devices disclosed in Japanese Patent Application Laid-Open No. 62-293104 may be arranged in the width direction of a steel sheet, or one device may be scanned by means having a mechanism capable of moving at high speed in the width direction. Even if the entire surface can be scanned by devising, the signal processing unit calculates the amplitude reflectance ratio tanΨ and the phase difference Δ of the ellipsometric parameter from the polarization intensity signals for all the measurement points, and uses an image processing device or the like to detect the defect. It is necessary to determine the seed and flaw grade. However, it is necessary to process polarization intensity signals of 1000 points or more in one line in the width direction. In particular, when the ellipsometric parameter operation is performed by software operation, it takes about several tens of seconds to perform the operation. It was not possible to continuously inspect the surface of a sheet-like product such as a steel sheet passing at a speed of a fraction of 100 m online on a continuous basis. For this reason, an arithmetic processing device such as a dedicated polarization parameter is required, and the device becomes expensive.
[0007]
In addition, this method is a method of measuring the film thickness or the physical property value from the beginning, and for that purpose, it was sufficient to measure the amplitude reflectance ratio tanΨ and the phase difference Δ of the ellipsometric parameters. However, these parameters were sufficient. The data does not always match the state seen by the human eyes, and even if a human can recognize the flaw, the ellipsoparameter cannot detect a flaw that does not change.
[0008]
The inspection method disclosed in Japanese Patent Application Laid-Open No. 4-58138 is a two-dimensional enlargement of an ellipsometer used for thin film evaluation and the like. In this case, the birefringence index for each pixel is reduced. Since it is obtained, the birefringence is measured as a different value between the normal part and the abnormal part, and there is a possibility that the normal part and the abnormal part can be distinguished from each other based on the difference. However, since the phase shifter and the analyzer are mechanically rotated for measurement, in order to measure the birefringence at each position of the test object, the test object must be stopped during at least one measurement. Had to be. For this reason, it has not been possible to continuously inspect online the surface of a sheet-like product, such as a steel plate, which is continuously manufactured and sent.
[0009]
Further, any of the inspection methods disclosed in JP-A-62-293104 and JP-A-4-58138 have the following disadvantages. In other words, these methods are very sensitive as inspection methods, and are relatively weak detection strengths from other types of flaws, dirt, oil unevenness, scale, etc. Was difficult to discriminate. In particular, when inspecting a steel sheet having a surface coated with an oil film and moving on a production line, polarization parameters including both the oil film unevenness and the surface defect to be originally detected are detected, and the surface defect is detected. Could not be discriminated and detected. For this reason, it is considered that there is no possibility that it can be used to detect the surface flaw of a normal steel sheet such as a cold-rolled steel sheet, which is often coated with an oil film particularly for rust prevention, and the pattern flaw of the steel sheet It has been impossible to detect the surface flaw by optical means and to determine the type and grade of the surface flaw.
[0010]
The present invention has been made in order to improve such disadvantages, and detects a pattern-like flaw or an uneven flaw on the surface of a sheet-like product continuously with a simple configuration, and detects the type and the type of the flaw. It is an object of the present invention to obtain a surface inspection device capable of discriminating the degree.
[0011]
[Means for Solving the Problems]
A surface inspection apparatus according to the present invention has a light projecting unit, a light receiving unit, and a signal processing unit. The light projecting unit emits polarized light having a long beam in the width direction to a surface to be inspected, and the light receiving unit inspects the light receiving unit. It has three analyzers provided in the optical path of the reflected light of the surface, each having a different azimuth angle, and a linear array sensor for receiving the light transmitted through each analyzer, and receives the reflected light from the surface to be inspected. The signal processing unit normalizes and flattens the output image signal from each linear array sensor, and indicates the amplitude reflectance ratio tanΨ and the phase difference Δ which are ellipsometric parameters from the flattened image signal. characterized in that calculating the cosΔ and relative values of the reflected light intensity I _0, to determine the presence or absence of abnormality of the surface of the inspection surface from the computed amplitude reflectance ratio tanΨ and phase difference cosΔ and relative values of the reflected light intensity I ₀ And
[0012]
BEST MODE FOR CARRYING OUT THE INVENTION
In the present invention, the light projecting unit is arranged so that polarized light is incident on the entire surface in the width direction of the surface to be inspected at a constant incident angle with respect to the surface to be inspected, and the light receiving unit for receiving the reflected light from the surface to be inspected is provided in a predetermined manner. To the position. The light receiving unit includes three sets of beam splitters having a beam splitter for separating incident light into three beams, and, for example, a CCD sensor for separately entering the three separated beams and outputting image signals. An array camera is provided between the beam splitter and each of the linear array cameras, and an analyzer is provided between the beam splitter and each linear array camera to convert reflected light from the surface to be inspected into polarized light of a different vibration surface. The three analyzers are arranged such that different azimuth angles, that is, angles formed by the transmission axis and the incident surface of the surface to be inspected are, for example, 0, π / 4, and -π / 4.
[0013]
The signal processing unit normalizes and flattens the output image signal from each linear array sensor so that the normal portion has the center luminance of all gradations, and converts the image signal into an image signal indicating a change relative to the normal portion. The ellipsometric from an image signal representing a relative change with respect to the normal portion parameters - calculates the cosΔ and reflected light intensity I ₀ represents the amplitude reflectance ratio tanΨ and the phase difference Δ is data, the amplitude reflectance ratio tanΨ and phase difference cosΔ And a relative value of the reflected light intensity I ₀ are calculated to form a relative value image of tanΨ, cos Δ, I ₀ . TanΨ from the relative value image, cos, detects the presence or absence of abnormality of the surface of the inspection surface such as steel to detect relative change of I _0.
[0014]
【Example】
1 and 2 show a configuration of an embodiment of the present invention. FIG. 1 is a configuration diagram of an optical system, and FIG. 2 is a block diagram showing a signal processing unit. As shown in FIG. 1, the optical system 1 has a light projecting unit 2 and a light receiving unit 3 that receives regular reflected light. The light projecting unit 2 is for emitting polarized light at an incident angle θ = 60 degrees, for example, over the entire width direction of the inspection object, for example, the steel plate 4, and includes a light source 5 and a polarizer 6 provided on the front surface of the light source 5. The light source 5 is formed in a rod shape and emits light over the entire width of the steel plate 4. Polarizer 6 is made of, for example, the polarizing plate or polarizing filter, as shown in arrangement diagram of Figure 3, the transmission axis P is arranged so that the incident surface and the angle alpha ₁ of the steel plate 4 is [pi / 4 . The light receiving section 3 receives specularly reflected light having a reflection angle θ from the steel plate 4, and includes, for example, beam splitters 7a, 7b, 7c made of optical flats, and linear array cameras 8a, 8b, 8c made of, for example, CCDs. And analyzers 9a, 9b, 9c provided on the front surfaces of the light receiving surfaces of the linear array cameras 8a to 8c. The linear array cameras 8a to 8c are adjusted so as to see the same position on the steel plate 4. Analyzer 9a~9c is made of, for example, the polarizing plate or polarizing filter, as shown in FIG. 3, the incident surface and the angle alpha ₂ is an analyzer 9a are alpha _{2 =} 0 the transmission axes steel 4 of the analyzer 9, test The photon 9b is arranged so that α ₂ = π / 4, and the analyzer 9c is arranged so that α ₂ = −π / 4.
[0015]
The signal processing unit 10 is the signal converting unit 11a, 11b, and 11c, frame - frame memory 12a, 12b, and 12c, a calculating unit 13, a tanΨ storage unit 14a and cosΔ storage unit 14b and _{I 0} storage unit 14c, parameters - data It has a comparing means 15 and an output means 16. The signal conversion units 11a to 11c use the signals of the normal parts of the polarized image signals output from the linear array cameras 8a to 8c as reference levels, and make the signals of the normal parts have the center luminance of all gradations. The polarization image signals output from 8a to 8c are normalized and flattened, and are converted into image signals indicating a change relative to a normal part. The image signals output from the signal conversion units 11a to 11c are developed in the frame memories 12a, 12b, and 12c for each pixel. The calculating means 13 sequentially reads out the image signals at the same position on the steel plate 4 from the frame memories 12a, 12b, 12c, and detects the ellipsometric parameters, that is, the amplitude reflectance ratio tanΨ and the cos .DELTA. the surface reflection intensity _{I 0} of the reflected light is calculated and stored in tanΨ storage unit 14a and cosΔ storage unit 14b and _{I 0} storage unit 14c. The parameter comparing means 15 determines the pattern flaws and irregularities on the surface of the steel plate 4 from the tan Ψ and cos 及 び stored in the tan Ψ storage means 14a, the cos 記憶 storage means 14b and the level change of the surface reflection intensity I ₀ stored in the I ₀ storage means 14c. The presence or absence of a flaw and its type are determined.
[0016]
In describing the operation of the surface inspection apparatus configured as described above, first, based on the light intensities detected by the three linear array cameras 8a, 8b, 8c, the amplitude reflectance ratio tanΨ, cosΔ, and the reflected light of the steel plate 4 the principle of calculating the surface reflection intensity I ₀ will be described.
[0017]
As shown in FIG. 3, when the reflected light from the steel plate 4 is incident on the analyzer 9, the angle between the transmission axis P of the polarizer 6 and the transmission axis A of the analyzer 9 with the incident surface of the steel plate 4 is α ₁ , Assuming α ₂ , the light intensity I (α ₁ , α ₂ ) when the p-polarized light component and the s-polarized light component incident on and reflected by the steel plate 4 at an arbitrary incident angle i pass through the analyzer 9 is the amplitude reflectance of the p and s components r _p, When r _s expressed by the following equation.
[0018]
(Equation 1)

[0019]
Here when the α ₁ = π / 4, the light intensity _{I 1} passing through the analyzer 9a of alpha ₂ = 0 _is passed through the analyzer 9b of _I 1 = I ₀ ρ ² becomes, α ₂ = π / 4 The light intensity I _{2 obtained} is I ₂ = I ₀ (1 + ρ ² + 2ρcosΔ) / 2, and the light intensity I ₃ passing through the analyzer 9c of α ₂ = −π / 4 is I ₂ = I ₀ (1 + ρ ² −2ρcosΔ). ) / 2. From these light intensities I ₁ , I ₂ , I ₃ , tan Ψ, cos Δ, and surface reflection intensity I ₀ are obtained by the following equations.
[0020]
(Equation 2)

[0021]
Next, the operation of the surface inspection apparatus using the above principle will be described. The polarized light emitted from the light projecting unit 2 and reflected on the surface of the steel plate 4 moving at a constant speed passes through the analyzers 9a, 9b, 9c and enters the linear array cameras 8a, 8b, 8c. When detecting the light intensity of the reflected light at the linear array camera 8a to 8c, the linear array camera 8a detects the light intensity I ₁ from the analyzer 9a of alpha _{2 =} 0 on the front is provided, the linear array camera 8b detects the light intensity I ₂ from the analyzer 9b of α _{2 =} π / 4 to the front is provided, the linear array camera 8c is provided with analyzer 9c of α _{2 =} -π / 4 to the front is for detecting the light intensity I ₃ because they. In order to make the output levels of the polarization image signals output from the linear array cameras 8a to 8c similar, the gain of the linear array camera 8a is set to の of the gain of the linear array cameras 8b and 87c. the light intensity I ₁ in the formula is twice the light intensity detected by the linear array camera 8a.
[0022]
The polarization image signals indicating the light intensities I ₁ , I ₂ , I ₃ detected by the linear array cameras 8a, 8b, 8c are sent to the signal conversion units 11a, 11b, 11c, respectively. The signal converters 11a to 11c use the transmitted polarization image signals indicating the light intensities I ₁ , I ₂ , and I ₃ with the image of the normal part as the reference level of the signal, and the image signal of the normal part has the central luminance of all gradations. Is normalized and flattened, and converted into an image signal indicating a change relative to a normal part. That is, for example, as shown in FIG. 4A, the polarization image signals indicating the light intensities I ₁ , I ₂ , I ₃ detected by the linear array cameras 8a, 8b, 8c are used as the reference level of the image of the normal part. As shown in FIG. 4B, the light intensity I ₁ , I ₂ , I ₃ is set to a predetermined value in the width direction such that the image signal of the normal part has 128 gradations, which is the central luminance of 255 gradations. The average brightness is obtained by the moving average width, normalized and flattened. For example, assuming that the light intensity of the j-th pixel in the linear array camera 8a is Iaj, and that the light intensity after performing the moving average process with a predetermined width is Iajm, a calculation of [G (Iaj-Iajm) / Iajm] +128 is performed. . G is a constant common to each camera and pixel. The image signals indicating the relative change with respect to the flattened normal portion are stored in the frame memories 12a, 12b and 12c, respectively. FIGS. 5A and 5B show images of the light intensities I ₁ , I ₂ , and I ₃ before and after such normalization. As shown in FIG. 5A, the polarization images showing the light intensities I ₁ , I ₂ , I ₃ detected by the linear array cameras 8 a, 8 b, 8 c correspond to the luminances of the light intensities I ₁ , I ₂ , I _3. As shown in FIG. 5B, in the image showing the normalized relative change, the abnormal part becomes brighter or darker based on the normal part showing the same density as shown in FIG. Denotes a change in density. By normalizing the light intensities I ₁ , I ₂ , I ₃ detected by the linear array cameras 8 a to 8 c with the normal part as a reference, individual drifts of the linear array cameras 8 a to 8 c do not become a problem. Thus, an image corresponding to the light intensities I ₁ , I ₂ , I ₃ can be obtained.
[0023]
Computing means 13 frame - frame memory 12a~12c reads an image showing the relative change in light intensity is deployed _{I 1,} I 2, _{I 3} for each pixel in an amplitude reflectance ratio tanΨ for each pixel The cos Δ and the surface reflection intensity I ₀ are sequentially calculated and stored in the tan Ψ storage unit 14a, the cos Δ storage unit 14b, and the I ₀ storage unit 14c as image data of tan Ψ, cos Δ, I ₀ . TanΨ obtained by this calculation, cos, images _{I 0,} the FIG. 6 (a), the result and (b), (c), the relative value image relative to the normal portion, for example, for cos delta = 90 degrees is generated based on the normal part, and the abnormal part becomes brighter or darker. The tan, cos, and displays on the display device (not shown) the relative value image _{I 0.}
[0024]
The parameter comparing means 15 determines the pattern flaws and irregularities on the surface of the steel plate 4 from the tan Ψ and cos 及 び stored in the tan Ψ storage means 14a, the cos 記憶 storage means 14b and the level change of the surface reflection intensity I ₀ stored in the I ₀ storage means 14c. The presence or absence of the flaw and its type are determined and output from the output means 16 to a recording device or a display device. That is, the abnormal portion of the steel plate 4, the internal FIG tanΨ of 5 (c), cos, as shown in the image I _0, different polarity relative to the normal portion by the type of abnormality such as i.e. flaws, for example steel plate 4 The flaw 21 caused by the inclusion has a positive polarity on the tanΨ screen and the cosΔ screen, and has a negative polarity on the _I0 screen. The positive in tan screen and cos screens and _{I 0} screen for oil stains, negative, a positive polarity or all the positive polarity can be discriminated tan, cos, the type of the abnormal portion by the level of _{I 0.}
[0025]
FIG. 7 shows the results of examining the change in polarity of tan Ψ, cos Δ, and I ₀ with respect to different types of flaws S, T, U, V, W, and X in the cold-rolled steel sheet including the flaws described above. In the figure, plus (+) means positive polarity, minus (-) means negative polarity, and 0 means no change. As is clear from FIG. 7, the flaw type cannot be determined by one or two combinations of the polarities of tan Ψ, cos Δ, and I ₀ , but the flaw type is determined by combining the three polarities of tan Ψ, cos Δ, and I _0. Can be clearly divided.
[0026]
Further, cold-rolled steel sheet for various flaw in plated steel tan, cos, were examined polarity after calculating the absolute value of I _0, was the same result as the computation of the relative value.
[0027]
In this way, by generating a relative image by normalizing the image signal of the normal part to be 128 gradations, which is the central luminance of 255 gradations, the change in the ellipso parameter tanΨ or cos Δ in the abnormal part is obtained. Can be clarified. That is, the gradation of the common image processing is a 255 gray level, for example, gradient of the normal portion of the light intensity I ₃ detected by the linear array camera 8c is reduced, the I ₃ image shown in FIG. 5 (a) As shown, the entire image is very dark. When the absolute values of the amplitude reflectance ratios tanΨ and cosΔ, which are ellipsometer parameters, are calculated using this image data, the change in the abnormal part of the ellipsometer parameters tanΨ and cosΔ may not be clear, and the flaw detection may occur. However, such inconvenience can be solved.
[0028]
Each parameter can theoretically take a value of ₀ ≦ tanΨ ≦ ∞, −1 ≦ cosΔ ≦ 1, 0 ≦ I ₀ ≦ 510, but in the surface flaw inspection of the steel sheet 4, for example, 0.4. Possible values may be limited, such as ≦ tanΨ ≦ 1.6, −0.28 ≦ cosΔ ≦ 0.28, 32 ≦ I ₀ ≦ 224. In such a case, the range is expanded so that the actual possible value corresponds to 0 to 255 gradations, so that the gradations of the frame memories 12a to 12c are effectively used, and for example, the threshold value of the flaw detection is used. The settings can be finer.
[0029]
In each of the above embodiments, the case where the reflected light from the steel plate 4 is split into three beams by the beam splitters 7a to 7c and received by the linear array cameras 8a, 8b and 8c has been described. As shown in the figure, the linear array cameras 8a, 8b, and 8c are arranged so as to be shifted with respect to the moving direction of the steel plate 4, and the reflected light from the same position of the steel plate 4 is taken into consideration in consideration of the shift amount L of the set position. It may be detected. Also, as shown in the top view of FIG. 9 and the side view of FIG. 10, the linear array cameras 8a, 8b, 8c are provided at the same height on the same line orthogonal to the moving direction of the steel plate 4, Light reflected from the same position may be simultaneously detected.
[0030]
In each of the above embodiments, the case where the polarized light emitted from the light projecting unit 2 is directly incident on the surface of the steel plate 4 and the reflected light is directly received by the linear array cameras 8a, 8b, 8c has been described. As shown, a linear array is provided so that light from a light source 5 passes through a polarizer 6 at a predetermined angle of incidence on the surface of the steel plate 4 and the reflected light is reflected by a mirror 17 so as to be orthogonal to the surface of the steel plate 4. Light may be received by the cameras 8a, 8b, 8c. By configuring the light projecting unit 2 and the light receiving unit 3 in this way, the installation space of the optical system 1 can be reduced, and the degree of freedom of online installation can be improved. In this case, since the relative change is not considered without considering the absolute change of tan Ψ, cos Δ, and I ₀ , the influence of the polarization by the mirror 17 such as aluminum does not need to be considered, and the optical system can be prevented. It is possible to increase the degree of freedom of the configuration and installation of 1.
[0031]
In the above-described embodiment, the case where each parameter of tan Ψ, cos Δ, and I ₀ is obtained by calculation is described. However, each parameter corresponding to the combination of the signal levels of the light intensities I ₁ , I ₂ , and I ₃ is described. A table containing values can be prepared in advance, and parameters can be obtained by referring to the table. In this case, if the possible values of I ₁ , I ₂ and I ₃ are limited, the capacity of the table is small and the processing time is short. When various flaws were examined for the cold-rolled steel sheet and the plated steel sheet, when I ₁ , I ₂ , and I ₃ were processed in 256 gradations, when all the flaws were targeted, a value in the range of 128 ± 64 was obtained. It has been confirmed that it is sufficient to assume a value in the range of 128 ± 32 except for some flaws of the lowest grade. Even if some of the flaws are calculated by replacing the value exceeding 128 ± 32 with 128 + 32 or 128−32, the polarity of the parameter is the same, and the value of the parameter used for class judgment is large. Confirmed that it does not change.
[0032]
【The invention's effect】
As described above, the present invention applies polarized light to the surface to be inspected, receives the reflected light through analyzers having different azimuth angles, measures the light intensity distribution of different polarized light, and measures the measured light intensity distribution. Is normalized so that the normal part has the center luminance of all gradations, and is converted to an image signal indicating a change relative to the normal part. An image corresponding to the light intensity can be generated without the need.
[0033]
In addition, since the ellipso parameter tan Ψ, cos Δ, and reflected light intensity I ₀ are calculated from the generated relative image, it is necessary to clarify changes in the abnormal part of the ellipso parameter tan Ψ, cos Δ, and reflected light intensity I _0. Thus, the presence or absence and type of an abnormal portion on the surface of the sheet-like product can be accurately detected online.
[0034]
Furthermore, since the relative value is measured without measuring the absolute value of the ellipsometric parameter, the strictness of the optical system and the signal processing unit is greatly reduced, so that the entire apparatus can be reduced in cost and installation and adjustment can be performed. Can be easier.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing an optical system according to an embodiment of the present invention.
FIG. 2 is a block diagram showing a signal processing unit of the embodiment.
FIG. 3 is a layout explanatory view showing the operation of the optical system.
FIG. 4 is an image density characteristic diagram showing a normalization process of the embodiment.
FIG. 5 is a screen diagram showing an image subjected to normalization processing.
[6] tan, cos, is a screen diagram showing an image of _{I 0.}
FIG. 7 is a polar characteristic diagram of a flaw type in a cold-rolled steel sheet.
FIG. 8 is a side view showing an optical system according to a second embodiment.
FIG. 9 is a top view illustrating an optical system according to a third embodiment.
FIG. 10 is a side view showing an optical system according to a third embodiment.
FIG. 11 is a side view showing an optical system according to a fourth embodiment.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Optical system 2 Light-emitting part 3 Light-receiving part 4 Steel plate 5 Light source 6 Polarizer 7 Beam splitter 8 Linear array camera 9 Analyzer 10 Signal processing part 11 Signal conversion part 12 Frame memory 13 Operation means 14a tan {storage means 14b cosΔ storage means 14c _{I 0} storage means 15 parameters - data comparison means

Claims (1)

A light emitting unit, a light receiving unit, and a signal processing unit,
A light-emitting unit for projecting polarized light having a long beam in the width direction onto a surface to be inspected, a light-receiving unit provided on an optical path of light reflected from the surface to be inspected, and three analyzers having different azimuth angles; A linear array sensor that receives the light transmitted through each analyzer, and converts the reflected light from the surface to be inspected into an image signal, and the signal processing unit normalizes the output image signal from each linear array sensor From the flattened image signal, the amplitude reflectance ratio tan あ る, which is an ellipsometric parameter, and the relative values of cos を indicating the phase difference 及 び and the reflected light intensity I ₀ are calculated, and the calculated amplitude reflectance ratio tan Ψ and the position are calculated. retardation cosΔ and surface inspection apparatus characterized by determining the presence or absence of abnormality on the surface of the inspected surface from the relative values of the reflected light intensity I _0.

Images